[Vampyr]

Hi. I am looking for sensible ideas. In a multiple teams event with two- or 3-board rounds, is scoring by VPs or total IMPs better? If the latter, how many boards would it take for VPs to be better?

I am not really sure what I mean by "better" LOL, but perhaps people who use each method could explain why they have made this choice?

Thanks

[gordontd]

The White Book says:

Quote

The use of VPs in matches of fewer than five boards is not recommended. In EBU competitions,
events such as this are normally scored by straight IMPs, with no upper or lower limit.

[blackshoe]

Yes, Gordon, but why does it say that? What's the rationale?

[gordontd]

blackshoe, on 2017-April-03, 08:06, said:

Yes, Gordon, but why does it say that? What's the rationale?

It's been like that as long as I can remember and I had nothing to do with it, but my reason would be that you only need to reduce the effect of one extreme match if that match is a significant part of the whole event. Two boards out of 24 or four out of 48 are not alone going to decide the overall event on the basis of a mis-match of abilities.

[nige1]

It seems better if each board has about the same weight, so that no single match or board has a disproportionate effect on the result. Hence

MPs > BAM > VPs > IMPs > aggregate

I've always wondered why match-points aren't used for multiple-teams. You would score each board at match-points over the field. The score for each team would be the sum of the MP scores for the pairs in the team.

You might also ensure that winning a match is more important than by how much. e.g. award a large winning bonus.

[Vampyr]

nige1, on 2017-April-04, 15:20, said:

I've always wondered why match-points aren't used for multiple-teams. You would score each board at match-points over the field. The score for each team would be the sum of the MP scores for the pairs in the team.

You could do this, of course. I am guessing that the idea doesn't hold the same appeal to other people, because you don't see this happening. Also I think it requires a very complicated strategy for each pair's approach in bidding and play. Also most teams would have a score near 50%, which would be kind of boring.

Quote

You might also ensure that winning a match is more important than by how much. e.g. award a large winning bonus.

Would this make sense when most of the winning margins will be very small?

[Jeremy69A]

No idea why we ever use victory points. It seems reasonable to have a cut off so you lose whatever you lose but you can't win more than whatever constituted 20-0.

[Cyberyeti]

I'd have said fewer than 4 boards rather than 5 as the limit, if you get the right boards against the right team, you can easily put 50 on in 4 boards which can seriously distort a competition, way more than a 20-0, if you limited it to 30-odd IMPs (whatever 20-0 is), that would be fine.

[PeterAlan]

nige1, on 2017-April-04, 15:20, said:

It seems better if each board has about the same weight, so that no single match or board has a disproportionate effect on the result.

Hence MPs > BAM > VPs > IMPs > aggregate.

Following this to its logical conclusion within the VP / IMP dichotomy, one wouldn't VP each 2- or 3-board match; instead one would VP each board within a match. But then one wouldn't need the IMP scale at all, just a revised VP one (aggregate difference -> IMP -> VP can trivially be compounded to get aggregate difference -> VP). The question then becomes: what's the appropriate degree of smoothing of (aggregate difference) results that the VP scale should be targeting, because it's not at all obvious that it's the one we have now by following that compounding process. And we might then end up back at the old IMP scale.

nige1, on 2017-April-04, 15:20, said:

I've always wondered why match-points aren't used for multiple-teams. You would score each board at match-points over the field. The score for each team would be the sum of the MP scores for the pairs in the team.

The essence of teams is a head-to-head contest, and a scoring system that reflects that (which IMP scoring does). This idea removes that entirely; it's just pair-of-pairs.

[nige1]

PeterAlan, on 2017-April-07, 03:22, said:

The essence of teams is a head-to-head contest, and a scoring system that reflects that (which IMP scoring does). This idea removes that entirely; it's just pair-of-pairs.

BAM scoring provides an excellent head-to-head contest. In the context of multiple-teams, however, BAM throws out data that is (arguably) relevant. MP scoring rescues that discarded data. It is Pairs-of-pairs but, IMO, the fairest multiple-teams scoring method.

[StevenG]

nige1, on 2017-April-07, 04:24, said:

BAM scoring provides an excellent head-to-head contest. In the context of multiple-teams, however, BAM throws out data that is (arguably) relevant. MP scoring rescues that discarded data. It is Pairs-of-pairs but, IMO, the fairest multiple-teams scoring method.

But we play MPs almost every session. The reason for multiple teams is so that people can play IMPs occasionally (and there's no demand for IMP pairs).

