PDA

View Full Version : Swiss Pairing software testing and Doeberl pairings



jenni
15-04-2006, 09:33 PM
I am parallel running swiss master 5 at Doeberl. Just the Premier and the Seniors. We are doing the seniors, because Charles wanted to see what happened in a small tournament.

So far (4 rounds), swiss master and swiss perfect have agreed and Charles has said the pairings were correct, so all 3 are in agreement.

If we have time during the Young Masters, I am going to parallel run against the Aus championships and one of the Girls tournaments. Charles and I will then document differences etc. At the end of all this a report will go to the ACF.

Bill Gletsos
17-04-2006, 12:11 AM
It will be interesting to hear what SP and swiss master 5 produced as round 6 pairings. ;)

jase
17-04-2006, 01:26 AM
So far (4 rounds), swiss master and swiss perfect have agreed and Charles has said the pairings were correct, so all 3 are in agreement.


Jenni, can you offer a few words on the manner of this trial? Specifically is it being done independently?

Obviously neither SwissPerfect or SwissMaster5 are doing their pairings with the assistance of verification of the other, but where Charles arrives at a different pairing to that of SP (and events in Brisbane showed that to be quite regularly beyond the 3rd round), is SwissMaster5 being used as a pairing tool, or is its output completely independent and truly being run as a trial process?

Garvinator
17-04-2006, 08:11 AM
the closet gm has said in his article that the draw was changed from Rogers/Johansen to Goldenberg/Rogers.

Details about why?

jenni
17-04-2006, 06:54 PM
So far (4 rounds), swiss master and swiss perfect have agreed and Charles has said the pairings were correct, so all 3 are in agreement.


Jenni, can you offer a few words on the manner of this trial? Specifically is it being done independently?

Obviously neither SwissPerfect or SwissMaster5 are doing their pairings with the assistance of verification of the other, but where Charles arrives at a different pairing to that of SP (and events in Brisbane showed that to be quite regularly beyond the 3rd round), is SwissMaster5 being used as a pairing tool, or is its output completely independent and truly being run as a trial process?

I am not quite sure what you mean Jase? Obviously when a change is made to swiss perfect, then the manual change has to be made to swiss master to keep it in sync. Otherwise we would rapidly be running 2 different tournaments.

Basically I typed in all the names for the Premier and the seniors into Swiss master and then paired the first round. It matched exactly. Results were put in and round 2 paired, once again matched exactly. Continued doing this. Charles had been instructed that no manual changes were to be done unless a major problem was perceived. This happened 3 times - once in the Premier and twice in the Seniors. The senior changes were due to the program not allowing a float twice. Shaun and Charles decided to ignore the float rule in order to get a correct pairing. Swiss mater gave the same pairing as swiss perfect.

Round 6 was changed for the premier. Shaun and Charles discussed it and pored over the rule book for about an hour and in the end felt that the change was correct, but Shaun wasn't 100% sure.

Swiss master gave the exact same pairing as swiss perfect in all cases.

The premier circumstances are going to be documented by Shaun and clarification requested

Bill Gletsos
17-04-2006, 11:44 PM
I am not quite sure what you mean Jase? Obviously when a change is made to swiss perfect, then the manual change has to be made to swiss master to keep it in sync. Otherwise we would rapidly be running 2 different tournaments.That all makes sense.

Basically I typed in all the names for the Premier and the seniors into Swiss master and then paired the first round. It matched exactly. Results were put in and round 2 paired, once again matched exactly. Continued doing this. Charles had been instructed that no manual changes were to be done unless a major problem was perceived. This happened 3 times - once in the Premier and twice in the Seniors. The senior changes were due to the program not allowing a float twice. Shaun and Charles decided to ignore the float rule in order to get a correct pairing. Swiss mater gave the same pairing as swiss perfect.With regards the seniors why did they feel the need to ignore the float rule. The claim to get the "correct pairing" seems spurious given it could be argued that by not following the pairing rules and applying the float rule they didnt get the "correct pairing".
Perhaps Shaun will explain the reasoning here.

Round 6 was changed for the premier. Shaun and Charles discussed it and pored over the rule book for about an hour and in the end felt that the change was correct, but Shaun wasn't 100% sure.

Swiss master gave the exact same pairing as swiss perfect in all cases.

The premier circumstances are going to be documented by Shaun and clarification requestedAs I understand it the maunal pairing of round 6 resulted in 5 pairings being changed (the top 5 boards).
In fact although the manual pairings seem logically correct, they do not in fact follow the Dutch pairing rules.

In my opinion the correct pairings are the ones generated by Swiss Master and Swiss Perfect.

Given my previous investigations of Swiss Master I'd be surprised if Swiss Master is wrong.

Bill Gletsos
17-04-2006, 11:52 PM
After 5 rounds the situation was as follows:

3. Johansen 4.5 WBWBW thus a Black colour pref
8. Goldenberg 4.5 BWBWB thus a White colour pref

1. Rogers 4 WBWBW thus a Black colour pref
2. Smerdon 4 BWBWB thus a White colour pref
4. Zhao 4 BWBWB thus a White colour pref and an upfloat in round 4
6. Bjelobrk 4 BWBWW thus a Black colour pref and an upfloat in round 5
15. Charles 4 WBWBW thus a Black colour pref
16. Ly 4 BWBWB thus a White colour pref
17. Hoffmann 4 WBWBW thus a Black colour pref
40. Ikeda 4 BWBWW thus a Black colour pref
45. Stojic 4 WBWBW thus a Black colour pref

The pairing for Round 6 is explained in the following.

Note that Johansen has played Goldenberg and Ikeda, Goldenberg has played Hoffmann and Stojic and Rogers has played Ly.

Since Johansen has played Goldenberg, they drop down to the next score group.
Johansen and Goldenberg comprise group S1 and all the remaining players on 4 points comprise S2 in accordnace with pairing rule A6.

Now as per C2 determine "x" as per A8.

w = those wanting a white colour preference. Thus w = 4.
b = those wanting a black colour preference. Thus b = 7.
q = (number of players in the s1+s2)/2 and rounded up. Thus q = 6.
If b > w then x = b - q. Thus x = 1.
This represents the number of pairings where colour preference cannot be met. Snce b > w it means that 2 players with a black colour preference will have to be paired.

Now as per C3 determine p from A6.

p = number of players in S1, Thus p =2.
This means that 2 pairings need to be made.

Following C4 we get:


S1 S2
Johansen Rogers
Goldenberg Smerdon
Zhao
Bjelobrk
Charles
Ly
Hoffman
Ikeda
Stojic
Thus:
Rogers V Johansen (both have a black colour pref and colour pref not met)
This is the pairing representing x = 1.
Now the remaining pairing must meet colour preferences.
Goldenberg V Smerdon does not meet colour pref
Goldenberg V Zhao doesnt meet colour pref
Goldenberg V Bjelobrk meets colour pref but Bjelobrk was an uplfloat last round
Goldenberg V Charles is a valid pairing.

Now all the remaining players are on on 4 points so in accordance with C6 restart at C2.
Therefore:
w = those wanting a white colour preference. Thus w = 3.
b = those wanting a black colour preference. Thus b = 4.
q = (number of players in the s1+s2)/2 and rounded up. Thus q = 4.
if b > w then x = b - q. Thus x = 0.

This means the remaining pairings should all get their colour preferences.

According to A6, S1 is 3 and S2 is 4 and thus q is now = 3.

So now we have:

S1 S2
Smerdon Ly
Zhao Hoffmann
Bjelobrk Ikeda
Stojic
Now by carrying out transpositions as per C7 we get:


S1 S2
Smerdon Hoffman
Zhao Ikeda
Bjelobrk Ly
Stojic

Therefore you get the pairings:
Smerdon V Hoffmann
Zhao V Ikeda
Ly V Bjelobrk

and Stojic floats down to the 3.5 score group.

That is what you get by following the Dutch pairing rules.
These are also the pairings as determined by Swiss Master 5 and Swiss Perfect.

The pairings as actually played were:
Smerdon V Johansen
Goldenberg V Rogers
Zhao V Hoffman
Ly V Bjelobrk
Charles V Ikeda

arosar
18-04-2006, 01:02 PM
So all we care about really is, was the change a correct course of action? Can somebody just tell us that?

Thanks,

AR

Carl Gorka
18-04-2006, 01:11 PM
The way I look at it is the players themselves weren't happy about the original pairings but no one was unhappy about the revised pairings. This suggests to me that revising the pairings in this instance wasn't a big issue. I was extremely happy as I had time for one more smoke before the round started:D

Bill Gletsos
18-04-2006, 01:22 PM
So all we care about really is, was the change a correct course of action? Can somebody just tell us that?The manual pairings do not conform to the Dutch pairing rules.

Both the computer programs had the pairings correct according to the Dutch pairing rules.

Bill Gletsos
18-04-2006, 01:23 PM
The way I look at it is the players themselves weren't happy about the original pairings but no one was unhappy about the revised pairings. This suggests to me that revising the pairings in this instance wasn't a big issue. I was extremely happy as I had time for one more smoke before the round started:DWhat the players believe the pairings should be isnt the issue.
The pairings should be carried out in accordance with the pairing rules.

Carl Gorka
18-04-2006, 01:32 PM
What the players believe the pairings should be isnt the issue.
The pairings should be carried out in accordance with the pairing rules.

But then the pairings rules need to be displayed in the rules of the tournament which is something I wasn't aware of, except for the usual proviso of the arbiters decision will be final. If a player thinks the pairings are not correct they have the right to complain about them and if the arbiters think the complaint valid, then they have the right to change the pairings. If no one is unhappy about the new pairings then that is surely the end of the story?

PHAT
18-04-2006, 01:33 PM
The way I look at it is the players themselves weren't happy about the original pairings but no one was unhappy about the revised pairings. This suggests to me that revising the pairings in this instance wasn't a big issue.

Ahh, but what if someone was upset?

I am usualy a fan of "common sense" when it comes to bending/breaking the rules. However, when it comes to pairings, there must be consistancy. Using SP religiously will give consistancy. OTOH, allowing DOPs to doctor at will, leads to inconsistancy and allagations of impropriety.

Bill Gletsos
18-04-2006, 01:35 PM
But then the pairings rules need to be displayed in the rules of the tournament which is something I wasn't aware of, except for the usual proviso of the arbiters decision will be final. If a player thinks the pairings are not correct they have the right to complain about them and if the arbiters think the complaint valid, then they have the right to change the pairings. If no one is unhappy about the new pairings then that is surely the end of the story?Of course the players have the right to query the pairings.
However the point is the changing of the pairings by the arbiters was incorrect as they broke the Dutch pairing rules.

Carl Gorka
18-04-2006, 01:36 PM
Ahh, but what if someone was upset?

I am usualy a fan of "common sense" when it comes to bending/breaking the rules. However, when it comes to pairings, there must be consistancy. Using SP religiously will give consistancy. OTOH, allowing DOPs to doctor at will, leads to inconsistancy and allagations of impropriety.

Like I just said above if no one complains about the new pairings then it is end of story. In this case no one did so it seems fair enough. If OTOH someone had shouted foul about the changes then the arbiters might have had a tough choice to make but would probably have to go with the SP pairings. As it is they didn't so I don't think it's a problem.

Garvinator
18-04-2006, 01:40 PM
Like I just said above if no one complains about the new pairings then it is end of story. In this case no one did so it seems fair enough. If OTOH someone had shouted foul about the changes then the arbiters might have had a tough choice to make but would probably have to go with the SP pairings. As it is they didn't so I don't think it's a problem.
as has been pointed out before quite a few times, swissperfect is more like swissimperfect, but not as much as we like to make out.

Swiss Master 5 has been proven to be correct time and again as far as I know and explains why it is on the list of fide endorsed pairing programs. Sp is not.

The main Swissperfect main weakness that I seem to notice is:

1) It will move a player down two score groups to match colours, when a legal pairing in the score group one below could have been obtained.

Maybe other arbiters can add to this list.

Bill Gletsos
18-04-2006, 01:44 PM
as has been pointed out before quite a few times, swissperfect is more like swissimperfect, but not as much as we like to make out.

Swiss Master 5 has been proven to be correct time and again as far as I know.

The main Swissperfect main weakness that I seem to notice is:

1) It will move a player down two score groups to match colours, when a legal pairing in the score group one below could have been obtained.

Maybe other arbiters can add to this list.Although this is a valid point it isnt relevant to this thread.

The computer pairings generated by SP and Swiss master followed the Dutch pairing rules and were correct.
The manual pairings didnt and as such should not have been made.

Carl Gorka
18-04-2006, 01:45 PM
Of course the players have the right to query the pairings.
However the point is the changing of the pairings by the arbiters was incorrect as they broke the Dutch pairing rules.

So you're saying that the pairings that were originally created by the computer should have stood even though an appeal was made by the players who were clearly unhappy about the draw. And that is preferable to changing the draw to something which all agree seems more valid, even though no complaints were made about the revised draw?

If this is the case then I have no arguments with your stance.

Garvinator
18-04-2006, 01:47 PM
Although this is a valid point it isnt relevant to this thread.

