Swiss Pairs mis-matches

[gordontd]

Here in the EBU we have regulations to deal with the situation where a pair has been assigned to the wrong opponents, usually as a consequence of them having entered their scores incorrectly, or of having played the match in the wrong polarity, and it not having been noticed in time. These regulations, contained in section 3.5 of our White Book, aim to compensate a pair who, through no fault of their own, play a match against a higher-ranking pair than they should, as well as to remove some benefit from a pair at fault who play a match against a lower-ranking pair.

 

These regulations work fine, but they are based on match-points being converted to Victory Points, our standard method here. Recently we have run a Swiss Pairs, and intend to repeat it, that is scored by match-points and not converted. What I wonder is whether Norway (or indeed anywhere else), as somewhere with lots of experience of running events of this type, has any regulations or methods for dealing with mis-matches?

[Vampyr]

Recently we have run a Swiss Pairs, and intend to repeat it, that is scored by match-points and not converted.

 

Please don't repeat it very often.

[pran]

Here in the EBU we have regulations to deal with the situation where a pair has been assigned to the wrong opponents, usually as a consequence of them having entered their scores incorrectly, or of having played the match in the wrong polarity, and it not having been noticed in time. These regulations, contained in section 3.5 of our White Book, aim to compensate a pair who, through no fault of their own, play a match against a higher-ranking pair than they should, as well as to remove some benefit from a pair at fault who play a match against a lower-ranking pair.

 

These regulations work fine, but they are based on match-points being converted to Victory Points, our standard method here. Recently we have run a Swiss Pairs, and intend to repeat it, that is scored by match-points and not converted. What I wonder is whether Norway (or indeed anywhere else), as somewhere with lots of experience of running events of this type, has any regulations or methods for dealing with mis-matches?

No, we don't bother.

 

Our experience is that some few "mismatches" will not have any significant impact on the final ranking.

 

The top "N" pairs where "N" is the number of rounds played (and not exceeding 40% of the number of pairs) will usually be "correctly" ranked in the end regardless of such mismatches.

[mycroft]

I have to admit, I've never understood "MPs converted to VPs" for Swiss Pairs. We don't do that for a straight matchpoint night, and unlike IMPs (where we are concerned about flattening out the huge swings), matchpoints are "same top on every board".

 

What is the benefit of limiting blitzing in 6 board rounds that is different from the 2, 3, or 4 board rounds?

[Vampyr]

I have to admit, I've never understood "MPs converted to VPs" for Swiss Pairs. We don't do that for a straight matchpoint night, and unlike IMPs (where we are concerned about flattening out the huge swings), matchpoints are "same top on every board".

 

What is the benefit of limiting blitzing in 6 board rounds that is different from the 2, 3, or 4 board rounds?

 

Not converting to VPs removes the whole "match" element of the event. In a match the boards are finished and you start afresh. If you just use matchpoints, one or two very good or very bad rounds could determine where you finish in the event.

 

Not too dissimilar to straight matchpoint events (especially as, if you are comparing to Swiss Pairs, you are talking about 7 or 8 boards per match). Why have two different events be so similar? There are plenty of each on the calendar.

 

An otherwise ordinary pairs event where each round you are matched up with opponents on a similar score score seems unappealing to me. Why is this being tried in the EBU, Gordon?

[Finch]

Please don't repeat it very often.

 

Please do. I much prefer it (with or without a cap on the number of MPs you can win in one match).

[Vampyr]

Please do. I much prefer it (with or without a cap on the number of MPs you can win in one match).

 

Why do you prefer it, Frances?

[gordontd]

Not converting to VPs removes the whole "match" element of the event. In a match the boards are finished and you start afresh. If you just use matchpoints, one or two very good or very bad rounds could determine where you finish in the event.

That's why it's better to use it when the rounds are shorter and the event is longer. The effect is that each board has equal value.

 

Not too dissimilar to straight matchpoint events (especially as, if you are comparing to Swiss Pairs, you are talking about 7 or 8 boards per match).

That doesn't follow at all. The event we ran had 12x4-board rounds.

 

Why have two different events be so similar? There are plenty of each on the calendar.

 

An otherwise ordinary pairs event where each round you are matched up with opponents on a similar score score seems unappealing to me. Why is this being tried in the EBU, Gordon?

It was used as the qualifier for the Two-Stars Pairs at the Autumn Congress. The existing formula of two sessions of match-pointed pairs no longer seemed very exciting and gave those who were doing badly little to play for. Since we give master points for matches won in Swiss Pairs, everyone has something to play for, even if they have been doing badly. Swiss Pairs gives some of the advantages of barometer scoring without the limitations of all-play-all.

[pran]

Not converting to VPs removes the whole "match" element of the event. In a match the boards are finished and you start afresh. If you just use matchpoints, one or two very good or very bad rounds could determine where you finish in the event.

 

Not too dissimilar to straight matchpoint events (especially as, if you are comparing to Swiss Pairs, you are talking about 7 or 8 boards per match). Why have two different events be so similar? There are plenty of each on the calendar.

