PDA

View Full Version : PROTOS experiment



pax
25-01-2006, 03:24 PM
Ok, so I am undertaking a PROTOS experiment on the Australian Championships draws (just completed).

The first round of difference with the official draw is Rd 7.

Official draw:


1 2537 2 GM Chandler, Murray (5) 0 - 1 GM Rogers, Ian (4.5) 2556 2631
2 2340 2324 FM Xie, George (4) ˝ - ˝ GM Schmaltz, Roland (4.5) 2544 2530
3 2421 2405 IM Smerdon, David (4.5) 1 - 0 IM Lane, Gary (4) 2445 2484
4 2462 2453 GM Johansen, Darryl (3.5) ˝ - ˝ GM Ftacnik, Lubomir (4) 2612 2690
5 2354 2296 FM Canfell, Greg (3.5) ˝ - ˝ IM Zhao, Zong-Yuan (3.5) 2461 2475
6 2439 2538 IM Wohl, Alex (3.5) ˝ - ˝ FM Bjelobrk, Igor (3.5) 2399 2408
7 2415 2391 IM Solomon, Stephen (3) ˝ - ˝ Smirnov, Vladimir (3.5) 2235 2237
8 2254 2189 FM Humphrey, Jonathan (3) ˝ - ˝ FM Guthrie, Aaron (3) 2251 2120
9 2112 2093 WIM Moylan, Laura (2.5) 0 - 1 Song, Raymond (3) 2051 2073
10 2146 2105 Bird, Andrew (2.5) 0 - 1 FM Goldenberg, Igor (2.5) 2335 2315
11 2162 2148 Ly, Moulthun (2.5) 1 - 0 WIM Caoili, Arianne (2.5) 2207 2293
12 2126 2079 Pyke, Malcolm (2) ˝ - ˝ Obst, James (2.5) 2012 2054
13 2223 0 FM Dougherty, Michael (2) 0 - 1 Booth, Stewart (2) 2179 2184
14 0 1850 Frame, Nigel (1.5) 0 - 1 Lakner, Jay (1.5) 2207 2145
15 2114 2055 Wallis, Christopher (1.5) 1 Bye

PROTOS:



Pairings of round No. 7

1 4 Chandler ( 5 ) - 2 Rogers ( 4.5)
2 13 Xie ( 4 ) - 3 Schmaltz ( 4.5)
3 9 Smerdon ( 4.5) - 7 Lane ( 4 )
4 5 Johansen ( 3.5) - 1 Ftacnik ( 4 )
5 11 Bjelobrk ( 3.5) - 6 Zhao ( 3.5)
6 8 Wohl ( 3.5) - 18 Smirnov ( 3.5)
7 10 Solomon ( 3 ) - 12 Canfell ( 3.5)
8 15 Humphrey ( 3 ) - 16 Guthrie ( 3 )
9 29 Obst ( 2.5) - 28 Song ( 3 )
10 19 Dougherty ( 2 ) - 14 Goldenberg ( 2.5)
11 24 Bird ( 2.5) - 22 Booth ( 2 )
12 25 Pyke ( 2 ) - 23 Ly ( 2.5)
13 30 Frame ( 1.5) - 20 Caoili ( 2.5)
14 27 Moylan ( 2.5) - 26 Wallis ( 1.5)
15 21 Lakner ( 1.5) - 0 to receive the bye

So the first question is this:

Why does PROTOS float Canfell instead of Smirnov?

The second question is: What on *earth* is going on with boards 9 onwards?? There isn't a single homogeneous pairing among them!!

pax
25-01-2006, 03:27 PM
Boards 10 onwards have done exactly what was descibed earlier: combined the 1.5-2.5 groups into a single group, and paired top vs bottom. This cannot be correct..

Garvinator
25-01-2006, 03:31 PM
Hello Pax,

I wouldnt look at round seven too closely. It contained an error which wasnt picked up until round nine. In round seven, Song was downfloated two rounds in a row, round six and seven. Clearly this is incorrect.

For testing protos, I would go to round nine. That is where the biggest difference lies and when the arbiting meeting took place as Jason, Charles and myself came up with different pairings. Again it all came down to B4 or C6 takes precedence.

auriga
25-01-2006, 04:05 PM
pax, maybe you could add the swissmaster view(http://www.schaakbond.nl/swissmaster/bestele.htm)
to this small experiment??

from my 5 minutes of googling it seems like swissmaster
is used at the big tournaments (eg. aeroflot open 2005)
whereas protos seems to be the more official (but less used) program.

the question would then be: why do we overide protos (and swissmaster)?

pax
25-01-2006, 04:06 PM
pax, maybe you could add the swissmaster view(http://www.schaakbond.nl/swissmaster/bestele.htm)
to this small experiment??

I'd love to. You wanna give me 50 Euros for a license?

pax
25-01-2006, 04:09 PM
Round 8 official:

Official:


1 2556 2631 GM Rogers, Ian (5.5) 1 - 0 GM Schmaltz, Roland (5) 2544 2530
2 2421 2405 IM Smerdon, David (5.5) 1 - 0 FM Xie, George (4.5) 2340 2324
3 2461 2475 IM Zhao, Zong-Yuan (4) 0 - 1 GM Chandler, Murray (5) 2537 2
4 2399 2408 FM Bjelobrk, Igor (4) 0 - 1 GM Ftacnik, Lubomir (4.5) 2612 2690
5 2235 2237 Smirnov, Vladimir (4) ˝ - ˝ GM Johansen, Darryl (4) 2462 2453
6 2445 2484 IM Lane, Gary (4) 1 - 0 FM Canfell, Greg (4) 2354 2296
7 2051 2073 Song, Raymond (4) 0 - 1 IM Wohl, Alex (4) 2439 2538
8 2251 2120 FM Guthrie, Aaron (3.5) 0 - 1 IM Solomon, Stephen (3.5) 2415 2391
9 2335 2315 FM Goldenberg, Igor (3.5) 1 - 0 FM Humphrey, Jonathan (3.5) 2254 2189
10 2012 2054 Obst, James (3) 0 - 1 Ly, Moulthun (3.5) 2162 2148
11 2179 2184 Booth, Stewart (3) ˝ - ˝ WIM Moylan, Laura (2.5) 2112 2093
12 2207 2145 Lakner, Jay (2.5) 0 - 1 Wallis, Christopher (2.5) 2114 2055
13 2146 2105 Bird, Andrew (2.5) 0 - 1 Pyke, Malcolm (2.5) 2126 2079
14 2207 2293 WIM Caoili, Arianne (2.5) 1 - 0 Frame, Nigel (1.5) 0 1850
15 2223 0 FM Dougherty, Michael (2) 1 Bye


PROTOS:


1 2 Rogers ( 5.5) - 3 Schmaltz ( 5 )
2 6 Zhao ( 4 ) - 9 Smerdon ( 5.5)
3 8 Wohl ( 4 ) - 4 Chandler ( 5 )
4 1 Ftacnik ( 4.5) - 13 Xie ( 4.5)
5 18 Smirnov ( 4 ) - 5 Johansen ( 4 )
6 7 Lane ( 4 ) - 12 Canfell ( 4 )
7 11 Bjelobrk ( 4 ) - 28 Song ( 4 )
8 16 Guthrie ( 3.5) - 10 Solomon ( 3.5)
9 14 Goldenberg ( 3.5) - 15 Humphrey ( 3.5)
10 29 Obst ( 3 ) - 23 Ly ( 3.5)
11 22 Booth ( 3 ) - 27 Moylan ( 2.5)
12 21 Lakner ( 2.5) - 26 Wallis ( 2.5)
13 24 Bird ( 2.5) - 25 Pyke ( 2.5)
14 20 Caoili ( 2.5) - 30 Frame ( 1.5)
15 19 Dougherty ( 2 ) - 0 to receive the bye

Here PROTOS does something strange by not pairing Smerdon with the highest ranked valid opponent (Xie). Instead it seems to be double floating in search of a valid colour preference..

auriga
25-01-2006, 04:09 PM
I'd love to. You wanna give me 50 Euros for a license?

you can download free can't you?

i downloaded and installed. having tried feeding in some data though!

pax
25-01-2006, 04:22 PM
Round 9:

Official:


1 2439 2538 IM Wohl, Alex (5) ˝ - ˝ GM Rogers, Ian (6.5) 2556 2631
2 2415 2391 IM Solomon, Stephen (4.5) ˝ - ˝ IM Smerdon, David (6.5) 2421 2405
3 2544 2530 GM Schmaltz, Roland (5) ˝ - ˝ GM Chandler, Murray (6) 2537 2
4 2612 2690 GM Ftacnik, Lubomir (5.5) ˝ - ˝ IM Lane, Gary (5) 2445 2484
5 2340 2324 FM Xie, George (4.5) ˝ - ˝ FM Goldenberg, Igor (4.5) 2335 2315
6 2162 2148 Ly, Moulthun (4.5) ˝ - ˝ Smirnov, Vladimir (4.5) 2235 2237
7 2462 2453 GM Johansen, Darryl (4.5) ˝ - ˝ IM Zhao, Zong-Yuan (4) 2461 2475
8 2354 2296 FM Canfell, Greg (4) 1 - 0 FM Bjelobrk, Igor (4) 2399 2408
9 2114 2055 Wallis, Christopher (3.5) 0 - 1 Song, Raymond (4) 2051 2073
10 2254 2189 FM Humphrey, Jonathan (3.5) 1 - 0 Booth, Stewart (3.5) 2179 2184
11 2126 2079 Pyke, Malcolm (3.5) 1 - 0 FM Guthrie, Aaron (3.5) 2251 2120
12 2112 2093 WIM Moylan, Laura (3) ˝ - ˝ WIM Caoili, Arianne (3.5) 2207 2293
13 2207 2145 Lakner, Jay (2.5) 1 - 0 FM Dougherty, Michael (3) 2223 0
14 0 1850 Frame, Nigel (1.5) 0 - 1 Obst, James (3) 2012 2054
15 2146 2105 Bird, Andrew (2.5) 1 Bye


PROTOS:


1 8 Wohl ( 5 ) - 2 Rogers ( 6.5)
2 10 Solomon ( 4.5) - 9 Smerdon ( 6.5)
3 3 Schmaltz ( 5 ) - 4 Chandler ( 6 )
4 1 Ftacnik ( 5.5) - 7 Lane ( 5 )
5 13 Xie ( 4.5) - 14 Goldenberg ( 4.5)
6 23 Ly ( 4.5) - 18 Smirnov ( 4.5)
7 5 Johansen ( 4.5) - 6 Zhao ( 4 )
8 12 Canfell ( 4 ) - 11 Bjelobrk ( 4 )
9 26 Wallis ( 3.5) - 28 Song ( 4 )
10 15 Humphrey ( 3.5) - 19 Dougherty ( 3 )
11 27 Moylan ( 3 ) - 20 Caoili ( 3.5)
12 29 Obst ( 3 ) - 22 Booth ( 3.5)
13 25 Pyke ( 3.5) - 21 Lakner ( 2.5)
14 30 Frame ( 1.5) - 16 Guthrie ( 3.5)
15 24 Bird ( 2.5) - 0 to receive the bye


Once again, it looks like PROTOS can't be arsed after board 9.

pax
25-01-2006, 04:22 PM
you can download free can't you?

i downloaded and installed. having tried feeding in some data though!