The computer pairings generated by SP and Swiss master followed the Dutch pairing rules and were correct.
The manual pairings didnt and as such should not have been made.
i know it wasnt relevant to the actual discussion, but I was hoping a few more ppl would add to the flaws of sp and dispel some of the myths too.

I agree the pairings shouldnt have been changed under the dutch pairing rules and have said so, but also have said why errors do occur when under time pressure.

1min_grandmaster
18-04-2006, 02:36 PM
Thank you Bill for the clear explanation for the pairings as generated by SwissPerfect. It shows that indeed the program was correct.

It is very easy for anyone to be unhappy with the pairings generated and complain. Maybe what arbiters should be demanding is a written proof of why the pairings that they publish is incorrect, and otherwise, they simply ignore the complaint, or if they feel generous, explain why the pairing is correct. The onus should be on the person who wants to change the draw why the current pairing is incorrect.

arosar
18-04-2006, 02:49 PM
I hope to have an exclusive reply from the arbiters! Keep eye on the blog for further developments.

I really should educate myself on this pairing business. During the Sinner Dinner, an attendee wondered if I knew what the change was all about. I just had to shrug my shoulders as everything appeared kosher to me.

AR

jase
18-04-2006, 02:53 PM
as has been pointed out before quite a few times, swissperfect is more like swissimperfect, but not as much as we like to make out.

Swiss Master 5 has been proven to be correct time and again as far as I know and explains why it is on the list of fide endorsed pairing programs. Sp is not.

You choose a curious moment to make these statements Garvin.
The pairing program simulation conducted during the Doeberl Cup has provided substantial kudos for SwissPerfect (and SwissMaster), with both shown to be excellent at implementing the Dutch pairing rules.

Given that this is our first experience with SwissMaster, it's rather presumptuous to make such grand claims about its accuracy. To use one tournament and one IA columnist as gospel is insufficient.

I hope that Bill's post detailing the correct procedures for the Dutch rules stimulate some consideration of the difficulties of churning out pairings quickly using this system manually, and the resultant misunderstandings of some arbiters and most players.

The simulation is a useful tool for the following discussions:
1. The correctness of SwissPerfect pairings

My view, and that of other arbiters such as Shaun Press, has been that SwissPerfect gets it right. Doeberl supports this assertion.

2. The merit of the Dutch pairing system

This simulation demonstrates that implementing Dutch rules is somewhat more complicated that the top v bottom with respect to coloursmethodology that is widely considered appropriate. This latter methodology, closely aligned to the Lim pairing system, is the methodology we are all more familiar with, and may produce pairings regarded by players and arbiter alike to be in accordance with our understanding of Swiss pairing principles. However there may at this time be no adequate program to implement it.

Garvinator
18-04-2006, 02:53 PM
I really should educate myself on this pairing business.
i would advise against this ;) you might get accused of being an arbiter :whistle:

Igor_Goldenberg
18-04-2006, 02:55 PM
The computer pairings generated by SP and Swiss master followed the Dutch pairing rules and were correct.
The manual pairings didnt and as such should not have been made.

May be "the Dutch pairing rules" aren't perfect?

Carl Gorka
18-04-2006, 02:56 PM
It is very easy for anyone to be unhappy with the pairings generated and complain. Maybe what arbiters should be demanding is a written proof of why the pairings that they publish is incorrect, and otherwise, they simply ignore the complaint, or if they feel generous, explain why the pairing is correct. The onus should be on the person who wants to change the draw why the current pairing is incorrect.

Odd turn of phase, "if they feel generous"?

So what if, as in this case, the draw did look incorrect and the arbiters agreed with the players doubts about the original draw? Should the arbiter "generously" tell the player that the computer says so and that is that?

Bill Gletsos
18-04-2006, 02:58 PM
May be "the Dutch pairing rules" aren't perfect?Whether they are "prefect" or not is immaterial.
They are the FIDE endorsed rules.
They were followed in rounds 1-5 and round 7.
There was no valid reason they should not have been followed in round 6 as well.

PHAT
18-04-2006, 03:00 PM
Like I just said above if no one complains... As it is they didn't so I don't think it's a problem.

I find it very hard to believe, (because it's imposible impossible) that very player got an easier game out of the new doctored pairing It could be that those players who got harder games may have felt pressure not to object to the new pairing. After all not many people have the balls to stare down the top players at the Doeberl. If that is the case, shaaaaame. If not the case, it only highlights the fact that doctoring has the potential to lead to shaaaaame, and ought not be done.

Therefore, I am still in favour of SwPerf pairings every time. The occational oddies of pairing is a small price to pay for long-term transparency and consistancy.

jase
18-04-2006, 03:01 PM
Maybe what arbiters should be demanding is a written proof of why the pairings that they publish is incorrect, and otherwise, they simply ignore the complaint, or if they feel generous, explain why the pairing is correct. The onus should be on the person who wants to change the draw why the current pairing is incorrect.

Sounds great! :owned:
However I do believe the onus lies with the arbiter to be able to explain pairings to the player/s. Pairings are their domain. The difficulty of doing this concisely and quickly is clear from Bill's explanation of the Dutch system.

I am occasionally sent SwissPerfect files from organisers or arbiters asking if I could verify the pairings produced by SwissPerfect. I find it difficult flicking between various windows on my monitor; often I'll write up all the pairing cards and pair the round/s manually.

In Brisbane following the Round 7 debacle (where Charles changed many pairings without informing the players, or indeed other arbiters) it took me close to an hour to be able to prove that the original SwissPerfect pairings were correct.

Carl Gorka
18-04-2006, 03:09 PM
I find it very hard to believe, (because it's imposible impossible) that very player got an easier game out of the new doctored pairing It could be that those players who got harder games may have felt pressure not to object to the new pairing. After all not many people have the balls to stare down the top players at the Doeberl. If that is the case, shaaaaame. If not the case, it only highlights the fact that doctoring has the potential to lead to shaaaaame, and ought not be done.

Therefore, I am still in favour of SwPerf pairings every time. The occational oddies of pairing is a small price to pay for long-term transparency and consistancy.

I am also happy to go with the arbiters decision and have never complained about a pairing in my life. However, I feel that if arbiters don't have the right to interfere with pairings which seem inaccurate then this should be stated on the entry. Otherwise, the final decision of the arbiters has to be accepted. Pax has just said in another thread that when there is more than one game in a day it can be difficult to work out all the intricacies of the pairing system in such a limited time. The arbiters in Doeberl gave the matter consideration and came to their decision. They were within their rights to do so.

Bill Gletsos
18-04-2006, 03:12 PM
I am also happy to go with the arbiters decision and have never complained about a pairing in my life. However, I feel that if arbiters don't have the right to interfere with pairings which seem inaccurate then this should be stated on the entry. Otherwise, the final decision of the arbiters has to be accepted. Pax has just said in another thread that when there is more than one game in a day it can be difficult to work out all the intricacies of the pairing system in such a limited time. The arbiters in Doeberl gave the matter consideration and came to their decision. They were within their rights to do so.True the arbiters were within their rights to do so, however it could be argued they should have stuck with the computerised pairings. By changing the pairings they got it wrong.

Carl Gorka
18-04-2006, 03:19 PM
True the arbiters were within their rights to do so, however it could be argued they should have stuck with the computerised pairings. By changing the pairings they got it wrong.

Swings and roundabouts. Either they stick with the computer pairings which in the short time they have to get the draw out they can't justify. Or they change the pairings to what looks right and potentially make a mistake. Either way they can't win, so to me they make a decision and as long as there are no complaints then their decision is final and any further debate is spurious.

Bill Gletsos
18-04-2006, 03:26 PM
Swings and roundabouts. Either they stick with the computer pairings which in the short time they have to get the draw out they can't justify. Or they change the pairings to what looks right and potentially make a mistake. Either way they can't win, so to me they make a decision and as long as there are no complaints then their decision is final and any further debate is spurious.It isnt spurious at all.
If they did not follow the pairing rules in making the manual pairing (and they clearly did not) then they should not have changed the computerised pairings as they had no means to support the claim that the computerised pairings were incorrect.

In fact if they had followed the pairing rules in doing the manual pairing they would not have come up with the mainual pairing that they did but instead would have gotten the same pairings as the computer programs.

1min_grandmaster
18-04-2006, 03:57 PM
So what if, as in this case, the draw did look incorrect and the arbiters agreed with the players doubts about the original draw? Should the arbiter "generously" tell the player that the computer says so and that is that?



In Brisbane following the Round 7 debacle (where Charles changed many pairings without informing the players, or indeed other arbiters) it took me close to an hour to be able to prove that the original SwissPerfect pairings were correct.

If the arbiter agrees with the players doubts then of course they should change the draw, provided that they are still following the rules of the pairing system.

I said "generously" because, as has already been stated, it can be difficult to explain why the pairings are correct, especially to people who do not understand the pairing system and just claim that the pairings are incorrect.

Would you consider it generous if the arbiters at the Doeberl Cup explained the pairing of each division for each round, given that they are pushed for time to create the pairings, and that there is always the possibility of complaints? If they had to do this, they really are going out of their way to explain pairings when they should be arbiting the tournament.

Finally, I do not consider telling the player "that the computer says so and that is that" as an explanation. A proof, such as that shown in Bill's detailed post, is what is required.

arosar
18-04-2006, 04:09 PM
We should remember Marx's "division of labour". Players play and arbiters do their thing.

I'll take on all comers. Even Kasparov.

AR

PHAT
18-04-2006, 06:42 PM
Finally, I do not consider telling the player "that the computer says so and that is that" as an explanation. A proof, such as that shown in Bill's detailed post, is what is required.

When the checkout chick says the total, do you question the functionality of the till electronics? No, it is not reasonable. The same thing with SP pairings.

You could, however, challenge the input - ie. opperator error - eg the number of bar code scans of item X. By all means check/correct the info in the crosstable. But if you want to argue with a laptop, you are as mad as a meat axe.

Oepty
18-04-2006, 07:15 PM
It isnt spurious at all.
If they did not follow the pairing rules in making the manual pairing (and they clearly did not) then they should not have changed the computerised pairings as they had no means to support the claim that the computerised pairings were incorrect.

In fact if they had followed the pairing rules in doing the manual pairing they would not have come up with the mainual pairing that they did but instead would have gotten the same pairings as the computer programs.

Bill. I have looked at the Dutch pairings rules a couple of times in detail and the exact situation which has happened here has come into mind as a possible case where the rules are not really 100% clear. It never actually says, as far as I could see, that you have to deny colour preference to the first instance where a player is paired against their colour preference. Although it is seems to me that the most logical and most likely interpretation is that you do do as SP and Swiss master did, I am not sure it is the only possible interpretation.
I am short of time so will leave it at that.
Scott

Bill Gletsos
18-04-2006, 07:24 PM
Bill. I have looked at the Dutch pairings rules a couple of times in detail and the exact situation which has happened here has come into mind as a possible case where the rules are not really 100% clear. It never actually says, as far as I could see, that you have to deny colour preference to the first instance where a player is paired against their colour preference. Although it is seems to me that the most logical and most likely interpretation is that you do do as SP and Swiss master did, I am not sure it is the only possible interpretation.
I am short of time so will leave it at that.
ScottI disagree.

In this case they are 100% clear.

p = 2 thus 2 pairings are required (Johansen & Goldenberg) from S1
x + 1 and b > w hence the pairing where colour preferences are not matched is the player with the black colour preference. That is Johansen.

As such it is clear that Johansan V Rogers meets that condition as it is impossible for Goldenberg to meet it as he has a White colour preference.

Denis_Jessop
18-04-2006, 09:09 PM
I am also happy to go with the arbiters decision and have never complained about a pairing in my life. However, I feel that if arbiters don't have the right to interfere with pairings which seem inaccurate then this should be stated on the entry. Otherwise, the final decision of the arbiters has to be accepted. Pax has just said in another thread that when there is more than one game in a day it can be difficult to work out all the intricacies of the pairing system in such a limited time. The arbiters in Doeberl gave the matter consideration and came to their decision. They were within their rights to do so.


True the arbiters were within their rights to do so, however it could be argued they should have stuck with the computerised pairings. By changing the pairings they got it wrong.

I'd be a little stronger about this. If a tournament is being run under the Dutch System pairing rules with the assistance of a computer using those rules, the arbiter is not within his or her rights to change the computer pairings unless completely satisfied that the computer has misapplied the Dutch rules.

I understand that, if two different people or computers correctly apply the Dutch rules to a given situation, the result will be the same. That is, there is a unique solution to any given situation upon correctly applying the Dutch rules. Moreover, unlike the Laws of Chess, the Dutch rules do not contain any statement that they "cannot cover all possible situations". On the contrary, they are written so as to apply to all possible situations.

If it is difficult in the time available to work out all the intricacies of applying the Dutch rules manually, as Fireeater says, that is all the more reason to rely on the computer pairings and not to change them.

Whether he means it or not, the consequence of Fireeater's argument seems to be that one is justified in making pairings based on intuition which could give rise to some interesting situations if carried too far; and once you start along that road where do you stop?

