[matmat]

If I had to guess I'd say that someone has already tried this, or, maybe this is even played somewhere.

We have IMP and BAM team matches.

How about a mixed match?

say, 28 boards, seven four board segments. Each four board segment is scored using IMPS, but the winner of the match is determined by the score in segments.
(i.e. 4-3, 5-2 or whatever).

thoughts?

[Free]

Sounds interesting because tactics may change a lot, but it's probably just another way of giving weak teams a better chance to win against good teams.

[barmar]

This is similar to win-loss Swiss Teams, versus the more common victory points system.

[matmat]

barmar, on Sep 24 2008, 12:12 AM, said:

This is similar to win-loss Swiss Teams, versus the more common victory points system.

similar, maybe, but in the win-loss swiss you play each segment against a different team, no?

[rogerclee]

I think this method of scoring succeeds in randomizing the results of the match without significantly affecting the strategy.

[EricK]

Free, on Sep 22 2008, 12:13 PM, said:

Sounds interesting because tactics may change a lot, but it's probably just another way of giving weak teams a better chance to win against good teams.

Does it? Suppose in a single match the weak team bids a specualtive grand slam which makes. Now it is hard for the stronger team to catch up. But if the match is divided up into segments then the weak team need to do enough things to win the majority of segments.

[NickRW]

EricK, on Sep 24 2008, 09:39 AM, said:

Free, on Sep 22 2008, 12:13 PM, said:

Sounds interesting because tactics may change a lot, but it's probably just another way of giving weak teams a better chance to win against good teams.

Does it? Suppose in a single match the weak team bids a specualtive grand slam which makes. Now it is hard for the stronger team to catch up. But if the match is divided up into segments then the weak team need to do enough things to win the majority of segments.

Yeah, I think it does swing things in favour of the underdog slightly.

If we play a game with dice - you throw higher = you win, I throw higher = I win and we simulate me being a stronger player by me winning all draws. We each put a dollar, pound, sheckel or whatever in a hat and repeat the game 6 times - if I win at least 4 of the 6 I take the money - if you win at least 4 you take the money and if it is a draw we each get our stake returned.

I won't bore you with the math, but I rate, on average to come away with $/£1.31

If we alter the game so that we play two sets of 3 and sum the result, rather than one set of 6, I rate to come out of it with $/£1.25

Nick

[EricK]

Maybe if you treat everything as purely random variables it will work out this way. But the point is that in bridge the weaker team, if they know they are weaker can and should adopt a high variance strategy. Now this will almost certainly make their losses bigger in terms of how many IMPs they lose each match by, but it will also mean that they win more matches than they would otherwise. My feeling is that a high variance strategy is more likely to win one medium size match than a majority of shorter matches.

Since that is just my intuition, it may very well be wrong, but your figures don't really cover this scenario. For one thing, it doesn't model the fact that some individual board wins are worth more IMPS than others, nor does it give any scope for high-variance strategy.

[NickRW]

EricK, on Sep 25 2008, 12:51 PM, said:

Maybe if you treat everything as purely random variables it will work out this way. But the point is that in bridge the weaker team, if they know they are weaker can and should adopt a high variance strategy. Now this will almost certainly make their losses bigger in terms of how many IMPs they lose each match by, but it will also mean that they win more matches than they would otherwise. My feeling is that a high variance strategy is more likely to win one medium size match than a majority of shorter matches.

Since that is just my intuition, it may very well be wrong, but your figures don't really cover this scenario. For one thing, it doesn't model the fact that some individual board wins are worth more IMPS than others, nor does it give any scope for high-variance strategy.

Well, we could alter the dice game such that whoever throws a 6 wins by 2 points, but keeping it so that I still win (by 1 point) the draw of a double 6 - this would keep my advantage the same. This would probably be an adequate simulation of what you're proposing.

The math is more tedious and I can't be bothered to do it just now - I guess that it would tip things so that I had less overall advantage. So I think you're right that a high variance strategy helps the underdog - so long as they pick the right sort of things to it with - for example finessing against a queen with 9 trumps and so on - maybe deciding not to bid a thin game that opps probably will and hope the lie of the cards is bad etc - not mad play and bidding which will just negate what skill they really do have.

Nick

[helene_t]

If you want some compromise between IMP and BAM I would go for Patton. Intuitively I would say Patton must have higher statistical power than your format, although I am not quite sure why I think that ....

[Free]

EricK, on Sep 24 2008, 10:39 AM, said:

Free, on Sep 22 2008, 12:13 PM, said:

Sounds interesting because tactics may change a lot, but it's probably just another way of giving weak teams a better chance to win against good teams.

Does it? Suppose in a single match the weak team bids a specualtive grand slam which makes. Now it is hard for the stronger team to catch up. But if the match is divided up into segments then the weak team need to do enough things to win the majority of segments.

If a weak team can only bid 1 speculative slam which makes, they were going to lose the match anyway. However, (extreme example coming up) if a strong team bids 4 grand slams in a row while a weak team doesn't even bid game on these, the strong team will only win 1 point while in imps they were going to win 25-0 without problems.

Look at it as follows: what if you have 4 games which are laydown. On 1 board the strong team goes for the safety play, while the weak team doesn't know it's safety plays. Obviously the safety play loses a trick. Result: the weak team wins 1 imp, and 1 point.
However, if the safety play was necessary, the strong team wins lots of imps, but still only 1 point.

I think this is a realistic scenario, so yes, I think the format gives weaker players more chance to win than at an imps match.

[barmar]

matmat, on Sep 24 2008, 04:19 AM, said:

barmar, on Sep 24 2008, 12:12 AM, said:

This is similar to win-loss Swiss Teams, versus the more common victory points system.

similar, maybe, but in the win-loss swiss you play each segment against a different team, no?

True.

What this really reminds me of is tennis. A tennis match is built out of a hierarchy of these. First you have a game, which requires winning at least 4 points with a 2-point lead. Then you have sets, which requires winning at least 6 games with a 2-game lead. Then you have a match, which requires winning the majority of sets (3 of 5 and 4 of 7 are common).

At each level of the hierarchy, the margin of victory in the lower levels is ignored. Thus, if you have a match that goes 6-0, 4-6, 6-0, 4-6, 4-6, the player who won 18 games will beat the one who won 24.