You can only do three rounds in the free version.

auriga
25-01-2006, 04:28 PM
You can only do three rounds in the free version.

but we only want to do 1 round :)

oh well, i'm off to the pub.

pax
25-01-2006, 04:34 PM
but we only want to do 1 round :)

oh well, i'm off to the pub.

Yes, but we want to do round nine. A bit tricky when you can only enter the first three...

Bill Gletsos
25-01-2006, 05:09 PM
Note round 7 as actually played was not the pairings as generated by Swiss Perfect.

Neither were rounds 9 or 11.

The Sp regressio test shows:

Round 1
pairings identical
Round 2
pairings identical
Round 3
pairings identical
Round 4
pairings identical
Round 5
pairings identical
Round 6
pairings identical
Round 7
pair 5: existing 12-6, proposed by SP 11-6
pair 6: existing 8-11, proposed by SP 8-18
pair 7: existing 10-18, proposed by SP 10-12
pair 9: existing 27-28, proposed by SP 29-28
pair 11: existing 23-20, proposed by SP 27-20
pair 12: existing 25-29, proposed by SP 25-23
Round 8
pairings identical
Round 9
pair 12: existing 27-20, proposed by SP 21-20
pair 13: existing 21-19, proposed by SP 27-19
Round 10
pairings identical
Round 11
pair 1: existing 28-2, proposed by SP 9-28
pair 2: existing 9-14, proposed by SP 13-2
pair 3: existing 10-4, proposed by SP 4-8
pair 4: existing 8-3, proposed by SP 14-3
pair 5: existing 1-13, proposed by SP 1-18
pair 6: existing 7-18, proposed by SP 10-7
pair 10: existing 23-22, proposed by SP 23-19
pair 11: existing 20-16, proposed by SP 24-22
pair 12: existing 24-19, proposed by SP 20-16

Garvinator
25-01-2006, 05:12 PM
Note round 7 as actually played was not the pairings as generated by Swiss Perfect.

Neither were rounds 9 or 11.
Again Bill is correct and I should have added this. Interesting to note that it is the odd numbered rounds that had pairings changed from what swiss perfect generated.

The pairings used for the Australian Championships were manually done.

Bill Gletsos
25-01-2006, 05:39 PM
Round 7 as played:

5. Canfell [3.5] - Zhao [3.5]
6. Wohl [3.5] - Bjelobrk [3.5]
7. Solomon [3] - Smirnov [3.5]
8. Humphrey [3]] -Guthrie[3]
9. Moylan [2.5] - Song [3]
10. Bird [2.5] - Goldenberg [2.5]
11. Ly [2.5] - Caoili [2.5]
12. Pyke [2] - Obst [2.5]


Round 7 as generated by SP:

5. Bjelobrk [3.5] - Zhao [3.5] same as PROTOS
6. Wohl [3.5] - Smirnov [3.5] same as PROTOS
7. Solomon [3] - Canfell [3.5] same as PROTOS
8. Humphrey [3]] - Guthrie[3] same as PROTOS
9. Obst [2.5] - Song [3] same as PROTOS
10. Bird [2.5] - Goldenberg [2.5]
11. Moylan [2.5] - Caoili [2.5]
12. Pyke [2] - Ly [2.5] same as PROTOS

Round 9 as played:

Moylan [3] - Caoili [3.5]
Lakner [2.5] - Dougherty [3]

Round 9 as generated by SP:

Lakner [2.5] - Caoili [3.5]
Moylan [3] - Dougherty [3]

jenni
25-01-2006, 06:25 PM
pax, maybe you could add the swissmaster view(http://www.schaakbond.nl/swissmaster/bestele.htm)
to this small experiment??

from my 5 minutes of googling it seems like swissmaster
is used at the big tournaments (eg. aeroflot open 2005)
whereas protos seems to be the more official (but less used) program.

the question would then be: why do we overide protos (and swissmaster)?

We had a lot of discussion about this in Brisbane. A copy of swiss master 5 is being purchased and it will be parallel run at Doeberl to examine differences.

Garvinator
25-01-2006, 06:30 PM
We had a lot of discussion about this in Brisbane. A copy of swiss master 5 is being purchased and it will be parallel run at Doeberl to examine differences.
i wonder how many requests you will get for the reg key;) I can think of one right now:lol:

jenni
25-01-2006, 06:36 PM
i wonder how many requests you will get for the reg key;) I can think of one right now:lol:

Depends how ethical we are - my understanding is that the 50 euros only allows you to put the program on 1 computer.....

I think we need to do the trial at Doeberl first - rumour has it that it might not make any difference... From what CHarles has told me it appears there is a twofold problem - that (some) Aussie arbiters feel that local conditions are different to Europe and that the rules need clarification or amending. Charles has indicated that he will be seeking an official clarification from FIDE on the rules.

I d feel that it is imperative that we move to a situation where no manual intervention is necessary (apart from an extraordinary situation).

Garvinator
25-01-2006, 07:07 PM
I think we need to do the trial at Doeberl first - rumour has it that it might not make any difference... From what CHarles has told me it appears there is a twofold problem - that (some) Aussie arbiters feel that local conditions are different to Europe and that the rules need clarification or amending.
And in this situation, Charles is correct. In the Europe tournaments that is often referred to from when the dutch pairings rules were written and devised, the difference in ratings between top and bottom players is something like 400 points. Therefore each player receiving their 'due' colour is very important to a balanced tournament and 'fair' pairings. More important than say S1 v S2 for instance.

Out here in Australia, it is very common to have tournaments where the top to bottom is 1500 rating points at least. Therefore pairing changes because of colours can change pairing rating differentials of a few hundred points, making the change of colours rather academic.

My personal opinion is that I dont care much for S1 and S2 pairings at all as I think it leads to large pairing rating differentials for each pairing in each score group. But this is a personal opinion and I do try and follow the dutch pairing rules while they written as they are.


Charles has indicated that he will be seeking an official clarification from FIDE on the rules. I would go further and say that this asking for official clarification or re-writing of the dutch pairing rules should come from the acf ie national federation, not just one person.


I feel that it is imperative that we move to a situation where no manual intervention is necessary (apart from an extraordinary situation).
Would be great.

Denis_Jessop
25-01-2006, 07:47 PM
We had a lot of discussion about this in Brisbane. A copy of swiss master 5 is being purchased and it will be parallel run at Doeberl to examine differences.

Good luck Jenni. You may find you need a Doctorate in Computer Sciences to run it from my experience. Swiss Manager, which is widely used in top events, may now have the edge and looks better but it costs a lot - 99 Euros for the light version limited to 50 players and 11 rounds and 199 Euros for the full version (about $AUD162 and $AUD325 respectively).

DJ

PS I also draw your attention to this post of mine of 27.5.05

Changing computer pairings/ Pairing Programs
Not long ago there was a mighty debate about this which seems to be buried somewhere.

The participants may be interested in the following extract from Geurt Gijssen's Arbiter's Notebook for May:

"Question 1 Dear Geurt Gijssen, I have some questions regarding Swiss tournaments. In a ten round FIDE rated Swiss system tournament the following rule was instituted:
A half point bye will be given for any two rounds in a ten round FIDE rated classical chess tournament of two hours duration for each player if the player applies in advance before the commencement of the first round.

Is this rule correct, can a player receive one point without playing two rounds in a FIDE rated tournament? Won’t this disturb the ranking order of the players after the tenth round?

Question 2 Which is the best Swiss pairing software available for a FIDE rated classical chess tournament?

Question 3 The Swiss pairing program SP98 allows the pairings to be changed manually and this could be exploited by an unethical organizer to favor certain players in order to obtain a higher FIDE rating; meaning that one could pair someone against weaker players by manually changing the pairing. Can you suggest a foolproof Swiss pairing program to avoid such a situation? S. K. Talwar (India)

Answer 1 The rule regarding half-point byes is not forbidden. However, these byes are, of course, excluded from the rating calculations. I know of several tournaments in Holland that use the same rule.

Answer 2 It is very difficult to recommend a Swiss pairing program. At the moment there are two Windows-based programs and both have demos available for download:

* The Austrian program Swiss Manager
* The Dutch program Swissmaster


I have worked with both programs and I was very satisfied.

Answer 3 I understand your problem, but all Swiss pairing programs allow for the pairings to be changed manually. In fact, it is a necessity. There are instances in which a pairing may need to be changed for political reasons or if players from the same federation cannot be paired against each other. These are the only reasons to change pairings as far as I know.

It is unethical to change the pairings to benefit certain players and I can assure you that the majority of arbiters don’t do it. FIDE has issued a statement that they will not rate a tournament if it is discovered that an arbiter has been cheating in such a way."

DJ
__________________
Protos doesn't get a mention, I assume, as it is a DOS (pre-historic) program.

jase
25-01-2006, 08:43 PM
Charles has indicated that he will be seeking an official clarification from FIDE on the rules.

This has been done already - I believe it was published in the Arbiter's Notebook column on Chess Cafe (this column is written by the author of the Dutch pairing system). I'll try to source the relevant column.

It had been claimed some time ago that the rules were not as clearly written as they could be; however in some instances an arbiter has an overwhelming preference for a different system (the Lim pairing system, which was in vogue until a decade ago) and is less than capable of complying with the Dutch pairing rules. It was demonstrated in Brisbane via the pairing rules in his own rulebook, supported in practice by several independent arbiters, that Charles made fundamental errors in his manual adjustment of the round 7 pairings for the Australian Championships. I think it likely that similar errors were therefore made in pairing the Juniors event.



I d feel that it is imperative that we move to a situation where no manual intervention is necessary (apart from an extraordinary situation).