[PeterAlan]

nige1, on 2017-April-07, 04:24, said:

PeterAlan, on 2017-April-07, 03:22, said:

The essence of teams is a head-to-head contest, and a scoring system that reflects that (which IMP scoring does). This idea removes that entirely; it's just pair-of-pairs.

It is Pairs-of-pairs but, IMO, the fairest multiple-teams scoring method.

About the only head-to-head element left in your scheme is that both pairs of one "team" are playing the same boards against both pairs of the other and in opposite directions (I assume that you're positing the same multiple-teams movements). Consider (as I'm sure you're aware):

A's N/S bid and make 6♠ on a board against B's E/W; at the other table B's N/S are -1 in 7NT against A's E/W (there's a cashing A).

(1) The rest of the field is in 6NT=. A's N/S get 2 MPs and B's E/W get 100% - 2MPs. B's N/S get 0 and A's E/W get 100%. Overall A gets an average of 50% + 1 MP, B gets 50% - 1 MP.

(2) Same result at both A vs B tables, but the field is in 6minor= instead of 6NT. Now at the first A/B table A's N/S get 100% and B's E/W get 0. The MP result at the other A/B table is unchanged; overall A gets 100%, B 0%.

Nothing has changed in the A-B "match", but the overall result is determined by what all the other teams do. This is just a specialised and restricted form of matchpoints, so let's not pretend it's anything else.

[nige1]

PeterAlan, on 2017-April-07, 07:48, said:

About the only head-to-head element left in your scheme is that both pairs of one "team" are playing the same boards against both pairs of the other and in opposite directions (I assume that you're positing the same multiple-teams movements). Consider (as I'm sure you're aware):

A's N/S bid and make 6♠ on a board against B's E/W; at the other table B's N/S are -1 in 7NT against A's E/W (there's a cashing A).

(1) The rest of the field is in 6NT=. A's N/S get 2 MPs and B's E/W get 100% - 2MPs. B's N/S get 0 and A's E/W get 100%. Overall A gets an average of 50% + 1 MP, B gets 50% - 1 MP.

(2) Same result at both A vs B tables, but the field is in 6minor= instead of 6NT. Now at the first A/B table A's N/S get 100% and B's E/W get 0. The MP result at the other A/B table is unchanged; overall A gets 100%, B 0%.

Nothing has changed in the A-B "match", but the overall result is determined by what all the other teams do. This is just a specialised and restricted form of matchpoints, so let's not pretend it's anything else.

I think I understand Peter Alan's argument but I disagree with it. I'm happy to concede that Match-points is Match-points. What is the virtue of excluding results from other matches on the same boards? What about cross-imps and Butler scoring?

[mycroft]

Jeremy69A, on 2017-April-06, 10:42, said:

No idea why we ever use victory points. It seems reasonable to have a cut off so you lose whatever you lose but you can't win more than whatever constituted 20-0.

So, I hit the swingy team (or the World Class team) on the swingy boards, and as expected, get crushed. They get MAX IMPS (fine), but we get -ACTUAL? But our competition got them next round where even with the utmost skill, the boards only got the champions a 30-1 win? They get 2xMAX, but our competition gets -29 and we get -50. Is that really fair?

Why not a cutoff where you can't win - or lose - more than MAX? Why is it important to crush the spirits of the 80-IMP losers as opposed to the 35-IMP ones?

The IMP league in Toronto worked this way - sort of (I didn't like that either, and it was the "expected winners" who liked this method over VPs). You got your IMP margin for the 24-board result, capped at +/- 40.

Now in short matches, who knows? I like IMPs just as much as the next guy (and think Matchpoints is a lovely game, a challenging game; but sometimes would like to play *bridge* instead); and know that board volatility is inherent to the game. But some amelioration of the swings and roundabouts seems appropriate. What? Well, I haven't answered the OP; there might be a reason for that.

[paulg]

nige1, on 2017-April-07, 08:24, said:

I think I understand Peter Alan's argument but I disagree with it. I'm happy to concede that Match-points is Match-points. What is the virtue of excluding results from other matches on the same boards? What about cross-imps and Butler scoring?

Nigel,

I think you have this completely the wrong way around. In any teams event, I only want the results between my team and the other team to count on any individual board - I do not care what other teams do on the board and I do not want their results changing mine: if I did want this situation then I'd play pairs, not teams.