Earlier, Igor said that perhaps the Dutch rules are not perfect. That may be so, but if the event is being run under them they are what must be applied. Incidentally, assuming one has a definition of "perfect" for a situation, which is well-nigh impossible to start with, it is equally well-nigh impossible to draft a perfect set of rules for anything, not just Swiss pairings.

DJ

Carl Gorka
18-04-2006, 10:11 PM
It isnt spurious at all.


Ok, so what can be done about it now besides perhaps implementing legislation to make the creation of pairings a role purely for the computers?

Carl Gorka
18-04-2006, 10:19 PM
If the arbiter agrees with the players doubts then of course they should change the draw, provided that they are still following the rules of the pairing system.

I said "generously" because, as has already been stated, it can be difficult to explain why the pairings are correct, especially to people who do not understand the pairing system and just claim that the pairings are incorrect.

Would you consider it generous if the arbiters at the Doeberl Cup explained the pairing of each division for each round, given that they are pushed for time to create the pairings, and that there is always the possibility of complaints? If they had to do this, they really are going out of their way to explain pairings when they should be arbiting the tournament.

Finally, I do not consider telling the player "that the computer says so and that is that" as an explanation. A proof, such as that shown in Bill's detailed post, is what is required.


This whole thread seems to be an argument between practice and theory. In limited time the arbiters at Doeberl made a decision that nobody except a few people on this BB were unhappy with. I personally cannot see the problem with that.

Arbiters should explain the pairings to anyone who queries them or the arbiter is not being accountable. That isn't generosity. As for explaining every round I don't really know what you're on about. There usually aren't many complaints in a tournament about the pairings but the ones there are need to be dealt with.

Carl Gorka
18-04-2006, 10:28 PM
If it is difficult in the time available to work out all the intricacies of applying the Dutch rules manually, as Fireeater says, that is all the more reason to rely on the computer pairings and not to change them.

Whether he means it or not, the consequence of Fireeater's argument seems to be that one is justified in making pairings based on intuition which could give rise to some interesting situations if carried too far; and once you start along that road where do you stop?

DJ

I don't think the arbiters make pairings on intuition and I'm not sure how you come about that from my arguments. At Doeberl the computer generated a set of pairings that seemed to some to be off the mark. The arbiters changed the pairings to what appeared more correct. In the time they had I wouldn't blame them for making those changes. Referees in other sports get no comeback for close call decisions so why should they in chess? Once there is a codified practice for tournaments stating that pairing systems will be solely generated by computers then this sort of debate will be over and players will just have to lump whoever they get. Until then, the arbiters decision is final.

Garvinator
18-04-2006, 10:36 PM
This whole thread seems to be an argument between practice and theory. In limited time the arbiters at Doeberl made a decision that nobody except a few people on this BB were unhappy with. I personally cannot see the problem with that.
Maybe because at the time most of the people were not in full possession of the facts.
At the time, the facts were that two different pairing programs were saying one set of pairings and these were published. Only when these pairings were challenged were they more closely looked at. Clearly this means that upon first inspection the arbiters had no problem with the pairings.

Please answer this question:
1) what would your reaction have been if someone did have a problem with the change of pairings and then it was later to be found out that the change of pairings was incorrect?


Arbiters should explain the pairings to anyone who queries them or the arbiter is not being accountable. That isn't generosity. As for explaining every round I don't really know what you're on about. There usually aren't many complaints in a tournament about the pairings but the ones there are need to be dealt with.
Sometimes to check pairings when the arbiter(s) disagree with two pairing programs takes a forensic examination and takes heaps of time.
How many players would be willing to sit around for about half an hour waiting for the arbiters to confirm/change pairings?

I see that no one has really addressed the issue of time pressure on the arbiters.

Alan Shore
18-04-2006, 10:36 PM
Triple post from fireeater.... he's getting close to that CL access. ;)

Bill Gletsos
18-04-2006, 10:36 PM
Ok, so what can be done about it now besides perhaps implementing legislation to make the creation of pairings a role purely for the computers?Thats no good as computer pairings should not be assumed to be unquestionable. However just because a computer pairing doesnt look right should not mean it is automatically assumed to be wrong either.

As Denis said "If a tournament is being run under the Dutch System pairing rules with the assistance of a computer using those rules, the arbiter is not within his or her rights to change the computer pairings unless completely satisfied that the computer has misapplied the Dutch rules."

If the arbiter be it due to lack of time between rounds or some other factor cannot prove the computer pairing is wrong, he should go with the computer pairing.

However if he can clearly demonstrate the computer has not paired according to the Dutch rules there is no question that he should change the pairings.

All that is being shown here is that the pairings for the 6 round should not have been changed.

Garvinator
18-04-2006, 10:37 PM
I don't think the arbiters make pairings on intuition and I'm not sure how you come about that from my arguments. At Doeberl the computer generated a set of pairings that seemed to some to be off the mark. The arbiters changed the pairings to what appeared more correct. In the time they had I wouldn't blame them for making those changes. Referees in other sports get no comeback for close call decisions so why should they in chess? Once there is a codified practice for tournaments stating that pairing systems will be solely generated by computers then this sort of debate will be over and players will just have to lump whoever they get. Until then, the arbiters decision is final.
My bolding- the arbiters had actually made their decision when they posted the pairings in the first place. Under your interpretation, there should have been no challenge.

Bill Gletsos
18-04-2006, 10:39 PM
This whole thread seems to be an argument between practice and theory. In limited time the arbiters at Doeberl made a decision that nobody except a few people on this BB were unhappy with. I personally cannot see the problem with that.Whether the players were happy or not is totally immaterial.
Pairings are not made based on whether players are happy with them or not.
Pairings should be made in accordance with the rules.
It is that simple.

The round 6 pairings were not in accordance with the rules.

Bill Gletsos
18-04-2006, 10:41 PM
Until then, the arbiters decision is final.Actually I'll think you will find that with the exception of Article 10.2 and Article D of the Laws of Chess all decisions of the arbiter can be appealed.

Garvinator
18-04-2006, 10:41 PM
Triple post from fireeater.... he's getting close to that CL access. ;)
that would make another about to be disappointed poster:P

Carl Gorka
18-04-2006, 10:44 PM
Please answer this question:
1) what would your reaction have been if someone did have a problem with the change of pairings and then it was later to be found out that the change of pairings was incorrect?



I said in an earlier post that if a challenge was put to the change of pairings then the arbiters would have to examine it and probably would go with the computerised version unless they could prove it was wrong. At Doeberl no challenge was put to the change of pairings as they appeared to everyone to be more correct than the original computerised version. As such I still don't see the problem here?

Carl Gorka
18-04-2006, 10:46 PM
Thats no good as computer pairings should not be assumed to be unquestionable. However just because a computer pairing doesnt look right should not mean it is automatically assumed to be wrong either.

As Denis said "If a tournament is being run under the Dutch System pairing rules with the assistance of a computer using those rules, the arbiter is not within his or her rights to change the computer pairings unless completely satisfied that the computer has misapplied the Dutch rules."

If the arbiter be it due to lack of time between rounds or some other factor cannot prove the computer pairing is wrong, he should go with the computer pairing.

However if he can clearly demonstrate the computer has not paired according to the Dutch rules there is no question that he should change the pairings.

All that is being shown here is that the pairings for the 6 round should not have been changed.

But that barely answers my question about what can be done about it now, does it?

Garvinator
18-04-2006, 10:47 PM
But that barely answers my question about what can be done about it now, does it?
of course nothing can be done about it now. Have fun in the tea sofa ;)

Carl Gorka
18-04-2006, 10:49 PM
My bolding- the arbiters had actually made their decision when they posted the pairings in the first place. Under your interpretation, there should have been no challenge.

To be honest, I actually agree with you and have never challenged a tournament draw in my life. Saying that I don't have any problems with the arbiters changing the draw in this case as it appeared on the surface to be a fairer draw and no one was bothered by it.

Bill Gletsos
18-04-2006, 10:49 PM
But that barely answers my question about what can be done about it now, does it?No one is suggesting or has been suggesting anything can be done about it now.
However the arbiters and others can all learn from it.

Carl Gorka
18-04-2006, 10:50 PM
of course nothing can be done about it now. Have fun in the tea sofa ;)

Thanks:lol:

Kevin Bonham
19-04-2006, 04:54 AM
DOPs should not tweak a formally correct draw to suit their personal opinions about how the rules should be. DOPs should implement the rules as they are and take up any issues with those rules as complaints to the FIDE Pairings Committee. [EDIT: See post 66 below - I was incorrect in suspecting this applied in this particular case.]

In this case the effect of the change was that the leaders played Rogers and Smerdon instead of Rogers and Charles (and also which leader played Rogers changed). However because the Swiss system pairs according to standing in the event first with seeding only relevant where there is no other way to separate players, it assumes things like colours and float history are more important. By that stage if player A is playing so much better than player B then player A should be on a higher score. Smerdon was not at that stage playing to his rating; he had had an early upset loss.

Tweaking to create closer matchups may seem popular with players but should be avoided in cases where the draw is technically correct. It can have unexpected consequences - it's a bit of a stretch to say the arbiter could have seen this one coming, but Gareth got a fourth white in six rounds instead of getting upfloated, going to 5/6 and giving Stojic a more realistic shot at 6/7 than he otherwise might have had ... which Stojic duly took. This is a mildly artificial result - I don't think Stojic, excellent as his performance was, would merit a top three result if you looked at performance ratings.

PHAT
19-04-2006, 07:41 AM
I'd be a little stronger about this.

How about being a leader instead of a lawyer. Get an ACF motion passed that compels DOPs to accept SP pairings unless a flaw can be identified by three independant IAs.

No can do? That would be right. You keep worrying about keeping in NSWCA's good books and keep banned players in their place. You will keep. You couldn't even drive 10 minutes to put in an appearence at Doeberl :hand:

BTW you had better have one of your filthy-lawyer mates have a good look at corporation law, rights of appeal and the ACF. Maybe if the ACF worried more about promoting chess than killing it ...

Brian_Jones
19-04-2006, 09:08 AM
DOPs should not tweak a formally correct draw to suit their personal opinions about how the rules should be. DOPs should implement the rules as they are and take up any issues with those rules as complaints to the FIDE Pairings Committee.

I have to agree with this.

Also, I have stressed many times that rankings are more important than ratings!

So many top players just don't understand the difference!

Ian Rout
19-04-2006, 09:48 AM
I certainly don't agree with it. The computer should be a tool, not a rule. Accepting it's suggestions most of the time is a good principle, doing it mindlessly is a ticking bomb. Don't forget that according to other posts SP got it wrong twice in the Seniors.

antichrist
19-04-2006, 09:51 AM
How about being a leader instead of a lawyer. Get an ACF motion passed that compels DOPs to accept SP pairings unless a flaw can be identified by three independant IAs.

No can do? That would be right. You keep worrying about keeping in NSWCA's good books and keep banned players in their place. You will keep. You couldn't even drive 10 minutes to put in an appearence at Doeberl :hand:

BTW you had better have one of your filthy-lawyer mates have a good look at corporation law, rights of appeal and the ACF. Maybe if the ACF worried more about promoting chess than killing it ...

Matt, where does Denis live and you don't think you are spreading yourself too thinly but taking on another contender with all the other unfinished business?

Garvinator
19-04-2006, 10:07 AM
I certainly don't agree with it. The computer should be a tool, not a rule. Accepting it's suggestions most of the time is a good principle, doing it mindlessly is a ticking bomb. Don't forget that according to other posts SP got it wrong twice in the Seniors.
How did swiss master 5 do in the seniors?

Bill Gletsos
19-04-2006, 10:13 AM
How did swiss master 5 do in the seniors?My understanding is that Swiss Master 5 generated the same pairings as Swiss Perfect.
As such it would appear that SP did not get it wrong in the Seniors.

Bill Gletsos
19-04-2006, 10:20 AM
How about being a leader instead of a lawyer. Get an ACF motion passed that compels DOPs to accept SP pairings unless a flaw can be identified by three independant IAs.You fail to realise that the responsability for the pairings rests with the arbiter, not the computer. On top of that the computer is not infalliable.
The last part of your suggestion is essentially worthless as at the vast majority of events 3 IA's are not present.

Igor_Goldenberg
19-04-2006, 10:35 AM
Do I understand correctly that the opinion of the experts can be summarised as this:

1. The pairing done by the SP software for round 6 Doeberl was correct.
2. The arbiter had a right to manually alter the draw, which he exercised.
3. The revised draw was not with strict accordance with Dutch pairing rules.
4. The revised draw was still acceptable, despite #3 (OK, I might be putting words into someone mouth, as IMHO it still makes more sense to me then the original one).
5. The arbiter should not have changed the draw.
6. However, it is not a big transgression that requires any action.

Please correct any points.

Brian_Jones
19-04-2006, 10:39 AM
I certainly don't agree with it. The computer should be a tool, not a rule. Accepting it's suggestions most of the time is a good principle, doing it mindlessly is a ticking bomb.

Whilst I agree with this sentiment in many walks of life, I think that in a weekend chess tournament there is insufficient time to do the calculations manually without error.

