PDA

View Full Version : Last round colour vs top/bottom (from Queenstown)



pax
24-01-2006, 08:23 AM
Rd 10 is now Froehlich-Smerdon and Hecht-Bjelobrk.


Interesting change.

The previous draw (Hecht-Smerdon, Bjelobrk-Froehlich) had all colour preferences satisfied in the 7 point group (once floaters were removed), but had 1v2 and 3v4. The new draw has Froehlich playing against his (strong) colour preference, and Bjelobrk against a (mild) colour preference but has 1v4 and 2v3.

So we have another top/bottom vs colour preference change.

Is the fact that it is the last round a factor here?

The FIDE rules say that the rule B2 (absolute colour preferences) is ignored in the last round for players over 50%. It does not say that B4 (other colour preferences) is ignored, but perhaps it is by implication since if you ignore absolute colour preference you must ignore weaker preferences?

Clearly the rules here are ambiguous. My interpretation would probably be to ignore B2 only where there is no substitute opponent with the same score (e.g if you have players on 8.5 and 8 ahead of everyone else then they will play regardless of colour preference). I would be inclined to otherwise enforce colour preferences within homogeneous groups where possible (including B4).

Garvinator
24-01-2006, 11:52 AM
Without looking at players histories etc, something to notice from the change is now Bjelobrk gets a GM instead of IM.

Was: Hecht-Smerdon, Bjelobrk-Froehlich
Is now: Froehlich-Smerdon and Hecht-Bjelobrk

Maybe the change was made to help get Igor his norm?

pax
24-01-2006, 12:04 PM
Without looking at players histories etc, something to notice from the change is now Bjelobrk gets a GM instead of IM.

Was: Hecht-Smerdon, Bjelobrk-Froehlich
Is now: Froehlich-Smerdon and Hecht-Bjelobrk

Maybe the change was made to help get Igor his norm?

a)It wouldn't make any difference to his norm, and
b)changing the draw in order to facilitate a norm in such a way would breach FIDE rules (and thus they may refuse to award the norm if that was the case).

pax
24-01-2006, 12:09 PM
But now that you mention it, the rating of the opponent may be significant.

The way the rating averages work, 6.5/9 is enough with Hecht as his last round opponent. With Froehlich as his last opponent, he would need 7/9 (i.e at least a draw in his last game).

Hmm - a happy coincidence perhaps??

Garvinator
24-01-2006, 12:22 PM
Hmm - a happy coincidence perhaps??
you be the judge :owned: :whistle: :uhoh:

Bill Gletsos
24-01-2006, 03:12 PM
Interesting change.

The previous draw (Hecht-Smerdon, Bjelobrk-Froehlich) had all colour preferences satisfied in the 7 point group (once floaters were removed), but had 1v2 and 3v4. The new draw has Froehlich playing against his (strong) colour preference, and Bjelobrk against a (mild) colour preference but has 1v4 and 2v3.

So we have another top/bottom vs colour preference change.

Is the fact that it is the last round a factor here?

The FIDE rules say that the rule B2 (absolute colour preferences) is ignored in the last round for players over 50%. It does not say that B4 (other colour preferences) is ignored, but perhaps it is by implication since if you ignore absolute colour preference you must ignore weaker preferences?No.
Ignoring B2 does not imply ignoring B4 etc.

Clearly the rules here are ambiguous. My interpretation would probably be to ignore B2 only where there is no substitute opponent with the same score (e.g if you have players on 8.5 and 8 ahead of everyone else then they will play regardless of colour preference). I would be inclined to otherwise enforce colour preferences within homogeneous groups where possible (including B4).As per your 8.5 V 8 above if their histories in the last two rounds were both WW or BB then pairing them would mean one of them gets 3 of the same colour in a row. In any round other than the last this would not be permitted. However the ability in the last round to be able to ignore B2 allows this pairing to occur.

Oepty
25-01-2006, 03:37 PM
But now that you mention it, the rating of the opponent may be significant.