Agreed. And I think that there are arbiters such as Shaun and myself who uphold this principle. Gary Bekker and I, along with Garvin, trialled PROTOS and found it to be incapable of satisfactorily pairing the bottom half of the field in later rounds; Swiss Manager we found to be excellent, however we were only able to investigate up until round 5, when our freeware licence cut out. I'll be very interested in how the program analysis goes at Doeberl.

I also think that some players, and parents, ought to focus more on chess, and less on pairings. The vagaries of a Swiss tournament are not unknown, and you hope that it all works out fairly in the end.
Which it does, more or less.
I received several complaints from parents during the Championships in Brisbane. I investigated each - one because it was an official complaint and I was compelled to, the others out of curiosity and a willingness to assist others in their understanding of pairing rules - and found each gripe to have no grounds. On each occasion, the aggrieved party thought they/their child should be playing somebody weaker. There were some young players getting very stressed about pairings - a condition I observed to be mirrored in the accompanying parent/s.

A little learning is a dangerous thing; drink deep, or taste not the Pierian spring: there shallow draughts intoxicate the brain, and drinking largely sobers us again. :idea:


In the Europe tournaments that is often referred to from when the dutch pairings rules were written and devised, the difference in ratings between top and bottom players is something like 400 points. Therefore each player receiving their 'due' colour is very important to a balanced tournament and 'fair' pairings. More important than say S1 v S2 for instance.
In my opinion the Lim pairing system produces more equitable pairings. However this system is less prescribed, and with increasing scrutiny (eg this thread) there is a prevailing view that there should be uniformity, even if it also brings inferiority. Unfortunately there is no program for the Lim system; it requires a good arbiter who has a sound understanding of pairings. Ahh, I do miss the days when I would turn up to run an event and it was the packet of index cards, not my laptop, that was the essential item to pack ...:rolleyes:

Garvinator
25-01-2006, 09:23 PM
In my opinion the Lim pairing system produces more equitable pairings. However this system is less prescribed, and with increasing scrutiny (eg this thread) there is a prevailing view that there should be uniformity, (my bolding and underlining) even if it also brings inferiority. Unfortunately there is no program for the Lim system; it requires a good arbiter who has a sound understanding of pairings. Ahh, I do miss the days when I would turn up to run an event and it was the packet of index cards, not my laptop, that was the essential item to pack ...:rolleyes:
I actually dont think there would be too many grumbles if the Lim Pairing system was used and a pairing program was devised for it.
I think I have said before that swiss sys, which runs I think under the USCF pairing rules, uses the Lim system for its pairings. I just had a quick look at it is USCF, but couldnt find Lim system actually mentioned though.

Bill has commented on this before I think.

pax
25-01-2006, 09:24 PM
Swiss Manager we found to be excellent, however we were only able to investigate up until round 5, when our freeware licence cut out. I'll be very interested in how the program analysis goes at Doeberl.


It looks to me that five rounds is not enough. PROTOS, SP and the Official pairings all agreed up to round six.

I also suspect that the Doeberl will not be a very good test of Swiss Manager, since Doeberl is a large Swiss with few rounds. More interesting would be to run it retrospectively over the smaller Brisbane events.



In my opinion the Lim pairing system produces more equitable pairings. However this system is less prescribed, and with increasing scrutiny (eg this thread) there is a prevailing view that there should be uniformity, even if it also brings inferiority. Unfortunately there is no program for the Lim system; it requires a good arbiter who has a sound understanding of pairings. Ahh, I do miss the days when I would turn up to run an event and it was the packet of index cards, not my laptop, that was the essential item to pack ...:rolleyes:

I think PROTOS claims to pair Lim, but I don't believe this can be done unambiguously..

Garvinator
25-01-2006, 09:28 PM
It looks to me that five rounds is not enough. PROTOS, SP and the Official pairings all agreed up to round six.

I also suspect that the Doeberl will not be a very good test of Swiss Manager, since Doeberl is a large Swiss with few rounds. More interesting would be to run it retrospectively over the smaller Brisbane events.
would like to do and I have three events coming up in the next couple of months that would be perfect;) I would prefer to actually Swiss Master 5 a go as it is cheaper for the licence and I dont think it is that difficult to use.

Bill Gletsos
25-01-2006, 10:15 PM
And in this situation, Charles is correct. In the Europe tournaments that is often referred to from when the dutch pairings rules were written and devised, the difference in ratings between top and bottom players is something like 400 points. Therefore each player receiving their 'due' colour is very important to a balanced tournament and 'fair' pairings. More important than say S1 v S2 for instance.

Out here in Australia, it is very common to have tournaments where the top to bottom is 1500 rating points at least. Therefore pairing changes because of colours can change pairing rating differentials of a few hundred points, making the change of colours rather academic.

My personal opinion is that I dont care much for S1 and S2 pairings at all as I think it leads to large pairing rating differentials for each pairing in each score group. But this is a personal opinion and I do try and follow the dutch pairing rules while they written as they are.Exactly.
What you or I may think of the Dutch rule is immaterial.
If you are going to run an event based on them then they should be followed.
They shouldnt be open to interpretation by different arbiters.

Now in reality it should be as simple as that.

Unfortunately the rules as written are no absolutely 10%% clear.
What FIDE needs to do is remove any ambiguity.

e.g. in C1 and C6 where it clearly and explicitly only mentions B1 and B2 does this mean that B3 to B6 have no bearing on C1 and C6 whatsoever or do they still apply to some degree. It appears that some local arbiters would say yes and some say no, however I suspect you would get a consistent answer from the members of FIDE Swiss Pairings Commission.

Bill Gletsos
25-01-2006, 10:18 PM
I actually dont think there would be too many grumbles if the Lim Pairing system was used and a pairing program was devised for it.
I think I have said before that swiss sys, which runs I think under the USCF pairing rules, uses the Lim system for its pairings. I just had a quick look at it is USCF, but couldnt find Lim system actually mentioned though.

Bill has commented on this before I think.I have commented on the LIM system before and that it is described in the FIDE handbook.

However I do not believe I have ever stated that Swiss-sys uses it.
I do know that it uses the USCF swiss pairing rules but as far as I am aware they differ from the Lim rules.

Bill Gletsos
25-01-2006, 10:19 PM
I think PROTOS claims to pair Lim, but I don't believe this can be done unambiguously..What do you mean by that.

Bill Gletsos
26-01-2006, 01:17 AM
Stewart Reuben in the Second Edition of his Chess Organiser's Handbook said:
I have never heard comments of the Petunia Program or the Dutch System giving the wrong pairings.

In the Third Edition he replaced this with:
I have never heard comments about the Swiss46 and higher numbers of the Dutch System giving the wrong pairings. This program is now available in Windows.

And later he says of the Dutch System:
This is now available as a Windows computerised program. I have never known it to provide an incorrect answer and am puzzled as to why anybody bothers with any other computerised seeded system.


Swiss46 is better known as Swiss Master 4.6.
This is the program previously know as Petunia.


So how does it go with generating pairings for the Australian Championship 2005-2006 and how did Swiss Perfect compare to it.

Also how did the manual pairings done in rounds 7, 9 and 11 over-riding Swiss perfect compare to Swiss Master.


Rounds 1-6 are identical to those generated by Swiss Perfect.

Round 7 is exactly as Swiss Perfect generated it, but not as round 7 was actually played due to it being manually paired.

Pairing of round 7 of Australian Championship 2005-2006
Table White - Black Results
----------------------------------------------------------------------- round 7
1 Chandler ( 5 ) - Rogers ( 4«) 4- 2
2 Xie ( 4 ) - Schmaltz ( 4«) 13- 3
3 Smerdon ( 4«) - Lane ( 4 ) 9- 7
4 Johansen ( 3«) - Ftacnik ( 4 ) 5- 1
5 Bjelobrk ( 3«) - Zhao ( 3«) 11- 6
6 Wohl ( 3«) - Smirnov ( 3«) 8- 18
7 Solomon ( 3 ) - Canfell ( 3«) 10- 12
8 Humphrey ( 3 ) - Guthrie ( 3 ) 15- 16
9 Obst ( 2«) - Song ( 3 ) 29- 28
10 Bird ( 2«) - Goldenberg ( 2«) 24- 14
11 Moylan ( 2«) - Caoili ( 2«) 27- 20
12 Pyke ( 2 ) - Ly ( 2«) 25- 23
13 Dougherty ( 2 ) - Booth ( 2 ) 19- 22
14 Frame ( 1«) - Lakner ( 1«) 30- 21
15 Wallis ( 1«) - Bye 26-

Round 8 pairings are identical to those generated by Swiss Perfect.

Round 9 differs from the generated SP pairings on boards 12 & 13.
It exactly matches the manual pairings for boards 12 & 13.


12 Moylan ( 3 ) - Caoili ( 3«) 27- 20
13 Lakner ( 2«) - Dougherty ( 3 ) 21- 19

Round 10 pairings are identical to those generated by Swiss Perfect.

Round 11 pairings match SP generated pairings on boards 1 to 9 and pairings on boards 12 and 13 match SP pairings on boards 13 & 14.

Note that the pairings done manually do not match with the first 6 boards of either SP or Swiss Master.