As regards the Seniors I haven't yet seen proof that SP got the pairings wrong. Where is this?

Bill Gletsos
19-04-2006, 10:41 AM
Do I understand correctly that the opinion of the experts can be summarised as this:

1. The pairing done by the SP software for round 6 Doeberl was correct.
2. The arbiter had a right to manually alter the draw, which he exercised.
3. The revised draw was not with strict accordance with Dutch pairing rules.
4. The revised draw was still acceptable, despite #3 (OK, I might be putting words into someone mouth, as IMHO it still makes more sense to me then the original one).
5. The arbiter should not have changed the draw.
6. However, it is not a big transgression that requires any action.

Please correct any points.
1. True.
2. True but they should only have changed it if the SP draw was incorrect.
3. True.
4. False. It contravenes the pairing rules.
5. True.
6. True, other than that arbiters and players learn from it and hopefully avoid it in future.

Kevin Bonham
19-04-2006, 10:45 AM
I've received a lengthy explanation from shaun to the effect that the DOPs simply genuinely believed that the original draw was incorrect and overrode it on that basis - it was not a case of tweaking to fit their prejudices but simply a mistake because they did not see why the program made the pairings it did (with a colour mismatch on board 1) until Bill explained it.

My experience in similar cases (where the score groups are correctly paired but there is something a bit odd about who is paired with who) is that on closer investigation the program just about always turns out to be right, and it is safer to simply declare in advance that the program will be upheld unless obviously and glaringly wrong (eg matches across score groups violate B3).

Ian Rout
19-04-2006, 11:08 AM
I agree with Brian and Kevin that it is a good idea to accept the computer pairings almost all the time. Some arbiters do seem to like to fiddle the pairings to handicap the leader or create marquee match-ups. I don't approve of that and I wish they wouldn't do it. In other cases like this there is simply not time for the arbiter to really ascertain that the computer is wrong, and on the balance of probabilities it isn't.

BUT if you say that the computer is always right you will eventually come across an error in a bizarre set of circumstances that were not tested, or SP will produce an odd effect under the next operating system or whatever. Officials in any sport have discretion, they just have to be able to justify using it.

As an analogy, suppose you were told that your flight was entirely computer-controlled and the staff could not over-ride it, they were just there to serve drinks. Most of the time you would see this as a good thing, the computer would avoid numerous human errors; but occasionally it might be convenient if the pilot could step in.

Kevin Bonham
19-04-2006, 11:19 AM
BUT if you say that the computer is always right you will eventually come across an error in a bizarre set of circumstances that were not tested, or SP will produce an odd effect under the next operating system or whatever. Officials in any sport have discretion, they just have to be able to justify using it.

Agreed. It's the standard problem with any set of rules - if you allow someone to make exceptions then they may make exceptions needlessly, but if you don't you may find yourself needing that power. Probably best to keep it flexible at the cost of incorrect overrides from time to time rather than risk being unable to override an unexpected case of serious program error.

Garvinator
19-04-2006, 11:37 AM
1. True.
2. True but they should only have changed it if the SP draw was incorrect.
3. True.
4. False. It contravenes the pairing rules.
5. True.
6. True, other than that arbiters and players learn from it and hopefully avoid it in future.
You could add here that swiss master 5, a fide endorsed pairing program, was also giving the same pairings as swiss perfect.

Denis_Jessop
19-04-2006, 12:59 PM
I don't think the arbiters make pairings on intuition and I'm not sure how you come about that from my arguments. At Doeberl the computer generated a set of pairings that seemed to some to be off the mark. The arbiters changed the pairings to what appeared more correct. In the time they had I wouldn't blame them for making those changes. Referees in other sports get no comeback for close call decisions so why should they in chess? Once there is a codified practice for tournaments stating that pairing systems will be solely generated by computers then this sort of debate will be over and players will just have to lump whoever they get. Until then, the arbiters decision is final.

I didn't say that arbiters make pairings on intuition. I said that what you were saying would apparently lead to pairings being so made. That was because you were saying that it was OK to change pairings if the changed result seemed correct to the players. But pairings are not correctly made on perceptions. They are made according to rules, being the rules applying to the event. For a pairing just to look OK without reference to the rules you can rely only on perception or intuition.

The need to use rules and to comply with them is pretty obvious. Among other things, it provides an established basis for pairings, it allows arbiters to justify their pairings by reference to objective criteria and it protects them from allegations of bias. Once they move outside the rules they are open to criticism, especially on the last-mentioned ground.

DJ

Denis_Jessop
19-04-2006, 01:13 PM
I agree with Brian and Kevin that it is a good idea to accept the computer pairings almost all the time. Some arbiters do seem to like to fiddle the pairings to handicap the leader or create marquee match-ups. I don't approve of that and I wish they wouldn't do it. In other cases like this there is simply not time for the arbiter to really ascertain that the computer is wrong, and on the balance of probabilities it isn't.

BUT if you say that the computer is always right you will eventually come across an error in a bizarre set of circumstances that were not tested, or SP will produce an odd effect under the next operating system or whatever. Officials in any sport have discretion, they just have to be able to justify using it.

As an analogy, suppose you were told that your flight was entirely computer-controlled and the staff could not over-ride it, they were just there to serve drinks. Most of the time you would see this as a good thing, the computer would avoid numerous human errors; but occasionally it might be convenient if the pilot could step in.


The problem with this argument is that I don't think anyone is contending that the computer is always right. What is contended is that, in the case of SP, it is right almost all of the time but it has its quirks, including that pointed out by Garvin though not relevant to the present example - I have experienced it too and altered the draw. In fact FIDE states that the arbiter is finally responsible and the computer is only a tool. But what a tool! It beats pairing cards and a manual draw hands down even if the arbiter gets it right.

Incidentally, thanks Bill for the explanation. At last I understand why the funny algebriac stuff is in the rules ;) It seems I was not alone in not knowing that.

DJ

Garvinator
19-04-2006, 01:58 PM
Incidentally, thanks Bill for the explanation. At last I understand why the funny algebriac stuff is in the rules ;) It seems I was not alone in not knowing that.
i wish there was an easier way than all the algebraic stuff, as you call it, it is damn confusing to try and work out, even in the cold light of day. Never mind trying to do it with quite a few people watching and running late:eek:

Oepty
19-04-2006, 05:57 PM
I disagree.

In this case they are 100% clear.

p = 2 thus 2 pairings are required (Johansen & Goldenberg) from S1
x + 1 and b > w hence the pairing where colour preferences are not matched is the player with the black colour preference. That is Johansen.

As such it is clear that Johansan V Rogers meets that condition as it is impossible for Goldenberg to meet it as he has a White colour preference.

Bill you are right. I was applying my point to a situation where is does not apply, sorry, although I think my point stands in a slightly different situation.
Scott

Carl Gorka
20-04-2006, 10:21 AM
I didn't say that arbiters make pairings on intuition. I said that what you were saying would apparently lead to pairings being so made. That was because you were saying that it was OK to change pairings if the changed result seemed correct to the players.

DJ

No, I have been saying that the players complained about the result and the arbiters in the limited time couldn't find a reason to justify the draw that was made and so changed it manually, not to fit the wishes of the players but in accordance with the laws as they saw it. And for this I don't blame them even if it wasn't strictly according to the book. I have never mentioned intuition or perception or any other esoteric concept in how the arbiters came about the draw. You are inferring too much from what I write.

Bob1
23-04-2006, 04:58 PM
When will the chess community believe that the author of SP98 has written a program that acurately computes the draw correctly?

Bill - have you ever been able to prove a SP draw is:

a: Incorrect
b: Dubious
c: Against the basic rules

Arbiters continually score against b !

surely the algorithm needed to write a program for a tournament draw is significantly simpler than that needed actually to play the game of chess - the latter have been proved over the board.

Garvinator
23-04-2006, 06:42 PM
When will the chess community believe that the author of SP98 has written a program that acurately computes the draw correctly?
when sp98 does the draw correctly;)

I have identified one flaw of sp98 and want to know more.

Sp will downfloat a player two scoregroups to match colours, when a legal pairing exists for that player only one score group below. It does this to attempt to maximise the amount of players who get their preferred colour, but is incorrect in regards to the dutch pairing rules.


From the fide handbook:

The basic principles of a Swiss System tournament are:

1. The number of rounds to be played is declared beforehand.
2. Two players may play each other only once.
3. Players are paired with others of the same score, or nearest score.
4. When possible, a player is given the white pieces as many times as he is given the black pieces.
5. When possible, a player is given the colour other than that he was given the previous round.
6. The final ranking order is determined by the aggregate of points won: 1 point for a win, 0.5 point for a draw and 0 point for a loss. A player whose opponent fails to appear for a scheduled game receives one point.

Also: B3 of the pairing criteria:

B.3 The difference of the scores of two players paired against each other should be as small as possible and ideally zero.

From my testing, swiss master 5, a fide endorsed program, handles this much better.

Normally this flaw for sp98 occurs in the last round.

Bob1
23-04-2006, 07:06 PM
Gg
And your (exact example) has been verified by Bill !

or is this just heresay

(btw Bill is not the only person that can do the maths - I just like his precise presentation of the facts)

give us the actual example (or resign)

Garvinator
23-04-2006, 07:32 PM
Gg
And your (exact example) has been verified by Bill !

or is this just heresay

(btw Bill is not the only person that can do the maths - I just like his precise presentation of the facts)

give us the actual example (or resign)
Hello Bob,

I referred to this situation in post 16 of this thread and then this was backed up by Bill in post 17.

Denis also refers to it in post 71.


What is contended is that, in the case of SP, it is right almost all of the time but it has its quirks, including that pointed out by Garvin though not relevant to the present example - I have experienced it too and altered the draw.

Bill Gletsos
23-04-2006, 08:05 PM
Hello Bob,

I referred to this situation in post 16 of this thread and then this was backed up by Bill in post 17.I'm sorry for the confusion but I wasnt backing it up and should have been clearer in my reply in post #17 by not quoting all of your post #16.

When I said you had a valid point I was refering to your comment:

as has been pointed out before quite a few times, swissperfect is more like swissimperfect, but not as much as we like to make out.and not specifically to your comment regarding down floating players two score groups.

I have seen others make the same claim and SP may well be getting it incorrect, but why it is getting it incorrect may not be for the reasons you believe.

Therefore like Bob, I would like to see an actual example of the problem presented.

Garvinator
23-04-2006, 09:37 PM
Hello Bill,

While not completely sure if this falls into the flaw I remarked earlier, I will give a difference between sp and sm5. I believe that all information is exactly the same except for the pairings.

Round one pairings from sp:


No Name Loc Total Result Name Loc Total

1 THOMAS, Brian (1) 1743 [0] 1:0 CURRAN, John (10) 1082 [0]
2 DE VERE, Cameron (12) 1044 [0] 0:1 KIRKMAN, Phillip (2) 1724 [0]
3 ROBINSON, Andrew J (3) 1587 [0] 0:1 CHADWICK, Marty (13) 1023 [0]
4 MARTIN, Shane (14) 994 [0] 0:1 HOWES, Tony (4) 1569 [0]
5 JENKINS, Deborah (5) 1400 [0] 1:0 TAYLOR, Mark (15) 984 [0]
6 WILSON, Stuart (16) 934 [0] 0:1 ADAMS, Ray (6) 1389 [0]
7 WATERS, Mick (7) 1283 [0] 1:0 HILLER, James (17) 594 [0]
8 JOHNSON, Tim (18) [0] 0:1 LYONS, Kieran C (8) 1276 [0]
9 GRAY, Garvin (9) 1232 [0] 1:0 LYONS, Russell (19) [0]
10 MARKSON, Ellis (11) 1047 [0] .5:0 BYE

sm5 agrees.