The way the rating averages work, 6.5/9 is enough with Hecht as his last round opponent. With Froehlich as his last opponent, he would need 7/9 (i.e at least a draw in his last game).

Hmm - a happy coincidence perhaps??

This kind of thing is not needed. The person involved in suggesting in the change Charles Zworestine has very strong views about getting what he believes to be the correct pairings so it is absolutely believable he would of suggested the pairings be changed on that basis only. To say he did it because he wanted to help Bjelobrk get his IM norm is a slur on his character unless there is any other evidence to back up the assertion.
Scott

pax
25-01-2006, 03:46 PM
This kind of thing is not needed. The person involved in suggesting in the change Charles Zworestine has very strong views about getting what he believes to be the correct pairings so it is absolutely believable he would of suggested the pairings be changed on that basis only. To say he did it because he wanted to help Bjelobrk get his IM norm is a slur on his character unless there is any other evidence to back up the assertion.
Scott

I did not suggest that the change was made with the intention to assist Igor. I merely pointed out that it did. Others can speculate on the intentions.

Oepty
25-01-2006, 03:51 PM
I did not suggest that the change was made with the intention to assist Igor. I merely pointed out that it did. Others can speculate on the intentions.

Nonsense, the last line of your post clearly showed you doubted that it was a happy coincidence. It clearly implied that you thought there was chance it was made to help Bjelobrk. The idea is just shit.
Scott

auriga
25-01-2006, 04:11 PM
Nonsense, the last line of your post clearly showed you doubted that it was a happy coincidence. It clearly implied that you thought there was chance it was made to help Bjelobrk. The idea is just shit.
Scott

chill out freddy.

i didn't read it as such a serious comment.

see gg reply with all the smiley things.

the question is: why did they overide protos?
current consensus is they had some legit doubts on it's results.

Trizza
25-01-2006, 04:25 PM
This kind of thing is not needed. The person involved in suggesting in the change Charles Zworestine has very strong views about getting what he believes to be the correct pairings so it is absolutely believable he would of suggested the pairings be changed on that basis only. To say he did it because he wanted to help Bjelobrk get his IM norm is a slur on his character unless there is any other evidence to back up the assertion.
Scott

How can it be a slur on the character of Charles Z? Look at the timing of the posts. Pax's post which you refer to was made more than a day before Ian Rogers posted that the pairings were changed after Charles Z made his point. How can you slur someone's reputation if you don't refer to them in any way?

pax
25-01-2006, 04:32 PM
Nonsense, the last line of your post clearly showed you doubted that it was a happy coincidence.

I made no judgment one way or the other.

However, I happen to agree with Bill Gletsos' analysis that the initial pairing was correct, and the changed pairing was incorrect (according to the way I read the FIDE pairing rules, which may or may not be correct).

Ian_Rogers
26-01-2006, 07:05 PM
Throughout the tournament, the arbiters had been overriding Swiss Perfect to ensure that top half plays bottom half wherever possible.
The reason the Swiss Perfect pairings in the last round were not corrected until Charles intervened was a simple oversight. (You might have guessed why from the fact that two and three on the score group were one point apart on rating.)
The 'Dutch' pairing rules which were being followed for this tournament (with the announced exception of two floats in three rounds in the same direction being allowed) make perfectly clear that every combination of top half versus bottom half should be tried - the other system accepted by FIDE allows more exchanges for colour.

I think people are losing sight of the point of the Swiss system - pairing the most appropriate opponent for each player - preferably with the leading player meeting the strongest possible opponent. Colours are a peripheral matter - balancig colours is desirable but hardly the most important matter in deciding who wins a tournament.
Ian

pax
26-01-2006, 08:54 PM
The 'Dutch' pairing rules which were being followed for this tournament (with the announced exception of two floats in three rounds in the same direction being allowed) make perfectly clear that every combination of top half versus bottom half should be tried - the other system accepted by FIDE allows more exchanges for colour.