 

An otherwise ordinary pairs event where each round you are matched up with opponents on a similar score score seems unappealing to me. Why is this being tried in the EBU, Gordon?

What is the "match" element in a Swiss for Pairs? (Or for that sake in any barometer style event for pairs)

[gordontd]

What is the "match" element in a Swiss for Pairs? (Or for that sake in any barometer style event for pairs)

It is that, when scores are converted to Victory Points, your actions towards then end of a match will be affected by how you think you have done earlier in the match.

[campboy]

I have to admit, I've never understood "MPs converted to VPs" for Swiss Pairs. We don't do that for a straight matchpoint night, and unlike IMPs (where we are concerned about flattening out the huge swings), matchpoints are "same top on every board".

 

What is the benefit of limiting blitzing in 6 board rounds that is different from the 2, 3, or 4 board rounds?

The benefit of limiting blitzing is that we don't want how much the top pairs beat weak pairs by to be a major factor in the final ranking of the top pairs. This is especially important if they will have played different weak pairs: how much we beat weak pair A by is at least a fair measure of something, but whether we beat weak pair A by more or less than you beat weak pair B by probably tells us more about the relative strengths of A and B than it does about ours.

 

In Swiss, pairs in contention to win won't be playing more than one or two weak pairs, and which weak pairs (if any) they play has a large amount of essentially random variability, most of which comes from the round 1 draw. In a normal pairs night, especially with a Howell or 3/4 Howell, this is less of an issue because you are much closer to comparing like with like.

 

[edited to change "teams" to "pairs" :)]

[pran]

It is that, when scores are converted to Victory Points, your actions towards then end of a match will be affected by how you think you have done earlier in the match.

Am I right then that this has nothing to do with calculating MatchPoints from comparisons of scores or converting Matchpoints further on to Victory Points?

 

It is about the players being kept continuously updated on how they are currently ranked in the event. (A most valued feature of barometer type events.)

[gordontd]

Am I right then that this has nothing to do with calculating MatchPoints from comparisons of scores or converting Matchpoints further on to Victory Points?

 

It is about the players being kept continuously updated on how they are currently ranked in the event. (A most valued feature of barometer type events.)

If I understand what you are saying here, I don't think you understood what I was saying :)

 

If you play an eight-board match converted to VPs, if you think you have done very well on the first six boards you should play safe on the remaining two. If you think you have done very badly on the first six boards, you should take risks on the final two. Not all boards have equal value in a match.

[StevenG]

I don't really see the match element kicking in - after all, you don't get a bonus for winning the match (discounting master points awards), as the score just goes 9-11, 10-10, 11-9 etc.. If winning the match was important, why aren't positions determined by matches won, with VPs (or MPs) as the tiebreaker?

[campboy]

I don't really see the match element kicking in - after all, you don't get a bonus for winning the match (discounting master points awards), as the score just goes 9-11, 10-10, 11-9 etc.. If winning the match was important, why aren't positions determined by matches won, with VPs (or MPs) as the tiebreaker?

It's not just about winning the match, it's because the VP scale isn't linear. If you're a fair way behind, an average on the last board may get you 3VPs, a top 8VPs and a bottom 1VP. So the odds favour going for a top -- more than they would normally at any rate -- because you stand to gain much more than you stand to lose.

[Vampyr]

If I understand what you are saying here, I don't think you understood what I was saying :)

 

If you play an eight-board match converted to VPs, if you think you have done very well on the first six boards you should play safe on the remaining two. If you think you have done very badly on the first six boards, you should take risks on the final two. Not all boards have equal value in a match.

 

There is that, but also there is the fact that boards in past matches are finished and no longer have an effect on your score. Thiis is important always, but especially in the first match when the pairs may be mis-matched.

 

I appreciate the difference with a shorter match though. I don't mind if this other format replaces ordinary pairs and not Swiss Pairs!

[pran]

I am confused and cannot see that any of this discussion has any relevance at all in Swiss events for pairs, at least not the way we play Swiss pairs in Norway:

 

Immediately after each round (typically 3 or 4 Boards played simultaneously at all tables) each board in that round is scored out, resulting in a Match Point score in the range 0% - 100%.

 

The MP scores for each pair (accumulated over all the rounds played so far) is used when drawing the seating for the next, or more commonly the round following the next.

 

So for instance as soon as round 3 has been completed all pairs shift for round 4, and early during round 4 each pair receives a report with their results in round 3, their total number of matchpoints made so far, and instruction on where they shall be seated in round 5.

 

Is my confusion justified?

[Vampyr]

I am confused and cannot see that any of this discussion has any relevance at all in Swiss events for pairs, at least not the way we play Swiss pairs in Norway:

 

....

 

Is my confusion justified?

 

Yes, it is, if you thought the discussion was about the way things are done in Norway.

[gordontd]

There is that, but also there is the fact that boards in past matches are finished and no longer have an effect on your score.

I'm not surprised Sven is confused! What on earth do you mean by this?

[Vampyr]

I'm not surprised Sven is confused! What on earth do you mean by this?