Round 11 pairings as manually paired and played:

No Name Feder Rtg Total Result Name Feder Rtg Total

1 Song, Raymond NSW 2051 [6] Rogers, Ian NSW 2556 [8]
2 Smerdon, David VIC 2421 [8] Goldenberg, Igor VIC 2335 [5.5]
3 Solomon, Stephen QLD 2415 [5.5] Chandler, Murray ENG 2537 [7.5]
4 Wohl, Alex NSW 2439 [5.5] Schmaltz, Roland GER 2544 [6.5]
5 Ftacnik, Lubomir SVK 2612 [6] Xie, George NSW 2340 [5.5]
6 Lane, Gary NSW 2445 [6] Smirnov, Vladimir NSW 2235 [5.5]
7 Johansen, Darryl VIC 2462 [5] Canfell, Greg NSW 2354 [5]
8 Zhao, Zong-Yuan NSW 2461 [5] Pyke, Malcolm VIC 2126 [5]
9 Humphrey, Jonathan QLD 2254 [5] Bjelobrk, Igor VIC 2399 [5]
10 Ly, Moulthun QLD 2162 [5] Booth, Stewart VIC 2179 [4.5]
11 Caoili, Arianne QLD 2207 [4] Guthrie, Aaron SA 2251 [4.5]
12 Bird, Andrew NSW 2146 [4.5] Dougherty, Michael CAN 2223 [3.5]
13 Frame, Nigel TAS [1.5] Moylan, Laura NSW 2112 [4.5]
14 Wallis, Christopher VIC 2114 [4] Obst, James SA 2012 [4]
15 Lakner, Jay WA 2207 [3.5] BYE

Round 11 pairings as generated by Swiss Perfect:

No Name Feder Rtg Total Result Name Feder Rtg Total

1 Smerdon, David VIC 2421 [8] : Song, Raymond NSW 2051 [6]
2 Xie, George NSW 2340 [5.5] : Rogers, Ian NSW 2556 [8]
3 Chandler, Murray ENG 2537 [7.5] : Wohl, Alex NSW 2439 [5.5]
4 Goldenberg, Igor VIC 2335 [5.5] : Schmaltz, Roland GER 2544 [6.5]
5 Ftacnik, Lubomir SVK 2612 [6] : Smirnov, Vladimir NSW 2235 [5.5]
6 Solomon, Stephen QLD 2415 [5.5] : Lane, Gary NSW 2445 [6]
7 Johansen, Darryl VIC 2462 [5] : Canfell, Greg NSW 2354 [5]
8 Zhao, Zong-Yuan NSW 2461 [5] : Pyke, Malcolm VIC 2126 [5]
9 Humphrey, Jonathan QLD 2254 [5] : Bjelobrk, Igor VIC 2399 [5]
10 Ly, Moulthun QLD 2162 [5] : Dougherty, Michael CAN 2223 [3.5]
11 Bird, Andrew NSW 2146 [4.5] : Booth, Stewart VIC 2179 [4.5]
12 Caoili, Arianne QLD 2207 [4] : Guthrie, Aaron SA 2251 [4.5]
13 Frame, Nigel TAS [1.5] : Moylan, Laura NSW 2112 [4.5]
14 Wallis, Christopher VIC 2114 [4] : Obst, James SA 2012 [4]
15 Lakner, Jay WA 2207 [3.5] BYE

Round 11 pairings as generated by Swiss Master:

Pairing of round 11 of Australian Championship 2005-2006
Table White - Black Results
----------------------------------------------------------------------- round 11
1 Smerdon ( 8 ) - Song ( 6 ) 9- 28
2 Xie ( 5«) - Rogers ( 8 ) 13- 2
3 Chandler ( 7«) - Wohl ( 5«) 4- 8
4 Goldenberg ( 5«) - Schmaltz ( 6«) 14- 3
5 Ftacnik ( 6 ) - Smirnov ( 5«) 1- 18
6 Solomon ( 5«) - Lane ( 6 ) 10- 7
7 Johansen ( 5 ) - Canfell ( 5 ) 5- 12
8 Zhao ( 5 ) - Pyke ( 5 ) 6- 25
9 Humphrey ( 5 ) - Bjelobrk ( 5 ) 15- 11
10 Bird ( 4«) - Ly ( 5 ) 24- 23
11 Guthrie ( 4«) - Booth ( 4«) 16- 22
12 Frame ( 1«) - Moylan ( 4«) 30- 27
13 Wallis ( 4 ) - Obst ( 4 ) 26- 29
14 Caoili ( 4 ) - Dougherty ( 3«) 20- 19
15 Lakner ( 4 ) - Bye 21-


I think what can be drawn from this is that Swiss Prfect did very good job compared to Swiss Master, only differing with 2 pairings on the lower boards in round 9 and 3 pairings on the lower boards in round 11.

Garvinator
26-01-2006, 02:25 AM
I do believe something has been missed by either myself ;) or the pairing programs.

In round seven, both computer pairing programs have downfloated Raymond Song. Isnt this an error:

B.5 No player shall receive an identical float in two consecutive rounds.

Therefore, shouldnt Song have stayed in his own scoregroup? In the manual pairings done by Charles for round seven, this was also missed.

Jase also makes note of this for round seven in post 20 I believe:


It was demonstrated in Brisbane via the pairing rules in his own rulebook, supported in practice by several independent arbiters, that Charles made fundamental errors in his manual adjustment of the round 7 pairings for the Australian Championships.

pax
26-01-2006, 07:26 AM
What do you mean by that.

I mean that two arbiters correctly following Lim could produce different pairings..

pax
26-01-2006, 08:22 AM
I do believe something has been missed by either myself ;) or the pairing programs.

In round seven, both computer pairing programs have downfloated Raymond Song. Isnt this an error:

B.5 No player shall receive an identical float in two consecutive rounds.

Therefore, shouldnt Song have stayed in his own scoregroup? In the manual pairings done by Charles for round seven, this was also missed.

Jase also makes note of this for round seven in post 20 I believe:

In order to claim this, you need to propose a pairing which achieves an equal or better colour balance, and floats a different player. I am looking into this now..

pax
26-01-2006, 08:30 AM
In order to claim this, you need to propose a pairing which achieves an equal or better colour balance, and floats a different player. I am looking into this now..

I actually think the R7 Song downfloat is correct.

You have four players on 3 (Humphrey, Solomon, Guthrie and Song) plus the downfloater Smirnov. Solomon and Humphrey are due white, the others black. Therefore the downfloater must be Song or Guthrie. Humphrey has played both Smirnov and Song, so must play Guthrie. Therefore Song must float.

Thoughts?

pax
26-01-2006, 08:45 AM
Jase also makes note of this for round seven in post 20 I believe:

He obviously wasn't referring to the downfloat of Song, since he said the manual adjustment was in error. The manual and the SP pairings (and SM and PROTOS) all downfloated Song.

Bill Gletsos
26-01-2006, 11:03 AM
I actually think the R7 Song downfloat is correct.

You have four players on 3 (Humphrey, Solomon, Guthrie and Song) plus the downfloater Smirnov. Solomon and Humphrey are due white, the others black. Therefore the downfloater must be Song or Guthrie. Humphrey has played both Smirnov and Song, so must play Guthrie. Therefore Song must float.

Thoughts?You are correct.

jenni
27-01-2006, 03:01 PM
I also think that some players, and parents, ought to focus more on chess, and less on pairings. The vagaries of a Swiss tournament are not unknown, and you hope that it all works out fairly in the end.
Which it does, more or less.
I received several complaints from parents during the Championships in Brisbane. I investigated each - one because it was an official complaint and I was compelled to, the others out of curiosity and a willingness to assist others in their understanding of pairing rules - and found each gripe to have no grounds. On each occasion, the aggrieved party thought they/their child should be playing somebody weaker. There were some young players getting very stressed about pairings - a condition I observed to be mirrored in the accompanying parent/s.

You always get paranoia developing in high level tournaments, particularly junior ones. I have to say I heard quite a few complaints from quite high rated adult players in the Championships. I only heard about one very vocal (and quite incorrect) complaint in the juniors, although no doubt there were others.

The problem is that once people get to know that pairings are being manually changed it is going to lead to them wondering if they are being disadvantaged, however untrue that might be.

While I am sure it would make arbiters life a lot easier if people treated them like gods and just accepted what happened that is not the way our society runs these days - talk to older doctors. They used to have it easy, but now everyone checks stuff on the net and cross questions their diagnoses. :)

Parents don't like being mushrooms, so they are going to want to understand what is going on.

I have been attacked occasionally over a pairing when running junior tournaments in Canberra. However as we never change pairings and the draw is done in the open, usually with a dozen kids watching, it is easy to just say "tough bikkies, that is what the computer did. "



Ahh, I do miss the days when I would turn up to run an event and it was the packet of index cards, not my laptop, that was the essential item to pack ...:rolleyes:
I don't - I can remember how long it used to take to do it. When you are running schools comps with 200 kids playing and a tight deadline (kids have to be back to their schools by 2:30, so they don't miss their bus home), being able to press a button and have a set of pairings in 10 seconds is a blessing.

Nothing worse than having a paring take 20 minutes to do in the last round and then just as everyone is seated someone screeches "we've played each other". :eek:

Rincewind
27-01-2006, 03:08 PM
I have been attacked occasionally over a pairing when running junior tournaments in Canberra. However as we never change pairings and the draw is done in the open, usually with a dozen kids watching, it is easy to just say "tough bikkies, that is what the computer did. "

This also works at senior events at my club.

Garvinator
28-01-2006, 03:11 AM
When you are running schools comps with 200 kids playing and a tight deadline (kids have to be back to their schools by 2:30, so they don't miss their bus home), being able to press a button and have a set of pairings in 10 seconds is a blessing.
Hello Jenni,

When a tournament has 200 players and 10 rounds, swissperfect will cope pretty well with this as normally in each round, each person's pairings are coming from the same score group.

Where sp starts to struggle is when it needs to go outside of a players' own scoregroup to find an acceptable pairing. Sp seems to struggle with which pairing rule is supposed to be employed at which time.

Quite a few arbiters have noticed that swissperfect- when attempting to pair a player with someone from a lower scoregroup- will sometimes go down two scoregroups in an attempt to balance colours, even though there was already a legal pairing in the scoregroup half a point below. It is even more likely to do this if, by taking this action, it will lead to all the other players playing opponents on the same scores.

This is just wrong. Players are supposed to be paired from highest score group first all the way to the bottom.

Ian Rout
28-01-2006, 08:01 AM
Quite a few arbiters have noticed that swissperfect- when attempting to pair a player with someone from a lower scoregroup- will sometimes go down two scoregroups in an attempt to balance colours, even though there was already a legal pairing in the scoregroup half a point below. It is even more likely to do this if, by taking this action, it will lead to all the other players playing opponents on the same scores.
It isn't exactly to match colours, because minimising the difference in scores takes priority over colours. However downfloating a player half a point to the lower group, and then floating another player from that group half a point, is the same in total as downfloating one player two groups. Thus if the program is aiming to minimise score difference in total rather than group by group the two options are equal and it will then go to the next priority (colour). Although that's a way to interpret the rules it's not one that most people would find desirable.

Because of quirks like this it isn't viable to say that you will always use the computer pairings, though such an approach is expedient and adequate most of the time.

Denis_Jessop
28-01-2006, 10:10 AM
You always get paranoia developing in high level tournaments, particularly junior ones. I have to say I heard quite a few complaints from quite high rated adult players in the Championships. I only heard about one very vocal (and quite incorrect) complaint in the juniors, although no doubt there were others.

The problem is that once people get to know that pairings are being manually changed it is going to lead to them wondering if they are being disadvantaged, however untrue that might be.

<snip>

Parents don't like being mushrooms, so they are going to want to understand what is going on.