I disagree with this point.

The Dutch system as specified by FIDE do make it clear that every combination of top half vs bottom half must be tried first, but they also permit exchanges between top and bottom half in order to achieve colour balance.

The relevant parts of the procedures are as follows:



C.4
Put the highest players in S1, all other players in S2.

C.5
Order the players in S1 and S2 according to A2.

C.6
Pair the highest player of S1 against the highest one of S2, the second highest one of S1 against the second highest on e of S2, etc. If now p pairings are obtained in compliance with B1 and B2 the pairing of this score bracket is considered complete.

* in case of a homogeneous score bracket: remaining players are moved down to the next score bracket. With this score bracket restart at C1.
* in case of a heterogeneous score bracket: only players moved down were paired so far. Start at C2 with the homogeneous remainder group.

C.7
Apply a new transposition of S2 according to D1 and restart at C6.

C.8
In case of a homogeneous (remainder) group: apply a new exchange between S1 and S2 according to D2. Restart at C5.

C.9
Drop criterion B6 and B5 (in this order) for downfloats and restart at C4.

C.7 pertains to transpositions (within one top/bottom arrangement), while C.8 pertains to exchanges between top and bottom. Both of these occur before you start ignoring the prioritised pairing criteria at C.9 (B4 pertains to strong and mild colour preference).

Note that in C6, "B1 and B2" is generally understood to mean "B1 to B6". If you don't read it that way, the rules make no sense at all.

jase
27-01-2006, 12:53 PM
Throughout the tournament, the arbiters had been overriding Swiss Perfect to ensure that top half plays bottom half wherever possible ...
The 'Dutch' pairing rules which were being followed for this tournament (with the announced exception of two floats in three rounds in the same direction being allowed) make perfectly clear that every combination of top half versus bottom half should be tried - the other system accepted by FIDE allows more exchanges for colour.

Ian's comments are at odds with the 'Dutch' pairing rules, and indicate that these rules were not being followed in Queenstown. The 'Dutch' system promotes a far greater set of exchanges for colour than "the other system" [the Lim system, previously in common use, taught to me by Cathy Rogers and Peter Parr. As per my post on another thread, it does have greater respect for top v bottom, and is therefore in my view a better system].

Some time ago Gary Bekker sought clarification from Geurt Gijssen, author of the 'Dutch' system, on this very point: [you can read his full reply here - http://www.chesscafe.com/text/geurt38.pdf]


Mr. Bekker objects to the fact that due to colours, there is an exchange between the first half, called S1, and second half of the group, called S2. In our example #1, #2 and #3 are in S1, #4, #5 and #6 are in S2. Well, I am not against these exchanges. I am sure that players much prefer the alternating or equalising of colours; they are especially unhappy when they have two blacks in a row. I will remember forever the face of the late GM Bagirov when this happened to him. He never complained verbally, but his face said enough. I would like to add something else: the ratings are used to make the pairings, but they are only used to be sure that the pairings are not made at random and that players can check them. Last remark: the players in a score group, this is a group of players with the same score, have proved that they have in the actual tournament more or less the same strength. Why not to pair them among each other?

It is therefore not correct to state that "colours are a peripheral matter" for tournaments using the 'Dutch' pairing system.


I think people are losing sight of the point of the Swiss system - pairing the most appropriate opponent for each player
I don't think this is the case. There are two debates here, which you are interchanging. One debate is the correct application of the 'Dutch' pairing system; the other is that system's merits.