The world has also voted with its feet that it does not want BAM scoring and it does not want win-loss Swiss events. They do not want BAM because it is too difficult and the strongest team almost always wins. They do not want win-loss Swiss because an early defeat means that you have little chance of winning the event. I doubt that many really care about the difference imps and victory points, although I'm in agreement with Jeremy about the futility of VPs - there is a natural alternative to using a scale where every imp counts.

I don't understand your reference to cross-imping. This is a pairs metric, not teams.

p

[nige1]

paulg, on 2017-April-07, 10:15, said:

The world has also voted with its feet that it does not want BAM scoring and it does not want win-loss Swiss events.

Paul is right, of course

[barmar]

paulg, on 2017-April-07, 10:15, said:

there is a natural alternative to using a scale where every imp counts.

This is why major Swiss team events have switched to the continuous VP scale. Every IMP up to some limit counts for something (so there's no "breakage"), but they get compressed.

[gordontd]

barmar, on 2017-April-07, 23:28, said:

This is why major Swiss team events have switched to the continuous VP scale. Every IMP up to some limit counts for something (so there's no "breakage"), but they get compressed.

I think the "natural alternative" Paul had in mind was to just use IMPs themselves.

[Zelandakh]

I have always found it ironic that we feel the need to impose an artifical scale to flatten the effects after choosing a scoring method that was specifically designed to flatten the effect of the original scoring method. Perhaps we could just design "Intergalactic Mathpoints" with an even flatter curve and be done with it?

The comment about continuous VP scales also reminds me that the same could easily be done for the IMP scale. So instead of defining 50-80 as 2 IMPs, we might have 50 = 1.5, 60 = 1.75, 70 = 2 and 80 = 2.25. This would have been much too complicated in the days of hand scoring but where everything is entered into a computer it is quite simple.

I suppose bridge might also consider learning from the experiences of other sports that have dealt with these issues for longer. Most sports went over to a method where a win counts as a certain number of points regardless of the size precisely to negate the effects of massive wins against lesser teams having too large of an impact. Where leagues got so small that the impact of these matches started becoming too important once more, they quickly moved to using head-to-head as the first tie break rather than goal difference. Is there any reason to suspect that a similar method would not work in bridge? Adding bonus points, such as they are curretly doing with rugby, would be a reasonable way to encourage stronger teams to continue to play well against the minnows and also for the weaker teams to try and hang on in there and keep the score down. I submit that this would both increase the excitement of events as well as being a more effective method of evening out extreme results than VPs.

Perhaps this is something for national bodies to consider at least in their trials, which are often dominated by how many the top pairs can run up against the weakest. For real competitions I guess such a proposal is far too radical even to be considered within my lifetime, even if bridge manages to survive beyond that timeframe.

[barmar]

Zelandakh, on 2017-May-05, 10:08, said:

The comment about continuous VP scales also reminds me that the same could easily be done for the IMP scale. So instead of defining 50-80 as 2 IMPs, we might have 50 = 1.5, 60 = 1.75, 70 = 2 and 80 = 2.25. This would have been much too complicated in the days of hand scoring but where everything is entered into a computer it is quite simple.

Lots of tournaments still do hand scoring for the matches, we only use the computer to convert from IMPs to VPs.

Quote

I suppose bridge might also consider learning from the experiences of other sports that have dealt with these issues for longer. Most sports went over to a method where a win counts as a certain number of points regardless of the size precisely to negate the effects of massive wins against lesser teams having too large of an impact. Where leagues got so small that the impact of these matches started becoming too important once more, they quickly moved to using head-to-head as the first tie break rather than goal difference. Is there any reason to suspect that a similar method would not work in bridge? Adding bonus points, such as they are curretly doing with rugby, would be a reasonable way to encourage stronger teams to continue to play well against the minnows and also for the weaker teams to try and hang on in there and keep the score down. I submit that this would both increase the excitement of events as well as being a more effective method of evening out extreme results than VPs.

Swiss Teams used to be win-loss scoring, the switch to VPs was done to make big wins more significant.

Yes, there are some anomalies -- if a top team starts out against a minnow, they'll get a huge initial boost. But the theory is that this will force them to play against other good teams pretty soon, sothe cream will all rise to the top. On the other hand, if you just use win-loss, they'll win this first match, and it's a toss-up whether they'll play another good team or a team that just managed to get lucky and squeak out a little win against a contender.

If you want an event where the size of the win doesn't matter, play KOs.