Round Two from sp:


No Name Loc Total Result Name Loc Total

1 KIRKMAN, Phillip (2) 1724 [1] 1:0 WATERS, Mick (7) 1283 [1]
2 HOWES, Tony (4) 1569 [1] 1:0 LYONS, Kieran C (8) 1276 [1]
3 CHADWICK, Marty (13) 1023 [1] 1:0 JENKINS, Deborah (5) 1400 [1]
4 ADAMS, Ray (6) 1389 [1] 1:0 GRAY, Garvin (9) 1232 [1]
5 MARKSON, Ellis (11) 1047 [.5] .5:.5 ROBINSON, Andrew J (3) 1587 [0]
6 CURRAN, John (10) 1082 [0] 0:1 WILSON, Stuart (16) 934 [0]
7 HILLER, James (17) 594 [0] 0:1 DE VERE, Cameron (12) 1044 [0]
8 LYONS, Russell (19) [0] 1:0 MARTIN, Shane (14) 994 [0]
9 TAYLOR, Mark (15) 984 [0] 1:0 JOHNSON, Tim (18) [0]
10 THOMAS, Brian (1) 1743 [1] .5:0 BYE

Sm5 agrees

Round 3 from sp:


No Name Loc Total Result Name Loc Total

1 ADAMS, Ray (6) 1389 [2] 0:1 KIRKMAN, Phillip (2) 1724 [2]
2 CHADWICK, Marty (13) 1023 [2] 0:1 HOWES, Tony (4) 1569 [2]
3 JENKINS, Deborah (5) 1400 [1] 0:1 THOMAS, Brian (1) 1743 [1.5]
4 WATERS, Mick (7) 1283 [1] 0:1 DE VERE, Cameron (12) 1044 [1]
5 LYONS, Kieran C (8) 1276 [1] 1:0 LYONS, Russell (19) [1]
6 GRAY, Garvin (9) 1232 [1] 1:0 TAYLOR, Mark (15) 984 [1]
7 WILSON, Stuart (16) 934 [1] 1:0 MARKSON, Ellis (11) 1047 [1]
8 ROBINSON, Andrew J (3) 1587 [.5] 1:0 CURRAN, John (10) 1082 [0]
9 MARTIN, Shane (14) 994 [0] 1:0 JOHNSON, Tim (18) [0]

Sm5 agrees

Round 4 from sp:


No Name Loc Total Result Name Loc Total

1 KIRKMAN, Phillip (2) 1724 [3] 0:1 HOWES, Tony (4) 1569 [3]
2 THOMAS, Brian (1) 1743 [2.5] 1:0 ADAMS, Ray (6) 1389 [2]
3 LYONS, Kieran C (8) 1276 [2] .5:.5 CHADWICK, Marty (13) 1023 [2]
4 DE VERE, Cameron (12) 1044 [2] 1:0 GRAY, Garvin (9) 1232 [2]
5 ROBINSON, Andrew J (3) 1587 [1.5] 1:0 WILSON, Stuart (16) 934 [2]
6 LYONS, Russell (19) [1] .5:.5 JENKINS, Deborah (5) 1400 [1]
7 TAYLOR, Mark (15) 984 [1] 0:1 WATERS, Mick (7) 1283 [1]
8 MARKSON, Ellis (11) 1047 [1] 0:1 MARTIN, Shane (14) 994 [1]
9 JOHNSON, Tim (18) [0] 0:1 CURRAN, John (10) 1082 [0]

Sm5 agrees

Now for round 5 and the disagreement:

Pairing from sp:


No Name Loc Total Result Name Loc Total

1 HOWES, Tony (4) 1569 [4] : THOMAS, Brian (1) 1743 [3.5]
2 KIRKMAN, Phillip (2) 1724 [3] : LYONS, Kieran C (8) 1276 [2.5]
3 CHADWICK, Marty (13) 1023 [2.5] : DE VERE, Cameron (12) 1044 [3]
4 WATERS, Mick (7) 1283 [2] : ROBINSON, Andrew J (3) 1587 [2.5]
5 MARTIN, Shane (14) 994 [2] : ADAMS, Ray (6) 1389 [2]
6 GRAY, Garvin (9) 1232 [2] : WILSON, Stuart (16) 934 [2]
7 JENKINS, Deborah (5) 1400 [1.5] : MARKSON, Ellis (11) 1047 [1]
8 JOHNSON, Tim (18) [0] : LYONS, Russell (19) [1.5]
9 CURRAN, John (10) 1082 [1] : TAYLOR, Mark (15) 984 [1]

Pairing from sm5:


No Name Loc Total Result Name Loc Total

1 HOWES, Tony (4) 1569 [4] : THOMAS, Brian (1) 1743 [3.5]
2 KIRKMAN, Phillip (2) 1724 [3] : LYONS, Kieran C (8) 1276 [2.5]
3 CHADWICK, Marty (13) 1023 [2.5] : DE VERE, Cameron (12) 1044 [3]
4 WATERS, Mick (7) 1283 [2] : ROBINSON, Andrew J (3) 1587 [2.5]
5 MARTIN, Shane (14) 994 [2] : ADAMS, Ray (6) 1389 [2]
6 GRAY, Garvin (9) 1232 [2] : WILSON, Stuart (16) 934 [2]
7 CURRAN, John (10) 1082 [1] : JENKINS, Deborah (5) 1400 [1.5]
8 TAYLOR, Mark (15) 984 [1] : LYONS, Russell (19) [1.5]
9 JOHNSON, Tim (18) [0] : MARKSON, Ellis (11) 1047 [1]

The difference is boards 7, 8 and 9

Cross table from sm5: (aware that it doesnt show colours easily)


No Name Loc Total 1 2 3 4 5

1 HOWES, Tony 1569 4 11:W 6:W 7:W 3:W 2:
2 THOMAS, Brian 1743 3.5 15:W 0:D 13:W 8:W 1:
3 KIRKMAN, Phillip 1724 3 4:W 9:W 8:W 1:L 6:
4 DE VERE, Cameron 1044 3 3:L 18:W 9:W 10:W 7:
5 ROBINSON, Andrew J 1587 2.5 7:L 16:D 15:W 12:W 9:
6 LYONS, Kieran C 1276 2.5 19:W 1:L 14:W 7:D 3:
7 CHADWICK, Marty 1023 2.5 5:W 13:W 1:L 6:D 4:
8 ADAMS, Ray 1389 2 12:W 10:W 3:L 2:L 11:
9 WATERS, Mick 1283 2 18:W 3:L 4:L 17:W 5:
10 GRAY, Garvin 1232 2 14:W 8:L 17:W 4:L 12:
11 MARTIN, Shane 994 2 1:L 14:L 19:W 16:W 8:
12 WILSON, Stuart 934 2 8:L 15:W 16:W 5:L 10:
13 JENKINS, Deborah 1400 1.5 17:W 7:L 2:L 14:D 15:
14 LYONS, Russell 1.5 10:L 11:W 6:L 13:D 17:
15 CURRAN, John 1082 1 2:L 12:L 5:L 19:W 13:
16 MARKSON, Ellis 1047 1 0:D 5:D 12:L 11:L 19:
17 TAYLOR, Mark 984 1 13:L 19:W 10:L 9:L 14:
18 HILLER, James 594 0 9:L 4:L 0: 0: 0:
19 JOHNSON, Tim 0 6:L 17:L 11:L 15:L 16:

Cross Table from sp:


No Name Loc Total 1 2 3 4 5

1 HOWES, Tony 1569 4 11:W 6:W 7:W 3:W 2:
2 THOMAS, Brian 1743 3.5 15:W 0:D 13:W 8:W 1:
3 KIRKMAN, Phillip 1724 3 4:W 9:W 8:W 1:L 6:
4 DE VERE, Cameron 1044 3 3:L 18:W 9:W 10:W 7:
5 ROBINSON, Andrew J 1587 2.5 7:L 16:D 15:W 12:W 9:
6 LYONS, Kieran C 1276 2.5 19:W 1:L 14:W 7:D 3:
7 CHADWICK, Marty 1023 2.5 5:W 13:W 1:L 6:D 4:
8 ADAMS, Ray 1389 2 12:W 10:W 3:L 2:L 11:
9 WATERS, Mick 1283 2 18:W 3:L 4:L 17:W 5:
10 GRAY, Garvin 1232 2 14:W 8:L 17:W 4:L 12:
11 MARTIN, Shane 994 2 1:L 14:L 19:W 16:W 8:
12 WILSON, Stuart 934 2 8:L 15:W 16:W 5:L 10:
13 JENKINS, Deborah 1400 1.5 17:W 7:L 2:L 14:D 16:
14 LYONS, Russell 1.5 10:L 11:W 6:L 13:D 19:
15 CURRAN, John 1082 1 2:L 12:L 5:L 19:W 17:
16 MARKSON, Ellis 1047 1 0:D 5:D 12:L 11:L 13:
17 TAYLOR, Mark 984 1 13:L 19:W 10:L 9:L 15:
18 HILLER, James 594 0 9:L 4:L 0: 0: 0:
19 JOHNSON, Tim 0 6:L 17:L 11:L 15:L 14:

Bill Gletsos
24-04-2006, 12:30 AM
SP is clearly wrong.


5. Jenkins 1.5 WBWB thus a Black colur pref and an up float in round 3
19. Lyons 1.5 BWBW thus a White colour pref

10. Curran 1 BWBB thus a White colour pref
11. Markson 1 -WBW thus a black colour pref
15. Taylor 1 BWBW thus a black colour pref

18. Johnson 0 WBBW thus a black colour prefThe Round 5 pairing occurs as follows:

Note Jenkins has played Lyons and taylor, Johnson has played Curran and Taylor.

Since Jenkins and Lyons has played they drop down to the next score group.

as per C2 and C3:
w=2, b=3, q=3, thus x=0 and p = 2

Following C4 we get:

S1 S2
Jenkins Curran
Lyons Markson
TaylorThus carrying out C7 we get

Jenkins V Markson
Curran V Lyons

and restart at C2 with Taylor.
w=0, b=1, q=1, thus x = 0 and p=1

However Taylor is alone in the homonenous score group of 1.
Therefore as per C6 move Taylor down to the 0 score group and restart at C1.
Unfortunately Taylor has played Johnson therefore as per C1, apply C13 since this is the lowest score group.

As per C13 unpair the previous score group.
This means Jenkins V Markson and Curran V Lyons are unpaired.

Also as per C13 try and find another pairing what was the penultimate score bracket (players on 1.5 and 1) which will allow a pairing in the lowest score bracket.

Now note Curran cannot drop to play Johnson but Markson can.
Therefore attempt to pair the penultimate score group and continue from C2.

w=2, b=3, q=3, thus x=0 and p=2

following C4 we get

S1 S2
Jenkins Curran
Lyons Markson
TaylorNow we know from C13 that Markson is the player who it is planned will drop from this score group.

No matter how one applies C6, C7 and C9 (C8 and C10 are not applicable as the group is hetrogenous) no pairings can be found that allows for p=2 and x=0.
This is because Jenkins cannot play Lyons or Taylor as they have already played.
Apply C11.
Thus p=2 and x=1.
Again no matter how one applies C6, C7 and C9 no pairings allow for p=2 and x=1.
Again apply C11.
Thus p=2 and x=2.

Hence the following pairings meet the criteria of p=2 and x=2:

Curran V Jenkins (as per A7a Curran must be White as he was Black in the the previous two rounds)
Taylor V Lyons (as per E4 since they have identical colour histories and Lyons is the higher ranked)

Now pairing Markson and Johnson we get
w=0, b=2, q=1, thus x=1 and p=1.

So the pairing is Johnson V Markson (as per E2 as MArkson's colour prefererence is strong whilst Johnson's is mild)

Garvinator
24-04-2006, 01:25 AM
Geez I would find all those calculations difficult to do in the manner you have Bill :eek:

Anyways, so out of all your calculations, you have agreed with sm5 and disagreed with sp.

pax
24-04-2006, 08:50 AM
surely the algorithm needed to write a program for a tournament draw is significantly simpler than that needed actually to play the game of chess - the latter have been proved over the board.

It's a different proposition though. A chess playing program can play the best move only 90% of the time (or less), and still beat Grandmasters, whereas a pairing program needs to find the right pairing 100% of the time to be deemed "correct".

auriga
24-04-2006, 11:20 AM
At the end of all this a report will go to the ACF.

could we publish the report on the internet somewhere??
would be interesting to read about the comparison of
swiss manager, protos, swiss perfect, etc.

Ian Rout
24-04-2006, 11:33 AM
I have seen the effect that Garvin describes a number of times over the years. I think one may have been the 2004 NSWCA May Weekender.

The analogy with chess-playing programs is interesting. The computer often comes up with counter-intuitive moves that a human wouldn't, but which on inspection are found to be good. Occasionally though it is just plain wrong, like the famous 3...Ke7. If you could always play a program's move you would (for most of us) do better than always playing your own; but combining the two would be better still. Similarly if you work on using the computer draw every time, especially if your knowledge of the pairing rules is limited, you won't go too far wrong, but if you keep an eye out for errors you will do better - especially when the big one hits.


When will the chess community believe that the author of SP98 has written a program that acurately computes the draw correctly?
My first job in the full-time work force was in a section that was severely handicapped by having to work with a programmer whose programs never contained mistakes. (In fairness to the author of SP I don't think he himself makes such extravagant claims.)

arosar
24-04-2006, 12:08 PM
We want to run a tournament using SP. Can we happily rely on it by simply announcing, "all pairings done by SP are final and will not be changed"? We are not yet experts on the finer points of Swiss system.

AR

auriga
24-04-2006, 01:31 PM
ar, in my view there are 2 types of events: fide-rated and non fide-rated.

for fide-rated we need to announce the chief arbiter and s/w used in advance. we also need to use a fide-approved program and not overide the s/w.

for non fide-rated events, swiss perfect, IMO, is 'good enough' despite the bugs and can be relied upon to be used (without need to overide the s/w).
of course, if the organisers announce that they will overide then that's ok and is up to them.

pax
24-04-2006, 01:35 PM
for fide-rated we need to announce the chief arbiter and s/w used in advance. we also need to use a fide-approved program and not overide the s/w.


I don't believe there is any compulsion to use a FIDE approved program for FIDE rated events.

auriga
24-04-2006, 01:41 PM
I don't believe there is any compulsion to use a FIDE approved program for FIDE rated events.

guess what i meant is that 'best practises' is to use a FIDE-approved program.
if it used in most important events worldwide (eg. aeroflot open, dubai open)
then we should at least maintain this at our big tournaments (doeberl, aust ch). doesn't this make sense.