Bill Gletsos
27-01-2006, 01:36 PM
I don't think this is the case. There are two debates here, which you are interchanging. One debate is the correct application of the 'Dutch' pairing system; the other is that system's merits.Exactly.
It appears arbiters are not following the 'Dutch' pairing system when making manual pairings, because they either dont agree with aspects of it or are misapplying it.
This should not be the case.
If arbiters are going to pair using the Dutch system then pair using the Dutch system by applying it correctly and not just using parts of the rules they personally agree with and disregarding the ones they dont.
FIDE expect that all arbiters following the pairing rules should get the same pairings.
One could assume that arbiters getting pairings that differ from Swiss Master which is the official program of the Dutch Chess Federation have simply gotten the pairings wrong either through ignorance of the rules or misapplication.

shaun
27-01-2006, 02:00 PM
I assume the in the case of NSWCA events, that the NSWCA will now instruct their arbiters to follow the FIDE swiss pairing rules.
It is certainly something that I will reccomend to the ACTCA.

Bill Gletsos
27-01-2006, 02:13 PM
I assume the in the case of NSWCA events, that the NSWCA will now instruct their arbiters to follow the FIDE swiss pairing rules.
It is certainly something that I will reccomend to the ACTCA.That would need to be a Council decision.
I will raise it for discussion at the next NSWCA Council meeting.

jenni
27-01-2006, 03:15 PM
Can someone just check if my understanding is correct.

1. there are numerous ways of running tournaments - swiss, round robin, knock out etc,

2. If you run a swiss then there are various pairing systems you can use Lottery, Seeded, Burstein etc

3. Of the Seeded Swiss pairing systems, two are recognised by FIDE and these are Lim and Dutch. Dutch is the one that seems to be computerised.

Presumably before the event, it can be publicised as to what sort of tournament you are running. Also you can announce what pairing system you are using. ie. If I am running a tournament at Belco I can choose to use the Lim rules (if I am prepared to manually pair). Also we can decide to use pretty much any format we like and not just a seeded swiss.

Presumably if you want to do a seeded swiss and want it FIDE rated you are obliged to use either Lim or Dutch?

Is it correct that if you are running a seeded swiss and it is a FIDE rated tournament and you then overide the Dutch rules to give preference to Lim rules, then FIDE might refuse to rate it? I understand this has been happening in Norway, where FIDE is not rating their tournaments because of manual changes to the Dutch system?

Sorry there are so many questions, but I am trying to get my head around this - I might be a parent, but do run tournaments!

jase
27-01-2006, 03:30 PM
Hi Jenni,

Your understanding is correct. Additionally that there is a 3rd Swiss system recognised by FIDE, which is untried here in Australia. It is called the DUBOV system.


Is it correct that if you are running a seeded swiss and it is a FIDE rated tournament and you then overide the Dutch rules to give preference to Lim rules, then FIDE might refuse to rate it? I understand this has been happening in Norway, where FIDE is not rating their tournaments because of manual changes to the Dutch system?

There should be no problem with manual intervention of a pairing program, so long as that intervention is appropriate. An example of an inappropriate intervention would be manipulation of the pairings to assist a player to achieve a norm.

FIDE makes clear that it is the arbiter who makes the pairings, and computerised pairing programs are simply a tool to be used for this purpose.

I am not familiar with the particular instances in Norway to which you refer, and could not unearth anything in a few online searches.

jenni
27-01-2006, 03:34 PM
Hi Jenni,

Your understanding is correct. Additionally that there is a 3rd Swiss system recognised by FIDE, which is untried here in Australia. It is called the DUBOV system.


Thank-you - I am on a big learning curve at the moment.




FIDE makes clear that it is the arbiter who makes the pairings, and computerised pairing programs are a tool used be the arbiter/s in making pairings. I am not familiar with the particular instances in Norway to which you refer, and could not unearth anything in a few online searches.

There were some overseas players in Queenstown who told Tony that - I also couldn't find any info on the net, but it concerned me that it might parallel our situation.

Bill Gletsos
27-01-2006, 03:44 PM
There should be no problem with manual intervention of a pairing program, so long as that intervention is appropriate.I think you have to be extremely careful by just saying "is appropriate".

As Geurt stated in one of his columns:

...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.

The bolding is mine.

I doubt any of those reasons exist in Australia.