I have been attacked occasionally over a pairing when running junior tournaments in Canberra. However as we never change pairings and the draw is done in the open, usually with a dozen kids watching, it is easy to just say "tough bikkies, that is what the computer did. "


<snip>


I fully agree that, if it becomes known that arbiters will change pairings, that leaves not only those arbiters open to charges of improperly changing pairings, but also places other arbiters in the position of having players ("adults" in this case) of saying "why don't you change the pairings; that's what they do in Sydney" as was put to me in a local club event a year or so ago.

On the other hand, except perhaps with some less-experienced or placid people, you can't just get away with saying "that's what the computer did". An arbiter needs to be able to justify the computer pairing and, in doing so, be able to explain how the computer has applied the relevant swiss pairing rules. As Jason has pointed out elsewhere, FIDE has made this clear (or as clear as FIDE seems to be able to make it, which is not necessarily really clear:confused: ) in saying that computer programs are only a tool and the arbiter must take final responsibility for the application of the rules.

DJ

jenni
28-01-2006, 10:22 AM
On the other hand, except perhaps with some less-experienced or placid people, you can't just get away with saying "that's what the computer did". An arbiter needs to be able to justify the computer pairing and, in doing so, be able to explain how the computer has applied the relevant swiss pairing rules. As Jason has pointed out elsewhere, FIDE has made this clear (or as clear as FIDE seems to be able to make it, which is not necessarily really clear:confused: ) in saying that computer programs are only a tool and the arbiter must take final responsibility for the application of the rules.

DJ

This is why I am now trying to understand the pairing rules. I think I drove Cathy crazy in Brisbane by walking around with my organisers handbook and questioning everything. I have been told the best way to learn is to actually do it manually and then compare against the computer program. After a few interations you will actually understand how the pairings are done. (or so I am told).

However I do feel strongly that the more manual changes are done, the more likely it is that people will start feeling paranoid and questioning things.

As you probably know both Libby and I have been very conscious over the years that we are in a vulnerable position, because we have run so many competitions where our children were in line for prizes. Being able to say that we have no input to the draw, has made it much easier for us to do this. In fact the only time I have been challenged over the draw in a non-schools junior tournament was by a NSW person, who was used to draws being changed. Fortunately that time I was able to detect it was because of colour balancing and they were happy with the explanation.

Ian Rout
28-01-2006, 01:00 PM
I fully agree that, if it becomes known that arbiters will change pairings, that leaves not only those arbiters open to charges of improperly changing pairings, but also places other arbiters in the position of having players ("adults" in this case) of saying "why don't you change the pairings; that's what they do in Sydney" as was put to me in a local club event a year or so ago.
If you were going to make decisions on the basis of not wanting complaints about them then cricket umpires would uphold every appeal by Shane Warne (as an interesting aside, cricketers can be fined thousands of dollars or suspended for complaining about decisions even when those decisions are incorrect).

Another point to note is that the word "change" in this context is not entirely appropriate, as it implies some sort of incumbency. If an arbiter did the pairings by hand and then checked them against the computer the change would be to use the computer pairings.

pax
28-01-2006, 05:41 PM
You always get paranoia developing in high level tournaments, particularly junior ones. I have to say I heard quite a few complaints from quite high rated adult players in the Championships. I only heard about one very vocal (and quite incorrect) complaint in the juniors, although no doubt there were others.

The problem is that once people get to know that pairings are being manually changed it is going to lead to them wondering if they are being disadvantaged, however untrue that might be.


It seems to me that there is a corollary to this in Australia.

Because there are a few minor flaws in the Swiss Perfect pairing algorithm many people get the impression that Swiss Perfect pairings are usually incorrect. This leads to players frequently questioning the Swiss Perfect pairings (often incorrectly). I am pretty sure this is a false impression, and that SP pairings are usually correct.

The impression is exacerbated by the frequency of manual changes. These are sometimes due to incorrect pairings, but mostly due to differing interpretations of Swiss rules or choosing not to apply certain rules (e.g the two downfloats in three rounds rule).

Ian_Rogers
28-01-2006, 07:41 PM
I'd have to add that Swiss Perfect pairings are most often inadequate in the last round of a seven or nine round tournament.
In a tournament where seed one may have already played and beaten seeds two, four and six, the score group just below the leader could easily be seeds three five seven and ten. Then, just to avoid consecutive colours - even though all players have usually had, say, three whites and three blacks - Swiss Perfect will probably pair seed one against 10.
In Australian swisses, where seed ten may be 400 or more rating points below seed three, this will tend to ruin the sporting nature of the final round.
This has happened frequently to me in Queensland tournaments, both as leader and as a player chasing the leader. If an arbiter does not change this sort of pairing - and many don't - they are wrecking the final round of the tournament.
In an even round tournament the question is more arguable but it would still be better for seed one to play seed three and have someone have imbalanced colours rather than wreck the sporting contest.

Notably, some NZ administrators conducted an informal survey of their top players after the last round of Queenstown, asking whether they would prefer to see top half against bottom half or balanced colours as a priority. The players (Dive, Smith, etc) were unanimous in preferring top half v bottom half. (The example given to them was a four player group and none were happy with 1v2 and 3v4, even in an even round to balance colours.)

Ian

Denis_Jessop
28-01-2006, 07:53 PM
If you were going to make decisions on the basis of not wanting complaints about them then cricket umpires would uphold every appeal by Shane Warne (as an interesting aside, cricketers can be fined thousands of dollars or suspended for complaining about decisions even when those decisions are incorrect).

Another point to note is that the word "change" in this context is not entirely appropriate, as it implies some sort of incumbency. If an arbiter did the pairings by hand and then checked them against the computer the change would be to use the computer pairings.

I wasn't talking about making decisions on the basis of not wanting complaints. What I was getting at is that, if it becomes the practice to change SP pairings for reasons other than that SP has wrongly applied the Swiss Pairing Rules - and it usually applies them correctly as mentioned in post#41 - certain kinds of people will be encouraged to question the pairings, thus making themselves more of a nuisance than usual.

I have no problems with explaining to those people how the program is right and they are wrong, nor do I have a problem with changing the pairings when SP gets them wrong as I'm sure it did in a club event for which I was arbiter a year or so ago.

As for the use of the word "change", as you know Ian, I am a lawyer, but, even with that advantage, the sublety of your point here completely escapes me.

DJ

PS: Re Ian Rogers' post #43, the wrong SP pairings to which I referred above were in just that kind of situation though also in a round a couple from the end. I think it was exacerbated by the field being small too. What SP did was to fail to pair 2 players on equal scores which it could lawfully have done, but instead to float them both down in one case across the score group immediately below. It was possible without infringing any of the Dutch system rules (as far as I could see after much checking) to make another draw that complied with the intention and the letter of the Dutch System Rules.

Garvinator
28-01-2006, 08:23 PM
Notably, some NZ administrators conducted an informal survey of their top players after the last round of Queenstown, asking whether they would prefer to see top half against bottom half or balanced colours as a priority. The players (Dive, Smith, etc) were unanimous in preferring top half v bottom half. (The example given to them was a four player group and none were happy with 1v2 and 3v4, even in an even round to balance colours.)

Ian
why werent they happy with 1v2 and 3v4, especially if it balanced colours?

Garvinator
28-01-2006, 08:31 PM
I'd have to add that Swiss Perfect pairings are most often inadequate in the last round of a seven or nine round tournament.
In a tournament where seed one may have already played and beaten seeds two, four and six, the score group just below the leader could easily be seeds three five seven and ten. Then, just to avoid consecutive colours - even though all players have usually had, say, three whites and three blacks - Swiss Perfect will probably pair seed one against 10.
In Australian swisses, where seed ten may be 400 or more rating points below seed three, this will tend to ruin the sporting nature of the final round.
This has happened frequently to me in Queensland tournaments, both as leader and as a player chasing the leader. If an arbiter does not change this sort of pairing - and many don't - they are wrecking the final round of the tournament.
In an even round tournament the question is more arguable but it would still be better for seed one to play seed three and have someone have imbalanced colours rather than wreck the sporting contest.

In this situation, I would want to know whether the dutch pairing rules have been used in previous rounds. If they have, then they should be used for the final round as well.

It may seem like it is wrecking the final round, but the previous rounds have been paired with the dutch pairing rules. Had those previous rounds been paired used a different system then maybe the pairings for the final round would be different.

For instance, in round six, Player 1 may have been due to play Player 6 and it looks like a mismatch, but then when B4 is used, the pairing changes to say Player 1 v Player 3, even though player 1 and 3 may have come from S1, instead of S1 v S2.

pax
28-01-2006, 09:03 PM
I'd have to add that Swiss Perfect pairings are most often inadequate in the last round of a seven or nine round tournament.
In a tournament where seed one may have already played and beaten seeds two, four and six, the score group just below the leader could easily be seeds three five seven and ten. Then, just to avoid consecutive colours - even though all players have usually had, say, three whites and three blacks - Swiss Perfect will probably pair seed one against 10.

I would just point out that the issue here is not with Swiss Perfect, but rather with the FIDE Dutch rules.

I agree completely that it makes sense to ignore mild colour preferences in the last round. Unfortunately I can't find any instruction to do so in the FIDE Dutch rules. The "special last round rules" are that B2, B5 and B6 be ignored when pairing score groups over 50%. Mild colour preferences are handled in B4, therefore these would seemingly still apply.

You can't blame Swiss Perfect for correctly applying the letter of the law in these situations.

Ian_Rogers
28-01-2006, 10:13 PM
"Why werent they happy with 1v2 and 3v4, especially if it balanced colours?"

For the obvious reason that colour isn't as important to a serious player as having top half v bottom half. (Nor should it be to any serious version of the Swiss system.)

Imagine if 1 is rated 2400, 2 is 2300, 3 is 2000 and 4 is 1200 and you begin to see why colours are a complete irrelevance to the results of these possible pairings.

Ian

Garvinator
28-01-2006, 10:35 PM
Imagine if 1 is rated 2400, 2 is 2300, 3 is 2000 and 4 is 1200 and you begin to see why colours are a complete irrelevance to the results of these possible pairings.

Ian
But then you would just get either of these two combinations:

1 v 3: 2400 v 2000
2 v 4: 2300 v 1200

1 v 2: 2400 v 2300
3 v 4: 2000 v 1200

Under any system, whether it be dutch paring system, Lim pairings or some combination of pairing system, one player, be it 1,2 or 3 is going to get the 'far easier' game in theory.

pax
28-01-2006, 10:41 PM
Imagine if 1 is rated 2400, 2 is 2300, 3 is 2000 and 4 is 1200 and you begin to see why colours are a complete irrelevance to the results of these possible pairings.