Bill Gletsos
24-04-2006, 02:08 PM
I believe Swiss Perfect was used for the FIDE rated 2006 Reykjavik Open.

Kevin Bonham
24-04-2006, 02:24 PM
It's useful that Bill's calculations prove SP is incorrect under section C as well in Garvin's example. Garvin's example is also another B3 double-downfloat bug issue - it pairs a 1.5 with a 0 (there is no 0.5 scoregroup) and a 1 with a 1 instead of a 1.5 with a 1 and a 1 with a 0. It would be interesting to know if other double-downfloat bug cases also proved to be incorrect under section C on more careful examination.

auriga
24-04-2006, 03:49 PM
It's useful that Bill's calculations prove SP is incorrect under section C as well in Garvin's example. Garvin's example is also another B3 double-downfloat bug issue - it pairs a 1.5 with a 0 (there is no 0.5 scoregroup) and a 1 with a 1 instead of a 1.5 with a 1 and a 1 with a 0. It would be interesting to know if other double-downfloat bug cases also proved to be incorrect under section C on more careful examination.

yes, i agree.

Garvinator
24-04-2006, 04:44 PM
It's useful that Bill's calculations prove SP is incorrect under section C as well in Garvin's example. Garvin's example is also another B3 double-downfloat bug issue - it pairs a 1.5 with a 0 (there is no 0.5 scoregroup) and a 1 with a 1 instead of a 1.5 with a 1 and a 1 with a 0. It would be interesting to know if other double-downfloat bug cases also proved to be incorrect under section C on more careful examination.
Would you or Bill like me to try and find more examples of pairings that have disagreed between sp/my judgement/sm5?

We do have round nine and maybe round eleven from 2006 Australian Championships as well.

pax
24-04-2006, 06:40 PM
SP is a *very* widely used program. It is used in many many big tournaments featuring lots of GMs.

I think people have this perception that SP is 'bad' because arbiters in Australia often change it's pairings. The fact is that often when the pairings are changed, the original SP pairing was 'correct' by FIDE standards.

Kevin Bonham
24-04-2006, 06:57 PM
Would you or Bill like me to try and find more examples of pairings that have disagreed between sp/my judgement/sm5?

We do have round nine and maybe round eleven from 2006 Australian Championships as well.

I've seen enough cases of the double-downfloat bug by now. If someone can find one they reckon is wrong but that has the pairings right in terms of which scoregroups are matched with which scoregroups then I would be interested to see that.

Bob1
24-04-2006, 09:35 PM
SP is clearly wrong.



Bill (& GG) thanks for the example and analysis.

However you must agress that ist is difficult to write a program that allows for the (random) removal of a player after rounds 1 & 2 .

I think that people withdrawing or entering after the start of an event are far more disruptive to the overall event than these minor calculation errors. ( and I hope they are minor because I didn't actually read the detail)

anyway keep up the investigation.

Bill Gletsos
24-04-2006, 11:50 PM
Bill (& GG) thanks for the example and analysis.

However you must agress that ist is difficult to write a program that allows for the (random) removal of a player after rounds 1 & 2 .It should not be difficult at all to handle such situations.

I think that people withdrawing or entering after the start of an event are far more disruptive to the overall event than these minor calculation errors. ( and I hope they are minor because I didn't actually read the detail)Some of these errors are anything but minor as they involve people playing others two score groups below them.

Garvinator
25-04-2006, 07:21 AM
I've seen enough cases of the double-downfloat bug by now. If someone can find one they reckon is wrong but that has the pairings right in terms of which scoregroups are matched with which scoregroups then I would be interested to see that.
see what i can do from previous tournaments or in the future. I was also thinking of where sp and sm5 disagree too, maybe from situations different to the ones we have shown.

Garvinator
25-04-2006, 05:43 PM
Tournament was played with sp pairings for round 5. I was not the arbiter;)

Ok same tournament as before, now for round 6 and there is a difference sp to sm5.

Most of the games have been played except two, i have assumed results for this situation where sp and sm5 differ. Thought it still relevant.

My opinion- I think sm5 is correct.

Results round 5:


No Name Loc Total Result Name Loc Total

1 HOWES, Tony (4) 1569 [4] 0:1 THOMAS, Brian (1) 1743 [3.5]
2 KIRKMAN, Phillip (2) 1724 [3] .5:.5 LYONS, Kieran C (8) 1276 [2.5]
3 CHADWICK, Marty (13) 1023 [2.5] 1:0 DE VERE, Cameron (12) 1044 [3]
4 WATERS, Mick (7) 1283 [2] 0:1 ROBINSON, Andrew J (3) 1587 [2.5]
5 MARTIN, Shane (14) 994 [2] 0:1 ADAMS, Ray (6) 1389 [2]
6 GRAY, Garvin (9) 1232 [2] 1:0 WILSON, Stuart (16) 934 [2]
7 JENKINS, Deborah (5) 1400 [1.5] 1:0 MARKSON, Ellis (11) 1047 [1]
8 JOHNSON, Tim (18) [0] 0:1 LYONS, Russell (19) [1.5]
9 CURRAN, John (10) 1082 [1] 0:1 TAYLOR, Mark (15) 984 [1]

Now sp pairings round 6:


No Name Loc Total Result Name Loc Total

1 THOMAS, Brian (1) 1743 [4.5] : KIRKMAN, Phillip (2) 1724 [3.5]
2 HOWES, Tony (4) 1569 [4] : ROBINSON, Andrew J (3) 1587 [3.5]
3 ADAMS, Ray (6) 1389 [3] : CHADWICK, Marty (13) 1023 [3.5]
4 LYONS, Kieran C (8) 1276 [3] : GRAY, Garvin (9) 1232 [3]
5 DE VERE, Cameron (12) 1044 [3] : JENKINS, Deborah (5) 1400 [2.5]
6 WILSON, Stuart (16) 934 [2] : LYONS, Russell (19) [2.5]
7 TAYLOR, Mark (15) 984 [2] : MARTIN, Shane (14) 994 [2]
8 CURRAN, John (10) 1082 [1] : WATERS, Mick (7) 1283 [2]
9 MARKSON, Ellis (11) 1047 [1] : JOHNSON, Tim (18) [0]

sm5 pairings:


No Name Loc Total Result Name Loc Total

1 THOMAS, Brian (1) 1743 [4.5] : KIRKMAN, Phillip (2) 1724 [3.5]
2 HOWES, Tony (4) 1569 [4] : ROBINSON, Andrew J (3) 1587 [3.5]
3 ADAMS, Ray (6) 1389 [3] : CHADWICK, Marty (13) 1023 [3.5]
4 LYONS, Kieran C (8) 1276 [3] : GRAY, Garvin (9) 1232 [3]
5 DE VERE, Cameron (12) 1044 [3] : JENKINS, Deborah (5) 1400 [2.5]
6 LYONS, Russell (19) [2.5] : TAYLOR, Mark (15) 984 [2]
7 WILSON, Stuart (16) 934 [2] : WATERS, Mick (7) 1283 [2]
8 CURRAN, John (10) 1082 [1] : MARTIN, Shane (14) 994 [2]
9 MARKSON, Ellis (11) 1047 [1] : JOHNSON, Tim (18) [0]

Cross Table through round 5:


No Name Loc Total 1 2 3 4 5

1 THOMAS, Brian 1743 4.5 16:W 0:D 10:W 6:W 2:W
2 HOWES, Tony 1569 4 13:W 7:W 5:W 3:W 1:L
3 KIRKMAN, Phillip 1724 3.5 9:W 12:W 6:W 2:L 7:D
4 ROBINSON, Andrew J 1587 3.5 5:L 17:D 16:W 15:W 12:W
5 CHADWICK, Marty 1023 3.5 4:W 10:W 2:L 7:D 9:W
6 ADAMS, Ray 1389 3 15:W 8:W 3:L 1:L 13:W
7 LYONS, Kieran C 1276 3 19:W 2:L 11:W 5:D 3:D
8 GRAY, Garvin 1232 3 11:W 6:L 14:W 9:L 15:W
9 DE VERE, Cameron 1044 3 3:L 18:W 12:W 8:W 5:L
10 JENKINS, Deborah 1400 2.5 14:W 5:L 1:L 11:D 17:W
11 LYONS, Russell 2.5 8:L 13:W 7:L 10:D 19:W
12 WATERS, Mick 1283 2 18:W 3:L 9:L 14:W 4:L
13 MARTIN, Shane 994 2 2:L 11:L 19:W 17:W 6:L
14 TAYLOR, Mark 984 2 10:L 19:W 8:L 12:L 16:W
15 WILSON, Stuart 934 2 6:L 16:W 17:W 4:L 8:L
16 CURRAN, John 1082 1 1:L 15:L 4:L 19:W 14:L
17 MARKSON, Ellis 1047 1 0:D 4:D 15:L 13:L 10:L
18 HILLER, James 594 0 12:L 9:L 0: 0: 0:
19 JOHNSON, Tim 0 7:L 14:L 13:L 16:L 11:L

arosar
01-05-2006, 10:45 AM
Last Saturday, there was an interesting little column in The Age. You can read a portion of it in the blog.

Thanks to our Victorian insider for the scoop!

AR

pax
01-05-2006, 12:26 PM
Thanks to our Victorian insider for the scoop!


That's a pretty broad definition of "insider" Amiel :D

PHAT
01-05-2006, 02:29 PM
That's a pretty broad definition of "insider" Amiel :D

Enough about the "insider." Now comment on the scoop. Is CZ a pushover?

Bill Gletsos
01-05-2006, 02:33 PM
Tournament was played with sp pairings for round 5. I was not the arbiter;)

Ok same tournament as before, now for round 6 and there is a difference sp to sm5.

Most of the games have been played except two, i have assumed results for this situation where sp and sm5 differ. Thought it still relevant.

My opinion- I think sm5 is correct.Your post lacks relevant information.
You need to post the round 5 pairing info screen from SP as it is impossible to determine colour allocations in previous rounds from the cross table.

PHAT
01-05-2006, 02:46 PM
Your post lacks relevant information.
You need to post the round 5 pairing info screen from SP as it is impossible to determine colour allocations in previous rounds from the cross table.
That was an unnecessary piece of knit picking. You have already done the good job of showing SP to be correct in this instance. Now shut up and let the humans decide the human course of action.

jenni
01-05-2006, 02:55 PM
Enough about the "insider." Now comment on the scoop. Is CZ a pushover?

Charles could be viewed as a pushover as he is a very kind person who likes to keep everyone happy. However Shaun (as I am sure you will agree), is not and he agreed to the change. (done under a huge amount of time pressure).

Bill Gletsos
01-05-2006, 03:00 PM
That was an unnecessary piece of knit picking. You have already done the good job of showing SP to be correct in this instance. Now shut up and let the humans decide the human course of action.Firstly garvin was asking a question as to which pairings were correct in his example. that cannot be determined without the additional information I requested.
As such your claim that I am nitpicking just demonstrates how clueless you are when it comes to swiss pairings.
Given you are clueless when it comes to the pairing rules and arbitering then if anyone should shut up it is you.

Igor_Goldenberg
01-05-2006, 03:42 PM
Last Saturday, there was an interesting little column in The Age. You can read a portion of it in the blog.

Thanks to our Victorian insider for the scoop!

AR

Being a journalist Chris wants to notch things up a bit, as it makes a good story. During the tournament I do not remember anyone who'd think that the amended (and actually played) pairing is incorrect.

Whilst technically the generated pairing strictly follows the rule, the revised pairing made much more sense logically.

If I was paired against Smerdon or Zhao instead of Gareth Charles, there would probably be no complain, as leaders would have to play the strongest opponents.

Let's look at the situation from Rogers's point of view. If he does not play with white against Johansen in round 6, he is very likely to play against him in round seven with black. The original pairing was more favourable to him the the amended!

Therefore I do not agree that Rogers protested the pairing for the personal gain. Given that the protest from him carries more weight, it was prudent for Ian to make a complaint.

PHAT
01-05-2006, 03:54 PM
Given you are clueless when it comes to the pairing rules ...

Show where I have debated the pairing rules - I have not. Therefore you cannot make that claim. So, STFU.


... and arbitering then if anyone should shut up it is you.

Show where I have stuffed up an arbiting job - or are you an "evidence free zone," hypercrit.

smurf
01-05-2006, 04:21 PM
Not that it makes much difference, but... In the Doeberl instance, the changed pairings meant that I had to play Johansen instead of Hoffman. Obviously this went against me, but I could see the logic of the original protest and the new pairings. I did grumble to the top boards that I was unhappy about the new pairings, but made no official protest, as my reason was simply that I had to play a harder player.

In Brisbane, again the new pairings in the last round made sense, although again I came out with the harder opponent.

Such is life!

David.