EGOR
27-01-2006, 03:49 PM
I think you have to be extremely careful by just saying "is appropriate".

As Geurt stated in one of his columns:
...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.

I doubt any of those reasons exist in Australia.

It's my understanding that in the Interleages tournament (Sydney, NSW) players from the same club are not supposed to play each other.

jase
27-01-2006, 03:59 PM
It is the controller who makes the pairings. The computer is just a tool which he may use

Supporting this state of play is the following analogous regulation:



Compilation of pairings with the aid of a computer is permissible. If, however, a discrepancy arises between the computer`s pairing and those worked out personally by the Pairings Committee, the final decision shall lie with the committee chairman.

Bill Gletsos
27-01-2006, 05:16 PM
It is the controller who makes the pairings. The computer is just a tool which he may use

Supporting this state of play is the following analogous regulation:



Compilation of pairings with the aid of a computer is permissible. If, however, a discrepancy arises between the computer`s pairing and those worked out personally by the Pairings Committee, the final decision shall lie with the committee chairman. Yes clearly it is the arbiter who makes the pairings and the computer is just a tool.
However I dont interpret either statement to imply that the arbiter can disregard the applicaple pairing rules at his whim.

If the arbiter is going to override the computer and it isnt one of the reason's Geurt stated then it should only be because the computer program has clearly got the pairing wrong.

eclectic
27-01-2006, 05:39 PM
i note with interest guert's contention that pairings may sometimes have to be changed for political reasons ...

a direct and flagrant violation of the fide motto

gens una sumus

Trent Parker
27-01-2006, 10:29 PM
i note with interest guert's contention that pairings may sometimes have to be changed for political reasons ...

a direct and flagrant violation of the fide motto

gens una sumus

which means?....:uhoh:

jase
27-01-2006, 10:47 PM
Gens una sumus = We are one family.

It's FIDE's motto. Did you look it up first?

Denis_Jessop
28-01-2006, 10:43 AM
Can someone just check if my understanding is correct.

1. there are numerous ways of running tournaments - swiss, round robin, knock out etc,

2. If you run a swiss then there are various pairing systems you can use Lottery, Seeded, Burstein etc

3. Of the Seeded Swiss pairing systems, two are recognised by FIDE and these are Lim and Dutch. Dutch is the one that seems to be computerised.

<snip>

Presumably if you want to do a seeded swiss and want it FIDE rated you are obliged to use either Lim or Dutch?

Is it correct that if you are running a seeded swiss and it is a FIDE rated tournament and you then overide the Dutch rules to give preference to Lim rules, then FIDE might refuse to rate it? I understand this has been happening in Norway, where FIDE is not rating their tournaments because of manual changes to the Dutch system?

<snip>


FIDE deals with this issue in part 04 of the Handbook. It describes 3 systems - the so-called Dutch based on rating, the FIDE Swiss Rules (?Lim) and the Dubov based on rating but in para 04.2 D19 it says

"As directed by the Kishinev Congress 1997, below is the list of endorsed programmes and their respective capabilities:

a.PETUNIA Dutch System

b.GMB Lim System

c.SWISS CHESS Dutch System

d.SVBOSS Dutch System

e.DUBOV

f.BURSTEIN System (GA '98)"

I note that "their respective capabilities" are not there.

It goes on to say, in para 20:

"20. Mandatory information to be provided in tournament's reports by arbiters

20.1 While reporting a tournament to FIDE, the Arbiter shall declare which of the official FIDE Swiss Systems was used. (GA '97).

20.2 If another system was used, the Arbiter has to submit the rules of this system for checking by the Swiss Pairing Committee (GA '97)

20.3 Where it can be shown that modifications of the original pairings were made in favour of a player to achieve a norm, a report may be submitted to the Qualification Commissionto initiate disciplinary measures through the Ethics Commission (GA '97)"

These passages seem to make it clear that quite a few swiss pairing systems are approved and others can be use subject to scrutiny. It also makes it clear that changing of approved rules to allow a norm can result in disciplinary action.