 

Well, I mean that you get your VPs and the boards no longer have an effect on your scores, for example, you can get a bottom on a board and still have a big win, even 20-0. You can have two bottoms and a big win. Or a tie or whatever. Without VPs, those 0% boards will drag your score down for the rest of the event.

[pran]

Well, I mean that you get your VPs and the boards no longer have an effect on your scores, for example, you can get a bottom on a board and still have a big win, even 20-0. You can have two bottoms and a big win. Or a tie or whatever. Without VPs, those 0% boards will drag your score down for the rest of the event.

 

Do you play Swiss pairs considering each round a separate and complete tournament on its own?

 

That is:

Here is the prize list for round 1

Here is the prize list for round 2

Here is the prize list for round 3

Here is the prize list for round 4

.....

And there is no prize list for the entire event consisting of say 15 rounds?

 

That is the only way I can see how the results in round 1, 2, 3 and so on have no effect on later rounds.

 

I am still confused.

 

PS.: And what scores are the base for assigning seats in for instance round 15?

[Mbodell]

Do you play Swiss pairs considering each round a separate and complete tournament on its own?

 

That is:

Here is the prize list for round 1

Here is the prize list for round 2

Here is the prize list for round 3

Here is the prize list for round 4

.....

And there is no prize list for the entire event consisting of say 15 rounds?

 

That is the only way I can see how the results in round 1, 2, 3 and so on have no effect on later rounds.

 

I am still confused.

 

PS.: And what scores are the base for assigning seats in for instance round 15?

 

I don't see how this can be confusing. You play round 1. Everyone has MP scores for that round. You have a mapping from various MP cutoffs to various VP awards (for simplicity lets say 70+% => 20 VP; 66-70% => 19 VP; 62-66% => 18VP; 60-62% => 17 VP; 58-60% => 16 VP; 56-58% => 15 VP; 54-56% => 14 VP; 53-54% => 13 VP; 52-53% => 12 VP; 51-52% => 11 VP; 49-51% => 10 VP; and so on for the losing VP percentages - I made these up for illustrative purposes, there is a real scale depending on number of boards in the various CoC and regulations). Now you pair people based on their VP totals. In round 2, you score just the round 2 boards by MP. You then convert round 2's MP score to VP again. The MP in round 1 have no direct effect on the MP in round 2 or the VP earned in round 2 (but do effect the pairing - assuming you are doing current round pairings and not r-1 pairings). So just like you often pair swiss teams by VP totals, not IMP totals, and the IMP=>VP earned in round X is calculated without any reference to rounds X-1 and behind, so it is the same in swiss pairs with match scoring. At the end of the night the winning pair is the one with the most VP.

 

It makes much more sense to me to do the match scoring and VP than to just do raw match points through out the night.

[blackshoe]

It makes much more sense to me to do the match scoring and VP than to just do raw match points through out the night.

Are you saying that it makes more sense to VP each "match" as it happens rather than to VP at the end of the session, or are you saying that it makes more sense to VP the matches rather than just score on the basis of raw match points?

 

Either way, why?

[pran]

I don't see how this can be confusing. You play round 1. Everyone has MP scores for that round. You have a mapping from various MP cutoffs to various VP awards (for simplicity lets say 70+% => 20 VP; 66-70% => 19 VP; 62-66% => 18VP; 60-62% => 17 VP; 58-60% => 16 VP; 56-58% => 15 VP; 54-56% => 14 VP; 53-54% => 13 VP; 52-53% => 12 VP; 51-52% => 11 VP; 49-51% => 10 VP; and so on for the losing VP percentages - I made these up for illustrative purposes, there is a real scale depending on number of boards in the various CoC and regulations). Now you pair people based on their VP totals. In round 2, you score just the round 2 boards by MP. You then convert round 2's MP score to VP again. The MP in round 1 have no direct effect on the MP in round 2 or the VP earned in round 2 (but do effect the pairing - assuming you are doing current round pairings and not r-1 pairings). So just like you often pair swiss teams by VP totals, not IMP totals, and the IMP=>VP earned in round X is calculated without any reference to rounds X-1 and behind, so it is the same in swiss pairs with match scoring. At the end of the night the winning pair is the one with the most VP.

 

It makes much more sense to me to do the match scoring and VP than to just do raw match points through out the night.

I am at a loss on how anybody can state that the results in rounds 1, 2, 3 and so on has no impact on later rounds if Match Point scores were converted to VP scores so long as the outcome of the event is decided by the accumulated sum of whatever kind of scores won in all rounds during the entire event.

 

Anyway, our experience is that the final results based on Match Points calculated as specified in Law 78A extremely seldom extend outside the range 40% - 60%. So the alleged advantage by converting Match Points to Victory Points seems rather obscure to me.

[Vampyr]

Are you saying that it makes more sense to VP each "match" as it happens rather than to VP at the end of the session, or are you saying that it makes more sense to VP the matches rather than just score on the basis of raw match points?

 

Either way, why?

 

I would say both, and the reason for me is that it makes a more enjoyable event.