Bill Gletsos
01-05-2006, 04:22 PM
Show where I have debated the pairing rules - I have not. Therefore you cannot make that claim. So, STFU.You have argued that the pairings generated by Swiss Perfect should be followed without any manual intervention. Only someone ignorant of the pairing rules could take this stance as it has been shown that there are times when SP gets the pairings wrong. In those circumstances it should be overridden to follow the pairing rules.

Show where I have stuffed up an arbiting job - or are you an "evidence free zone," hypercrit.I never said yiou had stuffed up, I said you were clueless when it came to arbitering. After all you have demonstrated your lack of knowledge of the laws of chess a number of times in various threads.

pax
01-05-2006, 04:39 PM
You have argued that the pairings generated by Swiss Perfect should be followed without any manual intervention. Only someone ignorant of the pairing rules could take this stance as it has been shown that there are times when SP gets the pairings wrong. In those circumstances it should be overridden to follow the pairing rules.
(emphasis mine)

That is not a fair judgment. In weekend events such as the Doeberl, there is seldom enough time to properly verify the draw from first principles (unless it it the first round in a day).

I would argue that unless the arbiter has intends (and has time) to properly verify the draw for every round of the tournament, a stance of not overriding the draw is perfectly reasonable.

Rincewind
01-05-2006, 04:46 PM
I would argue that unless the arbiter has intends (and has time) to properly verify the draw for every round of the tournament, a stance of not overriding the draw is perfectly reasonable.

I think there are two factors at play. You consider on of them, that is the expertise and time to scrutinise the draw and this is definitely a factor in adopting such a policy. However, what must be factored in is the confidence in the pairing software to produce a valid set of pairings without intervention. There must be some confidence or else why use it at all? However, the current process seems to be

- produce computer pairing,
- kill goat,
- investigate entrails
- adjust draw accordingly
- publish draw

Unless the confidence level in the software is increased will be difficult to do away with steps 2, 3 and 4.

arosar
01-05-2006, 04:52 PM
So to the top players are saying that the revised pairings were "logical" from, it seems to me, a sporting contest point of view. Looking at it that way, is not their position reasonable?

AR

Bill Gletsos
01-05-2006, 04:54 PM
(emphasis mine)

That is not a fair judgment.I suspected someone might take issue with it. ;)

In weekend events such as the Doeberl, there is seldom enough time to properly verify the draw from first principles (unless it it the first round in a day).Arbiters have to make the call to override or not based on the circumstances.
To dogmatically state that you would follow the computer pairings come what may is illogical. As Ian Rout said in another thread:

And for that reason those who want to mindlessly follow computer instructions are just, well, mindless.

I would argue that unless the arbiter has intends (and has time) to properly verify the draw for every round of the tournament, a stance of not overriding the draw is perfectly reasonable.I disagree as the double float error of Swiss Perfect is detectable and if detected should be rectified.

PHAT
01-05-2006, 05:41 PM
You have argued that the pairings generated by Swiss Perfect should be followed without any manual intervention.

Correct. I did say that and I stand be it. However, you said I have no understanding of pairing rules. You cannot say that because I have never argued pairing rules. Thus you statement that I do do not understand them is an evidence free slur. So V.. .


Only someone ignorant of the pairing rules could take this stance as it has been shown that there are times when SP gets the pairings wrong. Why "Only someone ignorant ..."? What matters most is not that SP pairs slightly incorrectly on the odd occations, but that DOPs fiddle with the SP pairings and 50-50 get it wrong. That means players get upset with perceived bias (or worse.) Players cannot allege bias in a laptop.

Therefore, if we have a chioce, it is better to follow a program that gets it wrong sometimes than to follow a DOP who gets it wrong sometimes.

If you cannot see that you should not hold official position that requires you to blue tooth interface with a human being.

Garvinator
01-05-2006, 05:54 PM
Your post lacks relevant information.
You need to post the round 5 pairing info screen from SP as it is impossible to determine colour allocations in previous rounds from the cross table.
Hello Bill,

I believe this is what you require.


No Name Loc Total Result Name Loc Total

1 HOWES, Tony (4) 1569 [4] 0:1 THOMAS, Brian (1) 1743 [3.5]
2 KIRKMAN, Phillip (2) 1724 [3] .5:.5 LYONS, Kieran C (8) 1276 [2.5]
3 CHADWICK, Marty (13) 1023 [2.5] 1:0 DE VERE, Cameron (12) 1044 [3]
4 WATERS, Mick (7) 1283 [2] 0:1 ROBINSON, Andrew J (3) 1587 [2.5]
5 MARTIN, Shane (14) 994 [2] 0:1 ADAMS, Ray (6) 1389 [2]
6 GRAY, Garvin (9) 1232 [2] 1:0 WILSON, Stuart (16) 934 [2]
7 JENKINS, Deborah (5) 1400 [1.5] 1:0 MARKSON, Ellis (11) 1047 [1]
8 JOHNSON, Tim (18) [0] 0:1 LYONS, Russell (19) [1.5]
9 CURRAN, John (10) 1082 [1] 0:1 TAYLOR, Mark (15) 984 [1]

Bill Gletsos
01-05-2006, 06:06 PM
Hello Bill,

I believe this is what you require.No, what is required is the pairing info for round 5. You obtain it from the "Pairing Info" option under View in SP.

It should look something like:

Place No Opponents Colours Float Score


1-2 2 : 17,11,7,1,10,9,4,3,8,5 BWBWBWBWBW D 8
9 : 24,1,3,5,4,2,7,13,10,8 WWBWBBWWBB D 8

3 4 : 19,15,29,7,9,1,2,6,3,12 BWBWWBWBBW D 7.5

4 3 : 18,10,9,25,6,7,13,2,4,1 WBWBWWBBWW U 6.5

5-7 1 : 16,9,6,2,8,4,5,11,7,3 WBWBWWBBWB D 6
7 : 22,27,2,4,11,3,9,12,1,13 WBWBWBBWBW uD 6
28 : 13,5,12,24,15,21,27,8,26,23 BWBWBWBWBW d 6

8-12 8 : 23,13,17,29,1,12,11,28,2,9 BWBWBBWBWW U 5.5
10 : 25,3,5,23,2,13,18,16,9,14 BWBWWBWBWB u 5.5
13 : 28,8,18,19,27,10,3,9,14,7 WBWBBWWBWB U 5.5
14 : 29,22,21,11,5,23,24,15,13,10 BWBWBWBWBW 5.5
18 : 3,25,13,15,21,27,10,5,23,6 BWBWBWBWBW D 5.5

13-19 5 : 20,28,10,9,14,11,1,18,6,2 WBWBWBWBWB Ud 5
6 : 21,23,1,27,3,29,12,4,5,18 BWBWBWBWBB U 5
11 : 26,2,19,14,7,5,8,1,12,20 WBWBBWBWBW 5
12 : 27,21,28,17,23,8,6,7,11,4 BWWBBWWBWB U 5
15 : 30,4,27,18,28,20,16,14,22,25 WBWBWBWBWB 5
23 : 8,6,16,10,12,14,20,29,18,28 WBWBWBWBWB 5
25 : 10,18,20,3,19,22,29,24,16,15 WBWWBBWBWW 5

20-23 16 : 1,24,23,21,29,19,15,10,25,30 BWBWBWBWBW D 4.5
22 : 7,14,26,30,20,25,19,27,15,21 BBWBWWBWBW 4.5
24 : 9,16,30,28,17,26,14,25,-,29 BBWBWBWW-B Ud 4.5
27 : 12,7,15,6,13,18,28,22,20,- WWBBWBWBW- uD 4.5

24-26 20 : 5,29,25,26,22,15,23,30,27,11 BWBWBWBWBB d 4
26 : 11,17,22,20,30,24,-,21,28,19 BWBBWW-BWB uD 4
29 : 14,20,4,8,16,6,25,23,30,24 WBWBWBBWBW D 4

27-28 19 : 4,30,11,13,25,16,22,-,21,26 WBBWWBW-BW Ud 3.5
21 : 6,12,14,16,18,28,30,26,19,22 WBWBWBBWWB u 3.5

29 30 : 15,19,24,22,26,-,21,20,29,16 BWBWB-WBWB U 1.5

but only show the first 5 rounds.

pax
01-05-2006, 06:15 PM
To dogmatically state that you would follow the computer pairings come what may is illogical.

Let me put it like this: suppose I am directing a tournament with 7 rounds on two days. I expect to have no more than 15 minutes for most rounds to generate the draw, print it out, and get all the players seated ready to start the next round.

I do not believe that 15 minutes is long enough to correctly detect a pairing error, and to manually generate a correct pairing with > 90% confidence. It is also not long enough to persuade a player who objects to the computer pairing that it is, in fact, correct. Therefore I will announce to the players prior to the start of the event that the computer pairing is final.

The double downfloat error is relatively minor compared with the tournament running more than an hour behind schedule.

Bill Gletsos
01-05-2006, 06:16 PM
Correct. I did say that and I stand be it. However, you said I have no understanding of pairing rules. You cannot say that because I have never argued pairing rules. Thus you statement that I do do not understand them is an evidence free slur. So V.. .Of course I can claim you have no understanding of the pairing rules becuase you showed your lack of understanding of the pairing rules when you queried in post #104 why I asked Garvin for the pairing information details. Anyone who understood the pairing rules would know why I requested that information.

Why "Only someone ignorant ..."? What matters most is not that SP pairs slightly incorrectly on the odd occations, but that DOPs fiddle with the SP pairings and 50-50 get it wrong. That means players get upset with perceived bias (or worse.) Players cannot allege bias in a laptop.Player's can however claim an error in the laptop and expect action from the arbiter in cases where the computer is wrong.

Therefore, if we have a chioce, it is better to follow a program that gets it wrong sometimes than to follow a DOP who gets it wrong sometimes.No. what is required is to educate the arbiters to ensure that if they pair manually they pair correctly. The FIDE rules make it clear the draw is the responsibility of the arbiter, not a computer.

If you cannot see that you should not hold official position that requires you to blue tooth interface with a human being.the opinion of someone as clueless as you does not concern me.

Bill Gletsos
01-05-2006, 06:20 PM
Let me put it like this: suppose I am directing a tournament with 7 rounds on two days. I expect to have no more than 15 minutes for most rounds to generate the draw, print it out, and get all the players seated ready to start the next round.

I do not believe that 15 minutes is long enough to correctly detect a pairing error, and to manually generate a correct pairing with > 90% confidence. It is also not long enough to persuade a player who objects to the computer pairing that it is, in fact, correct. Therefore I will announce to the players prior to the start of the event that the computer pairing is final.The arbiter should act based on the circumstances. He may well be able to determine there is an error and correct it in the time available. If he can he should.
As such I stand by my comment that following the computer pairings come what may is illogical.

The double downfloat error is relatively minor compared with the tournament running more than an hour behind schedule.I doubt players would feel that way if it impacted them winning a prize.

Garvinator
01-05-2006, 06:27 PM
No, what is required is the pairing info for round 5. You obtain it from the "Pairing Info" option under View in SP.
ahhh yes.

Showing only the first five rounds:


Place No Opponents Colours Float Score


1 1 : 10,-,5,6,4 W-BWB Ud 4.5

2 4 : 14,8,13,2,1 BWBBW D 4

3-5 2 : 12,7,6,4,8 BWBWW D 3.5
3 : 13,11,10,16,7 WBWWB uD 3.5
13 : 3,5,4,8,12 BWWBW U 3.5

6-9 6 : 16,9,2,1,14 BWWBB u 3
8 : 18,4,19,13,2 BBWWB U 3
9 : 19,6,15,12,16 WBWBW 3
12 : 2,17,7,9,13 WBBWB D 3

10-11 5 : 15,13,1,19,11 WBWBW D 2.5
19 : 9,14,8,5,18 BWBWB D 2.5

12-15 7 : 17,2,12,15,3 WBWBW U 2
14 : 4,19,18,11,6 WBWBW 2
15 : 5,18,9,7,10 BWBWB 2
16 : 6,10,11,3,9 WBWBB d 2

16-17 10 : 1,16,3,18,15 BWBBW 1
11 : -,3,16,14,5 -WBWB U 1

18-19 18 : 8,15,14,10,19 WBBWW U 0

PHAT
01-05-2006, 07:14 PM
Of course I can claim you have no understanding of the pairing rules becuase you showed your lack of understanding of the pairing rules when you queried in post #104 why I asked Garvin for the pairing information details.

I just lobbed in here for a look and did not see that GG was querying another event's pairing. I inadvertantely thought you were knit picking. I see now that you were not. Therefore, you still do not have any evidence that I do not understand pairing. Your statement remains evidence free.


Player's can however claim an error in the laptop and expect action from the arbiter in cases where the computer is wrong.

Does that mean that they cannot claim an error when the computer is right. Be careful in your answer.



What is required is to educate the arbiters to ensure that if they pair manually they pair correctly.

Exactly what has the ACF done to insure this happens? In the absence of such a program (including an exam) the ACF should rule that in Australia SP over ride DOPs.

the opinion of someone as clueless as you does not concern me.

Then stop replying to me DOG!

Garvinator
01-05-2006, 07:24 PM
the ACF should rule that in Australia SP over ride DOPs.
nooooo, the acf should be looking into different pairing programs, which analysis like this is doing by comparing the different pairing programs.