It does appear that the Dutch, Lim and Dubov are the most common as the minutes of the Swiss Pairing Committee at the 73rd FIDE Congress state that, of the attendees, 10 used the Dutch, though some of them also used Lim as an alternative, and 3 used Dubov.

DJ

Bill Gletsos
28-01-2006, 12:27 PM
"As directed by the Kishinev Congress 1997, below is the list of endorsed programmes and their respective capabilities:

a.PETUNIA Dutch System

b.GMB Lim System

c.SWISS CHESS Dutch System

d.SVBOSS Dutch System

e.DUBOV

f.BURSTEIN System (GA '98)"There have been updates to that at the 74th, 75th and 76th Congress meetings of the Swiss Pairings Committee.

The following was passed at the 74th in October 2003 in Greece:

The up to now used unlimited endorsement of Swiss Pairing Programs has the disadvantage that FIDE has no influence to get programmers to implement changes of FIDE rules into their programs.
Therefore the Committee agreed unanimously to the proposal that all programmers are called to renew the endorsement of their programs in 2004 and any of these new endorsements shall be valid for 4 years.
For endorsement 3 copies of a program shall be delivered to the chairman of the Swiss Pairing Committee who will distribute the copies to Committee members for examination.
This was passed at the 75th in October 2004 in Spain:

1. Endorsement of submitted programs.
Due to the timetable of the congress and overlapping of meetings there were only two members of the committee present. Therefore the following decisions were taken at the meeting of the Technical Commission:
There are three Swiss programs unanimously proposed for endorsement by the General Assembly:
- SWISS Master 5.0 from Netherlands (a progressed version of the SWISS Master 4.8)
- SWISS Manager from Austria
- PROTOS
The other two programs formerly endorsed (of Mr.Weber (GER) and of Mr.Bordonada(PHI)) are invited to submit their application to the next year.
This was passed at the 76th in August 2005 in Germany:

1. Endorsement of newly submitted Swiss Pairing programs.
1.1. Program WINSWISS ( Dutch System) programmed by Franz-Josef Weber. Messrs. Stubenvoll, Markkula and Krause reported that they had checked the new version of the program and have not found any problems which might prevent the endorsement of the program. The proposal to endorse the program was unanimously accepted.
1.2. Program VEGA (Dubov System) programmed by Luigi Foriano. Messrs. Stubenvoll, Markkula and Krause reported that they had checked the program to the result that it looks to be a promising new program using the Dubov System. To investigate the program more deeply the time until end of September 2005 is needed. The committee accepted unanimously the proposal to authorize the subcommittee Mr. Krause, Mr. Markkula, Mr. Stubenvoll to endorse the program if this deeper investigation verifies the quality of the program.

Denis_Jessop
28-01-2006, 08:24 PM
This is interesting, Bill, as I assumed what I quoted was referring to rules for swiss pairings, not to computer pairing programs. I can't see anything in the relevant bit of the Handbook that refers to computer programs. If you are right, this is another case of the FIDE Handbook being out of date as well as downright confusing.

Incidentally, I note that the Swiss Pairings Committee, at the 74th Congress (2003), looked at the Crouch pairing system and at drafting a general set of pairing rules with the following results:

"7. Crouch Pairing System
The Committee considered the Crouch Pairing System. They found it extremely original and interesting. However, they felt it would never be applied widely, and thus would have limited value.
8. General rules for swiss system tournaments
The Committee agreed that besides the given rules of different Swiss Systems a set of general rules valid for every Swiss System tournament shall be specified and listed. Some of this rules are mentioned in the description of the Dutch Swiss System (e.g.: in principle never change a published pairing; players who are absent for a round without notifying the arbiter will be considered to have withdrawn). These rules will be a part of the tournament rules."

It will be interesting to see what effect this will have on SP, in particular, and the other programs generally.

DJ