On the other hand, if 1 is rated 2421, 2 is 2397, 3 is 2396 and 4 2361 one could imagine that whether a player is white or black is likely to play a far more significant role in the result than whether he is playing a top half player or a bottom half player.

Bill Gletsos
29-01-2006, 12:30 AM
"Why werent they happy with 1v2 and 3v4, especially if it balanced colours?"

For the obvious reason that colour isn't as important to a serious player as having top half v bottom half. (Nor should it be to any serious version of the Swiss system.)Shouldnt IA's be required to follow the FIDE Swiss pairing rules and if the tournament is being played under the Dutch pairing rules then the Dutch pairing rules should be followed.

If players especially titled players have an issue with the Dutch pairing rules then shouldnt they be lobbying FIDE for a change to the Dutch rules, rather than just expecting/insisting that arbiters ignore aspects of them.

In Europe are computerised pairing programs using the Dutch rules used and if so are IA's over there changing pairings at their discretion, by ignoring aspects of the Dutch rules they dont agree with.


"Imagine if 1 is rated 2400, 2 is 2300, 3 is 2000 and 4 is 1200 and you begin to see why colours are a complete irrelevance to the results of these possible pairings.Yes, but as pax pointed out you could have the players having ratings as he suggested in which case colour is then likely to be more important than top V bottom.

Bill Gletsos
29-01-2006, 12:41 AM
I would just point out that the issue here is not with Swiss Perfect, but rather with the FIDE Dutch rules.Exactly.

I agree completely that it makes sense to ignore mild colour preferences in the last round. Unfortunately I can't find any instruction to do so in the FIDE Dutch rules. The "special last round rules" are that B2, B5 and B6 be ignored when pairing score groups over 50%. Mild colour preferences are handled in B4, therefore these would seemingly still apply.Agreed.

You can't blame Swiss Perfect for correctly applying the letter of the law in these situations.Correct.

Ian_Rogers
29-01-2006, 09:28 AM
"On the other hand, if 1 is rated 2421, 2 is 2397, 3 is 2396 and 4 2361 one could imagine that whether a player is white or black is likely to play a far more significant role in the result than whether he is playing a top half player or a bottom half player."

Yes, but you have to be consistent. In my example, pairing 1 against 2 will most likely be a disaster for the tournament, especially if this is a penultimnate round. In your example colour will be a minor inconvenience but in any case there is no way it should be 1 v 2 unless absolutely forced.

To take up Bill's point, while Iljumzhinov runs FIDE and allows Gijssen to do anything he likes - and in particular settle scores through the Rules Commission - there is no point lobbying to try to get consistency in the Swiss rules.
For example, the only reason players are no longer allowed to write down their moves before they play it is that Gijssen lost an embarrassing dispute with a player a few years ago, found out that he was wrong, and determined to change the rules so that he could retrospectively be proved right.

A good arbiter just has to announce before the tournament that certain of the more stupid FIDE rules will not be applied or enforced and conduct the event in the best interests of the players.

Ian

Ian Rout
29-01-2006, 10:45 AM
As for the use of the word "change", as you know Ian, I am a lawyer, but, even with that advantage, the sublety of your point here completely escapes me.

DJ
Sorry, I'll elaborate. The use of the word "change" in saying that pairings were changed might be read as implying the possibility of corruption or arbitrary intervention, or at least that it is a non-routine action requiring justification. For instance in another thread someone was demanding that if arbiters change the computer pairings they should provide a written explanation, or some such nonsense.

What I was saying is that the use of a set of pairings different to those generated by the computer is only a change if you start with the premise that those of the computer are the incumbent set which have some special status, rather than merely a convenient draft which may, and often is, the final.

Ian Rout
29-01-2006, 11:00 AM
I think Ian Rogers has a point in a tournament which isn't of any great significance beyond the event itself; in particular if the top seed is a titled player who has received free entry or appearance money and might be expected to jump through a few more hoops to take the prize.

On the other hand where you have championships, norms, prizes or GP points at stake I think inserting sticky fingers into the draw is questionable, especially as it is necessarily arbitrary. It isn't meant to be a handicap or exhibition, the pairing rules are intended to provide a balance whereby in the end the contenders have met similar fields so that the difference in their results is due to their performances. Deliberately giving someone a harder game in the last round (= giving their opponent a harder game and someone else an easier game) is a dangerous business.

jenni
29-01-2006, 11:04 AM
For instance in another thread someone was demanding that if arbiters change the computer pairings they should provide a written explanation, or some such nonsense.
.
Actually I said it and it was a joke, as I said to Scott a few posts later. Problem about having all the smiley faces missing at the time.

Bill Gletsos
29-01-2006, 11:37 AM
Sorry, I'll elaborate. The use of the word "change" in saying that pairings were changed might be read as implying the possibility of corruption or arbitrary intervention, or at least that it is a non-routine action requiring justification. For instance in another thread someone was demanding that if arbiters change the computer pairings they should provide a written explanation, or some such nonsense.

What I was saying is that the use of a set of pairings different to those generated by the computer is only a change if you start with the premise that those of the computer are the incumbent set which have some special status, rather than merely a convenient draft which may, and often is, the final.However the FIDE web site states The FIDE Swiss Rules pair the players in an objective and impartial way, and different arbiters following the pairing rules should arrive at identical pairings.
It follows from this that if the computer generated pairings differ from the manual pairings done by the arbiter at least one of them must be wrong.

Back in Sept 2003 Cathy Rogers wrote on the old ACF BB The opinion I wanted was from British IA Stewart Reuben who is my pick for the best arbiter in the world.

Given the statements regarding Swiss Master by Stewart Reuben that he has never seen Swiss Master give incorrect pairings then if the pairing program is Swiss Master and the manual pairings differ from its pairings, the wrong pairings must be the manual pairings.

I would take it that Reuben's when using Swiss Master isnt manually adjusting pairings.

Denis_Jessop
29-01-2006, 11:46 AM
Sorry, I'll elaborate. The use of the word "change" in saying that pairings were changed might be read as implying the possibility of corruption or arbitrary intervention, or at least that it is a non-routine action requiring justification. For instance in another thread someone was demanding that if arbiters change the computer pairings they should provide a written explanation, or some such nonsense.

What I was saying is that the use of a set of pairings different to those generated by the computer is only a change if you start with the premise that those of the computer are the incumbent set which have some special status, rather than merely a convenient draft which may, and often is, the final.

Thanks, Ian. That's now clear. The reason I was puzzled is that I take the situation to be as you describe it in your second paragraph. That is, the pairings put out by the computer program are more than a convenient draft. In my view they have the status of the proper pairings unless there is a demonstrated need to change them, for example, if they don't properly apply the relevant swiss rules. That's not to say that the arbiter is not finally responsible for the pairings (FIDE asserts he is) and the arbiter always needs to be able to explain the reason for a pairing. But, if the computer pairings had any lesser status that that put by me, it would mean that arbiters should manually check the correctness of all the pairings before publishing the draw in which case the value of having a computer program is dramatically reduced.

DJ

Denis_Jessop
29-01-2006, 11:52 AM
However the FIDE web site states The FIDE Swiss Rules pair the players in an objective and impartial way, and different arbiters following the pairing rules should arrive at identical pairings.
It follows from this that if the computer generated pairings differ from the manual pairings done by the arbiter at least one of them must be wrong.

Back in Sept 2003 Cathy Rogers wrote on the old ACF BB The opinion I wanted was from British IA Stewart Reuben who is my pick for the best arbiter in the world.

Given the statements regarding Swiss Master by Stewart Reuben that he has never seen Swiss Master give incorrect pairings then if the pairing program is Swiss Master and the manual pairings differ from its pairings, the wrong pairings must be the manual pairings.

I would take it that Reuben's when using Swiss Master isnt manually adjusting pairings.

On the other hand I see that, in his second edition (I am still awaiting the third), Reuben seems to think that it's OK for an arbiter to adjust pairings so as to give a player a chance for a norm and FIDE has said that's tabu. (Not on the topic except as to SR's status as the world's greatest arbiter.;) )

DJ

jenni
29-01-2006, 12:24 PM
On the other hand I see that, in his second edition (I am still awaiting the third), Reuben seems to think that it's OK for an arbiter to adjust pairings so as to give a player a chance for a norm and FIDE has said that's tabu. (Not on the topic except as to SR's status as the world's greatest arbiter.;) )

DJ

If we are talking about "The Chess Organiser's Handbook" I have edition 3 and I haven't seen that comment (although it might be buried somewhere).

He does state under "The Dutch System"

"This is now available as a Windows computerised program. I have never known it to provide an incorrect answer and am puzzled as to why anybody bothers with any other computerised seeded system."

Bill Gletsos
29-01-2006, 01:48 PM
On the other hand I see that, in his second edition (I am still awaiting the third), Reuben seems to think that it's OK for an arbiter to adjust pairings so as to give a player a chance for a norm and FIDE has said that's tabu. (Not on the topic except as to SR's status as the world's greatest arbiter.;) )The full text of what is says is as follow:

It is commonplace to make minor adjustments in Swiss daws to help increase the possabilirty of players achieving title norms or ratings .e.g. a player has met 8 rated opponents. He has a realtively low score and is due to meet an unrated opponent in the last round. Instead he can meet a rated opponent on the same score. There is no prize at stake and nobody is disadvantaged.
It is difficult to write set rules for such action. It is essential to ensure that nobody is disadvantaged. The arbiter must ensure no bias. However, it is illogical to say such actions are totally wrongwhen, for example, the first round of a swiss is frequently incorrect due to late entries or no-shows. In the 1992 Olympiad there was even the absurd sight of three teams winning 4-0 by deafult all in the fourth round.

The critical wording by Reuben in all this would seem to be "It is essential to ensure that nobody is disadvantaged. The arbiter must ensure no bias."

shaun
29-01-2006, 02:09 PM
Another important question/scenario

Assuming that there is no special announcment at the start of the tournament (or on the entry form) along the lines of "We are dropping pairing rule N.x from the swiss rules from this tournament", what happens if a player challenges the pairings and it turns out that the arbiter had manually changed the pairings from what the fide rules would have given. Does the arbiter back down on the grounds that what he did wasn't sanctioned by the official rules or does he use the "you agree with the decision of the arbiter" clause on the entry form as a defence? And if it escalates to an appeal to the sanctioning body (State Association of ACF), what happens to (a) the player (in the case that they refuse to play the game with ammended pairings) and (b) the arbiter?

Feel free to supply offical/non official answers

Denis_Jessop
29-01-2006, 09:20 PM
The full text of what is says is as follow:


The critical wording by Reuben in all this would seem to be "It is essential to ensure that nobody is disadvantaged. The arbiter must ensure no bias."