Maybe if after EXTENSIVE testing, another pairing program is shown to be superior and especially more trustworthy, a deal can be arranged with the makers of the pairing program and the acf.

By the way, nothing forces an arbiter or tournament to use swissperfect for pairing purposes, only that the pairings are converted to sp files to send in for ratings.

PHAT
01-05-2006, 10:12 PM
nooooo, the acf should be looking into different pairing programs, which analysis like this is doing by comparing the different pairing programs.


Agreed. But between now and that happening, he ACF should simply rule SP as the first and only pairing.

Bill Gletsos
02-05-2006, 12:30 AM
I just lobbed in here for a look and did not see that GG was querying another event's pairing. I inadvertantely thought you were knit picking. I see now that you were not. Therefore, you still do not have any evidence that I do not understand pairing. Your statement remains evidence free.What it means is that you didnt bother reading the posts in context and therefore without understanding what was being discussed did your usual and just sprouted rubbish.

Does that mean that they cannot claim an error when the computer is right. Be careful in your answer.Players can always query pairings.

Exactly what has the ACF done to insure this happens? In the absence of such a program (including an exam) the ACF should rule that in Australia SP over ride DOPs.Well I could understand why you would want to blindly follow the SP pairings.

Then stop replying to me DOG!Then stop making false statements.

Rhubarb
02-05-2006, 05:15 AM
I just lobbed in here for a look You know, it just occurred to me: after all that has happened; after all of Matty's ill-informed abuse; after his totally useless time on the NSWCA Council; after his idiotic bluster to try and blame his own faults on the people who removed him according to the NSWCA Constitution; after his subsequent clueless attempts to 'advise' the NSWCA; after his ridiculously incompetent attempts to weigh into ratings debates (OMG what an abject tard - how could he possibly delude himself into believing he had anything useful to say here); after getting himself banned from this place for continued, defamatory, unwarranted, unprovoked abuse; after doing the same and worse on his own ill-fated board; after getting himself banned from the NSWCA for just that; after then going into full-on martyr mode (what me? what did I do?); and after now showing up in the Arbiters Corner with his usual total ignorance and then abusing Bill after admitting that he hadn't even read the start of the thread. After all that, I cannot be swayed: I still think Matt's a moron.

PHAT
02-05-2006, 07:50 AM
You know, it just occurred to me: after all that has happened; ... I still think Matt's a moron.

:lol: priceless satire.

Rhubarb
02-05-2006, 10:09 AM
Given the voluntary nature of chess management in Australia (club, state and national level), anything of the magnitude can only be achived by individual efforts. As long as ACF and state maintain basic infrastructure and don't stay in the way of individual/club initiative (which, unfortunately, they do time to time), they'd bo doing the right thing.

If you look at the world-wide picture, when FIDE and most national federation try to do something outside of maintaining the basics, they stuff thighs up. Most of the spectacular events are organised by someone else (with FIDE mostly stiffing their nose and messing things up).

Matty, why don't you sic yourself upon the Goldenberg. Like you, he is utterly clueless when it comes to Swiss pairings and, remarkably, just like you, seems to claim to know how to solve all of Australia's chess ills at one sitting.

Igor_Goldenberg
02-05-2006, 12:37 PM
Matty, why don't you sic yourself upon the Goldenberg. Like you, he is utterly clueless when it comes to Swiss pairings and, remarkably, just like you, seems to claim to know how to solve all of Australia's chess ills at one sitting.

What about posting something of the substance?

PHAT
02-05-2006, 03:09 PM
What about posting something of the substance?

Perhaps you and I could post something of substance to Greg and ensure the Feds know about it. ;) OTOH we could ask him why he is like a cracked record about his always being paired as black in the last round.

PHAT
02-05-2006, 03:21 PM
Matty, why don't you sic yourself upon the Goldenberg. Like you, he is utterly clueless when it comes to Swiss pairings ...

Why would I want to savage a person who is as unlike Gletsos as it is possible to be? It would be a treachery against my species.


... and, remarkably, just like you, seems to claim to know how to solve all of Australia's chess ills at one sitting.

I make no such claim. But I will claim that a risky plan is better than no plan.

In any case, I would rather have a dymamic faulty brain than a mindless silicon chip.

Rhubarb
02-05-2006, 04:29 PM
What about posting something of the substance?

Well I'm buggered if I know what Igor meant by the above, but I'd sure as hell like to know why you

agreed to the faulty redraw.

Garvinator
02-05-2006, 04:37 PM
Well I'm buggered if I know what Igor meant by the above, but I'd sure as hell like to know why you

agreed to the faulty redraw.(back to normal size for my reply to save space)

Greg, which faulty draw;)

Igor_Goldenberg
02-05-2006, 04:42 PM
Well I'm buggered if I know what Igor meant by the above, but I'd sure as hell like to know why you

agreed to the faulty redraw.

I expected to play Rogers and I wanted to play him.

You know, some people are more intrested in the game then in the chasing result/prize and want to play a grandmaster when opportunity arises, not to avoid it.

jase
02-05-2006, 05:14 PM
One can see why Igor has returned to chess so strongly in the last 6 months.
Perhaps his humility in his Olympiad application won't be warranted next time around.

Rhubarb
02-05-2006, 05:40 PM
I expected to play Rogers and I wanted to play him.

You know, some people are more intrested in the game then in the chasing result/prize and want to play a grandmaster when opportunity arises, not to avoid it.
Oh sorry, I was just interested in the correct draw. Don't mind me.

jase
02-05-2006, 06:00 PM
I was just interested in the correct draw
What is correct? Are the Dutch pairing rules correct? I think they're rather inferior to the Lim system. The Dutch rules are in vogue because
a) Guert Gijssen, who designed the Dutch system, is the planet's most influential arbiter; and
b) Software such as SwissPerfect allows computer-generated pairings that obey the Dutch system, and file transfers for rating purposes.

Your point, Greg, rough-hew it how you will, is valid. However Igor did offer a splendid explanation in response.

Rincewind
02-05-2006, 06:11 PM
What is correct? Are the Dutch pairing rules correct?

Whatever pairing system the aribiter's decided to use throughout the event would be the correct one. If that is the Dutch system, then yes, those rules are the correct ones in this instance.

PHAT
02-05-2006, 06:59 PM
Whatever pairing system the aribiter's decided to use throughout the event would be the correct one.

So, if CZ used some newly invented "Canfeell-Cox" system, then that would be the correct one.

Rincewind
02-05-2006, 07:31 PM
So, if CZ used some newly invented "Canfeell-Cox" system, then that would be the correct one.

Such a hypothetical is meaningless. Whatever pairing system was decided before the event should be used throughout. In this case it was the Dutch system, and unquestionably that is the correct system in Greg's context.

Rhubarb
02-05-2006, 07:40 PM
What is correct? Are the Dutch pairing rules correct? I think they're rather inferior to the Lim system. The Dutch rules are in vogue because
a) Guert Gijssen, who designed the Dutch system, is the planet's most influential arbiter; and
b) Software such as SwissPerfect allows computer-generated pairings that obey the Dutch system, and file transfers for rating purposes.

Your point, Greg, rough-hew it how you will, is valid. However Igor did offer a splendid explanation in response.Well, Jase, we'd all like to let you believe that you're some mystical superior being, but sometimes you're required to make a decision.

I understand that you let Charles Zworestine change the Round 11 Australian Championship pairings. Why don't you describe, in your own words, and please take your time, what exactly happened in those fateful hours.

Garvinator
02-05-2006, 08:11 PM
I understand that you let Charles Zworestine change the Round 11 Australian Championship pairings. Why don't you describe, in your own words, and please take your time, what exactly happened in those fateful hours.
I know that you have directed the question to Jason, but I think I should reply for at least part of your question Greg.

I do believe I have answered some of it in the round 11 Australian Championships thread:


correct and I was there.

After the round seven goings on, it was agreed by all four arbiters that they would sit down together for each round and do the pairings together after each round had finished. I was allowed to participate in the discussions as a learning experience (someone has to learn somehow) and there wasnt any complaints about this.

From my memory,
I had heard that there was some discussion between Charles and Cathy regarding the round eleven pairings before the following (below) took place.

Charles said that the pairings for round eleven were wrong and a discussion was held about the pairings and everyone went through the chess organisers handbook in turn, plus looking at the pairing rules directly.

We only had protos (which we learnt wasnt reliable or trustworthy) and swiss perfect to work from (no swiss master 5 or swiss manager for round eleven).

Charles put his case forward that the pairings were wrong (I dont remember what grounds exactly, it was four months ago). Jason disagreed with this and they debated it a bit.

Charles then presented what he believed were the correct pairings. More discussions were held and there was a vote taken. It was decided by a majority of 3-1 to go with the Charles pairings.

Something I have learnt from recent experience is that it is possible to do pairings by mathematical formula (thanks Bill). In Brisbane, there was certainly no discussion at any time of lets check the maths.

I have bolded the part that I believe is relevant to your question. Jason in no way 'allowed' Charles to change the round 11 pairings on his own. There was a long discussion about the round 11 pairings after round 10 was completed involving four IA's and myself.

After approximately one hours discussion, going through the pairing rules, more discussion and back and forth, there was a vote taken about how the rules should be applied and Charles's pairings were the ones voted for by a majority of 3-1 (all IA votes). Jason was the against vote.

I do think that Jason could give a better explanation of how the manual pairings were arranged, but the best person to ask is Charles Zworestine, which I have noticed does not seem to happen much through a few threads now.

jase
02-05-2006, 09:31 PM
Well, Jase, we'd all like to let you believe that you're some mystical superior being, but sometimes you're required to make a decision.

Your use of a comma denotes a link between my superiority and decision-making. Care to explain that link (in your own words)?


I understand that you let Charles Zworestine change the Round 11 Australian Championship pairings. Why don't you describe, in your own words, and please take your time, what exactly happened in those fateful hours.
Whilst Garvin has covered the basics of your enquiry, I'm happy to jog your memory and answer in my own words, if it pleases you to enquire in the appropriate thread ;) (given that we have discussed those fateful hours over a few beers a couple of times now; admittedly at the end of the first such evening you were yelling at a certain Grandmaster "Don't f**king touch me!" when he attempted to bid you farewell).

Igor_Goldenberg
02-05-2006, 11:07 PM
(given that we have discussed those fateful hours over a few beers a couple of times now; admittedly at the end of the first such evening you were yelling at a certain Grandmaster "Don't f**king touch me!" when he attempted to bid you farewell).

I assume it was 9th of January?:D

Lucena
02-05-2006, 11:17 PM
What is correct? Are the Dutch pairing rules correct?

In the specific context of the tournament, wouldn't the "Dutch" pairing rules be the correct rules to use as they were being used to do the earlier pairings via Swiss Perfect?


I think they're rather inferior to the Lim system.

I think the question of what system is "better" to use becomes irrelevant to a tournament once you are several rounds into that tournament having used the one pairing system. On the other hand, the more general question which you have raised, of whether there is a superior pairing system to the "Dutch" pairing system is definitely an interesting one. Could you give more detail (if this hasn't already been discussed elsewhere) about the pros and cons of the "Dutch" and Lim pairing systems? I have only a relatively rudimentary understanding of Swiss pairings, but am hoping to remedy this.

PHAT
02-05-2006, 11:28 PM
... given that we have discussed those fateful hours over a few beers a couple of times now; admittedly at the end of the first such evening you [GC] were yelling at ...

Isn't that a bannable offence ;)

jase
03-05-2006, 12:22 AM
I assume it was 9th of January?:D
Indeed it was Igor ;)
That said, there are few people's company I enjoy more over a lager or three than Greg. He is a wise and entertaining scholar.

Igor_Goldenberg
03-05-2006, 09:40 AM
Indeed it was Igor ;)
That said, there are few people's company I enjoy more over a lager or three than Greg. He is a wise and entertaining scholar.

It must've been much more then three:)

arosar
04-05-2006, 09:54 PM
Where's Ian Rogers? What does he have to say for himself? The public demands a statement immediately.

AR

Garvinator
05-05-2006, 11:47 AM
Where's Ian Rogers? What does he have to say for himself? The public demands a statement immediately.

AR
The public isnt 'demanding' anything of the sort. Of course it would be good if Ian would reply to some or all of the comments made in this and the Round 11 Aus Champ thread.

The only one demanding an answer is you because you need something to write about and Ian Rogers reply would give you something to write about.

PHAT
05-05-2006, 12:35 PM
The public isnt 'demanding' anything of the sort.

Name one person here who is very intersted in hearing IR's version. Given that interest, I think it is reasonable to say that the public demands dot dot dot.

arosar
05-05-2006, 02:07 PM
The public isnt 'demanding' anything of the sort. Of course it would be good if Ian would reply to some or all of the comments made in this and the Round 11 Aus Champ thread.

The only one demanding an answer is you because you need something to write about and Ian Rogers reply would give you something to write about.

This is why your PR skills lack a certain savvy gray. Of course the public demands it. Just as the public demands to know what Tom Cruise had for brekky this morning.

AR