But SR also says " However, it is illogical to say such actions are totally wrong" whereas the FIDE handbook says that if you do it you'll likely as not find yourself before the Ethics Commission and Gijssen calls it cheating. Perhaps, as Ian Rogers says, all is not happy at FIDE :hmm:

DJ

Denis_Jessop
29-01-2006, 09:27 PM
Another important question/scenario

Assuming that there is no special announcment at the start of the tournament (or on the entry form) along the lines of "We are dropping pairing rule N.x from the swiss rules from this tournament", what happens if a player challenges the pairings and it turns out that the arbiter had manually changed the pairings from what the fide rules would have given. Does the arbiter back down on the grounds that what he did wasn't sanctioned by the official rules or does he use the "you agree with the decision of the arbiter" clause on the entry form as a defence? And if it escalates to an appeal to the sanctioning body (State Association of ACF), what happens to (a) the player (in the case that they refuse to play the game with ammended pairings) and (b) the arbiter?

Feel free to supply offical/non official answers

I don't think I'll answer this in any capacity at this point :doh: but I would like to know how a provision on the entry form that you agree with the decision of the arbiter sits with the usual right of appeal from an arbiter's decision to an appeals committee.

DJ

Phil Bourke
29-01-2006, 10:55 PM
Another important question/scenario

Assuming that there is no special announcment at the start of the tournament (or on the entry form) along the lines of "We are dropping pairing rule N.x from the swiss rules from this tournament", what happens if a player challenges the pairings and it turns out that the arbiter had manually changed the pairings from what the fide rules would have given. Does the arbiter back down on the grounds that what he did wasn't sanctioned by the official rules or does he use the "you agree with the decision of the arbiter" clause on the entry form as a defence? And if it escalates to an appeal to the sanctioning body (State Association of ACF), what happens to (a) the player (in the case that they refuse to play the game with ammended pairings) and (b) the arbiter?

Feel free to supply offical/non official answers

Am I allowed to ask for one thing only, consistency :lol:
I really think that once arbiters and DOP's start tinkering for any reason, they are only leaving themselves open for it to come back and bite them.
Sure, they may do it here to help someone out, but eventually they will come across a situation where someone else will need that bit of assistance, and they will be unable to deliver it for whatever reason. Then the shit will hit the fan, as the aggrieved party will surely remind them of occassions where they have done it previously. :uhoh:
It is the same with any sporting official, don't try and deal with every situation as a onceoff, try and find the common ground and be consistent with your rulings and you will find that it won't be long before you have the respect of everyone. Because that is all that any competitor asks, for officials that are fair and consistent.

arosar
30-01-2006, 08:31 AM
There was yet another alteration to the pairings in the last yesterday of the ADW. George Xie wasn't very happy about it and so expressed his unhappiness to Dr Z.

AR

Rhubarb
31-01-2006, 02:07 AM
The only reason players are no longer allowed to write down their moves before they play it is that Gijssen lost an embarrassing dispute with a player a few years ago, found out that he was wrong, and determined to change the rules so that he could retrospectively be proved right.Off-topic, but one is immediately reminded of certain trans-Tasman arbiters who, to avoid the embarassment and attendant accusations of incompetence, sent an urgent missive to FIDE in an attempt to hide the fact that they weren't au fait with the latest FIDE rules for IM norms, and then blithely tried to claim that they were right all along when they got one of their mates to agree with them!

Huh?

Oepty
31-01-2006, 02:58 PM
I have been doing some manual pairing on a made up tournament to try and see whether I do understand the rules.
When applying C13 where the penultimate score bracket is one which has been paired first as a hetrogeneous score bracket and then a homogeneous remainder group is the whole hetrogeneous score bracket unpaired or is just the homogeneous remainder group unpaired?
I think it is the whole score bracket as this is what I think Swiss Perfect does but I am not entirely sure.

Thanks
Scott

Ian_Rogers
31-01-2006, 03:46 PM
Off-topic, but one is immediately reminded of certain trans-Tasman arbiters who, to avoid the embarassment and attendant accusations of incompetence, sent an urgent missive to FIDE in an attempt to hide the fact that they weren't au fait with the latest FIDE rules for IM norms, and then blithely tried to claim that they were right all along when they got one of their mates to agree with them!

Huh?

Gary Bekker is the only Trans-Tasman arbiter I know with mates in FIDE and this comment would seem to be seriously slanderous to him. Since Greg was in Queenstown and presumably knew what happened there, I suspect that someone else is making up stuff and posting as Greg - every single fact in the quoted post is wrong!

Ian

jase
31-01-2006, 03:56 PM
I had interpreted Greg's post as being of the same matter as your own, Ian:

http://www.chesschat.org/showpost.php?p=85787&postcount=11

I thought to ask - how do you know the intentions of the Titles Commision?

[If there's a Moderator about, Greg's, Ian's, and my post probably belong in the thread I've quoted from].

Rhubarb
31-01-2006, 04:41 PM
Good to see you've finally worked out how to use the quote function, Ian. Let's hope it's not just pure dumb luck, as in the case of antichrist.


Gary Bekker is the only Trans-Tasman arbiter I know with mates in FIDE ...I believe Gary Bekker was the only arbiter present who actually knew what they were talking about in this case so it couldn't have been him. :hmm: See link in Jase's post above.


... and this comment would seem to be seriously slanderous to him. In such a forum it would be called defamation, not slander. You'd think you'd know that by now, since you're a journalist. Perhaps we should get Sandler to buy you a dictionary with the <****deleted by lawyers***> he made in the settlement. Gijssen might be interested as well.


Since Greg was in Queenstown and presumably knew what happened there, ...I know that certain people were stating that certain people made, or had the possibility to make, IM norms, based on outdated information. Despite your disdain for the Iljumzhinov/Gijssen-run Rules Commission, the new regulations in this case are, in my view, a huge improvement. Why on earth should a player's own rating or title have anything to do with them getting a norm?


I suspect that someone else is making up stuff and posting as Greg - Yeah, nice one. I can assure you it's me.


... every single fact in the quoted post is wrong!How can facts be wrong?


IanI know.

Ian_Rogers
31-01-2006, 05:24 PM
[QUOTE=greg]
I know that certain people were stating that certain people made, or had the possibility to make, IM norms, based on outdated information. Despite your disdain for the Iljumzhinov/Gijssen-run Rules Commission, the new regulations in this case are, in my view, a huge improvement. Why on earth should a player's own rating or title have anything to do with them getting a norm?

/QUOTE]

You are confusing the Rules Commission - for which I certainly have disdain for its chairman - and the Titles Commission, which formulated the new regulations. They have been weak in combatting title corruption - they plead inadequate resources to conduct investigations - but otherwise seem to do their job responsibly.

I completely agree that it is not logical that your own rating and title should count towards a norm but they system was developed so that everyone in a round robin would be shooting for the same norm.

Anyway, if you talk to the players involved (Bjelobrk, Lukey) you will find that they knew the precise situation regarding the validity or otherwise of their norms before they played their final games. Perhaps the arbiters simply forgot to give you the same information.

Ian

P.S. Sandler should have received precisely $0 from his defamation case against me and the Sun-Herald, while his lawyer had some of his costs covered. They were the terms of the settlement.
Any money Sandler claims to have received from the case should be refunded to the Sun-Herald - but he only makes those claims to people who don't know the details of the case, like you. Sandler knows that if he actually admitted to getting any money (presumably from his lawyer), he would have to return it, if the Sun-Herald could be bothered chasing him for it.

Rhubarb
31-01-2006, 06:30 PM
You are confusing the Rules Commission - for which I certainly have disdain for its chairman - and the Titles Commission, which formulated the new regulations. I did know this, but yes I did not make that clear.


I completely agree that it is not logical that your own rating and title should count towards a norm but they system was developed so that everyone in a round robin would be shooting for the same norm.Right.


Anyway, if you talk to the players involved (Bjelobrk, Lukey) you will find that they knew the precise situation regarding the validity or otherwise of their norms before they played their final games. Perhaps the arbiters simply forgot to give you the same information.In more ways than one, there really wouldn't have been any need, would there? (At the risk of exposing myself to ridicule given my collapse in the last 4 rounds, I already understood when I was on 5/6 that there was no way I could get a norm as I had not played any titled players (besides myself :) ).)


P.S. Sandler should have received precisely $0 from his defamation case against me and the Sun-Herald, while his lawyer had some of his costs covered. They were the terms of the settlement.
Any money Sandler claims to have received from the case should be refunded to the Sun-Herald - but he only makes those claims to people who don't know the details of the case, like you. Sandler knows that if he actually admitted to getting any money (presumably from his lawyer), he would have to return it, if the Sun-Herald could be bothered chasing him for it.You are correct, I don't know any of the details. The only thing I thought I knew about the case was that the settlement terms were not to be disclosed.

Rhubarb
31-01-2006, 06:49 PM
Having barged onto this thread, I feel honour-bound to at least contribute one (vaguely) on-topic post. Unfortunately, I know nothing of the respective merits of the pairing programs, but everyone I've spoken to does not a have a high opinion of Protos, to say the least. This is unfortunate as FIDE are now insisting that FIDE-rated tournaments be submitted in Protos-compatible data.

Garvinator
31-01-2006, 07:01 PM
This is unfortunate as FIDE are now insisting that FIDE-rated tournaments be submitted in Protos-compatible data.
:wall: :wall: :wall: :wall: :evil:

Ian_Rogers
31-01-2006, 07:07 PM
You are correct, I don't know any of the details. The only thing I thought I knew about the case was that the settlement terms were not to be disclosed.
Correct, but apparently I am allowed to correct claims made by Sandler, who first breached that term of the settlement.
Incidentally, you wouldn't mind if I gave your name to the Sun-Herald as a witness as to the amount Sandler claimed to have received from the case, would you? The lawyers there were pretty appalled when informed of the amounts he was claiming to have received but need something more definite to follow up.

Ian

themovingman
02-02-2006, 11:10 AM
We had a lot of discussion about this in Brisbane. A copy of swiss master 5 is being purchased and it will be parallel run at Doeberl to examine differences.
It would be indeed useful to know through actual experience with the programs in question and you chaps should be commended, if found 'correct' then putting through the last few BIG tournaments and we'll see (possibly) who is better - SP; Protos; SM and if human-corrections are ultimately wise/correct or just 'doesn't look right - we can do better' type of thing.

bravo

arosar
13-02-2006, 02:33 PM
On the other hand where you have championships, norms, prizes or GP points at stake I think inserting sticky fingers into the draw is questionable, especially as it is necessarily arbitrary. It isn't meant to be a handicap or exhibition, the pairing rules are intended to provide a balance whereby in the end the contenders have met similar fields so that the difference in their results is due to their performances. Deliberately giving someone a harder game in the last round (= giving their opponent a harder game and someone else an easier game) is a dangerous business.

Well that's right!

Now, I only just read this thread in full in the last half hour. And I am reminded of something.

Was there a change to the pairings of boards 1 and 2 in the last round of the Aus Ch in Brissy? It maybe mentioned somewhere but can't find it.

Thanks,

AR

Bill Gletsos
13-02-2006, 02:41 PM
Well that's right!

Now, I only just read this thread in full in the last half hour. And I am reminded of something.

Was there a change to the pairings of boards 1 and 2 in the last round of the Aus Ch in Brissy? It maybe mentioned somewhere but can't find it.If you read this post in full then you should have been able to deduce the answer from post #27 of this thread.

arosar
13-02-2006, 02:51 PM
Ah, yes, totally missed that. Thanks Bill.

Now Bill, you are a very wise and learned person. What is your view on the change to boards 1 and 2 in round 11?

I could ask another question. Who requested that change?

AR

Bill Gletsos
13-02-2006, 03:30 PM
Ah, yes, totally missed that. Thanks Bill.

Now Bill, you are a very wise and learned person. What is your view on the change to boards 1 and 2 in round 11?I would have stuck with the SP pairings on boards 1-9.

I could ask another question. Who requested that change?Jase was an arbiter at the event. He should be able to answer this.

jase
13-02-2006, 04:27 PM
I would have stuck with the SP pairings on boards 1-9.
Jase was an arbiter at the event. He should be able to answer this.

Pairings were altered from those initially produced by SP for the final round after consultation among all arbiters. Charles Zworestine, after discussing the matter with Cathy Rogers (who was in the official's room in her role as Publicity Officer, and is also an IA), was vehement in his opinion that the pairings must be changed. Gary Bekker and Roly Eime were persuaded by his arguments.

As Gary was the Chief Arbiter, and a majority of the arbiters agreed with Charles, the published pairings were in accordance with Charles' opinions.

arosar
13-02-2006, 06:29 PM
OK...what was the basis of Charles' change request? And I thought it was already mentioned somewhere in this thread that Charles made fundamental errors in his change to round 7. How in the world did we permit him to make further changes in such a crucial round?

This guy seems to have a lot of power. He gets the pairings changed and he gets someone sacked!

Oh and one more thing. Was Cathy Rogers employed as arbiter or only as Press Officer? What was she and Charles doing discussing pairings?

AR

ursogr8
13-02-2006, 06:42 PM
Ah, yes, totally missed that. Thanks Bill.

Now Bill, you are a very wise and learned person. What is your view on the change to boards 1 and 2 in round 11?

I could ask another question. Who requested that change?

AR

hi AR

This comes as a bit of a surprise. I thought previous posts on IAs changing pairings were always self-initiated by the IA. That is, I understood various posters to say that it was common practice for the IA to examine the SP pairings and judge if they complied with the Dutch rules.
Have you new information that IAs also take requests on board?

regards
starter

ps
Nice article of yours on Queenstown, in Chess_whatever.
Well done, btw.

arosar
13-02-2006, 06:55 PM
It may be common practice. But the question we want answered here is what was the basis of that change for round 11 - specifically boards 1 and 2? Given that jase himself already wrote that Dr Z made fundamental errors in round 7, what was his argument for the round 11 change that so persuaded a couple of perfectly competent arbiters?

AR

jase
14-02-2006, 04:02 PM
That is, I understood various posters to say that it was common practice for the IA to examine the SP pairings and judge if they complied with the Dutch rules.

Some arbiters make changes to computerised pairings based on whether they comply with their own understanding of the intent of the swiss pairing system. The practice is not common by weight of numbers; it is "common" if one takes into account the volume of tournaments officiated by one particular arbiter.


Given that jase himself already wrote that Dr Z made fundamental errors in round 7, what was his argument for the round 11 change that so persuaded a couple of perfectly competent arbiters?
The pairing error made by Charles in altering the draw for Round 7 is a different issue to that debated for Round 11. In all diplomacy am I the appropriate person to answer your query, as the one arbiter who did not approve the modified draw?

themovingman
16-02-2006, 12:08 PM
at the time there were a few small grumbles about the Moylan-Bird pairing round 2( in that "that explains why XXYZ changes the computer draw"), there had been a few innocent/casual questions before (when my answer was no - not unless something were obviously askew).

I had a quick perusal, it seemed reasonable to me - what do others with more skill think ?

thanks

Round 1


No Name Loc Total Result Name Loc Total

1 Young, Jeffrey 1654 [0] .5:.5 Xie, George 2324 [0]
2 Rej, Tomek 2243 [0] 1:0 Canfell, Mike J 1648 [0]
3 Ramachandran, Raj 1602 [0] 0:1 Jones, Brian A 2151 [0]
4 Bird, Andrew 2105 [0] .5:.5 Phillips, Andrew 1564 [0]
5 Dickson, Ian C 1546 [0] 0:1 Charles, Gareth 2105 [0]
6 Bolens, Johny 2095 [0] 0:1 Parker, Trent 1490 [0]
7 Tracey, Michael J 1470 [0] .5:.5 Moylan, Laura A 2093 [0]
8 Broekhuyse, Paul 2026 [0] 1:0 Greenwood, Norman 1468 [0]
9 Smit, George 1464 [0] 0:1 Castor, David 2008 [0]
10 Zvedeniouk, Ilia 2001 [0] 1:0 Losh, Gary 1456 [0]
11 Berry-Porter, Zach 1447 [0] 0:1 Wright, Neil S 1945 [0]
12 Mendes da Costa, Alex 1930 [0] 1:0 Dorhauer, Peter K 1443 [0]
13 Whiteside, Benjamin 1366 [0] 0:1 Illingworth, Max 1786 [0]
14 Kordahi, Nicholas 1743 [0] 1:0 Deacon, Graeme 1360 [0]
15 Porter, Leslie 1049 [0] 0:1 Kodzoman, Ivan 1732 [0]
16 Hvistendahl, Robert 1701 [0] 1:0 Rajendran, Prem 679 [0]
17 Coyle, Mick 632 [0] 0:1 Rachmadi, Herman 1701 [0]
18 Escribano, Jose 1669 [0] 1:0 Chiddy, Rory 383 [0]
19 Mehen, Alex [0] 0:1 Boyce, Jamie 1665 [0]


Round 2



No Name Loc Total Result Name Loc Total

1 Illingworth, Max 1786 [1] 0:1 Rej, Tomek 2243 [1]
2 Jones, Brian A 2151 [1] 1:0 Kordahi, Nicholas 1743 [1]
3 Charles, Gareth 2105 [1] 1:0 Kodzoman, Ivan 1732 [1]
4 Rachmadi, Herman 1701 [1] 0:1 Broekhuyse, Paul 2026 [1]
5 Castor, David 2008 [1] 1:0 Hvistendahl, Robert 1701 [1]
6 Boyce, Jamie 1665 [1] 0:1 Zvedeniouk, Ilia 2001 [1]
7 Wright, Neil S 1945 [1] 1:0 Escribano, Jose 1669 [1]
8 Parker, Trent 1490 [1] 0:1 Mendes da Costa, Alex 1930 [1]
9 Xie, George 2324 [.5] 1:0 Tracey, Michael J 1470 [.5]
10 Moylan, Laura A 2093 [.5] 1:0 Bird, Andrew 2105 [.5]
11 Phillips, Andrew 1564 [.5] 1:0 Young, Jeffrey 1654 [.5]
12 Dorhauer, Peter K 1443 [0] 0:1 Bolens, Johny 2095 [0]
13 Canfell, Mike J 1648 [0] 1:0 Whiteside, Benjamin 1366 [0]
14 Deacon, Graeme 1360 [0] 0:1 Ramachandran, Raj 1602 [0]
15 Porter, Leslie 1049 [0] 0:1 Dickson, Ian C 1546 [0]
16 Greenwood, Norman 1468 [0] 1:0 Coyle, Mick 632 [0]
17 Rajendran, Prem 679 [0] 0:1 Smit, George 1464 [0]
18 Losh, Gary 1456 [0] 1:0 Mehen, Alex [0]
19 Chiddy, Rory 383 [0] 0:1 Berry-Porter, Zach 1447 [0]

pax
16-02-2006, 12:42 PM
Er, you just posted round 1 twice, so it's a bit hard to judge!!

PHAT
16-02-2006, 12:42 PM
After lighly skimming this tread, it has become obvious to me that The ACF should simply state as part of conditions for ratimng of games, that;

"For events using Swiss draws, and SwissPerfect for pairing purposes, pairings generated by the SP application may be changed only for accommodation of additions, requested byes, and withdrawals."

My arguments:

Consistancy. Transparency. Simplicity.

Anyone who needs me to justify the merrits of these, is a fool.

Furthermore, anyone wishing to allow DOPs to continue to fiddle with the draw must show, that fiddling can result in a benefit so great that it can trump: consistancy, transparency, and simplicity. If their cannot do it, they should STFU.


The ACF is asleep at the wheel, still. :wall:

themovingman
16-02-2006, 12:45 PM
Er, you just posted round 1 twice, so it's a bit hard to judge!!
aagh what a dolt, I did too - ta, I will try to edit above or re-post - ta

have fixed it up - board 10 is the pairing in question

pax
16-02-2006, 12:53 PM
However, it's not hard to see the issue:

The 0.5 group is as follows (rating order, colour preference in brackets):

Xie (w)
Bird (b)
Moylan (w)
Young (b)
Phillips (w)
Tracey(b)

Swiss Perfect correctly follows the FIDE pairing rules in giving colour preferences a higher priority than "top vs bottom". As a result, you get a pairing such as:

Xie-Tracey
Moylan-Bird
Phillips-Young

satisfying all colour preferences. Certain arbiters who prefer "top/bottom" to colour preference will pair something like:

Xie-Tracey
Young-Bird
Moylan-Phillips

resulting in minor colour breaches for two players. Note that this approach has consequences for round 3, since two players will have absolute colour preferences.

themovingman
16-02-2006, 08:45 PM
Thanks Pax - that was my thinking, as far as my thinking went, so (although I didn't consider affecting the following round)

so it is not a black eye for SP - in that instance anyway