A good example of World Computer Championship

[Laskos]

World Computer-Bridge Championship, XVII
September 23 - 28, 2013
at the 41stWorld Bridge Team Championships Nusa, Bali, Indonesia (September 16-29, 2013)

http://www.ny-bridge.com/allevy/compute ... s2013.html

and the results so far

http://www.ny-bridge.com/allevy/compute ... cores.html

6 engines, 48 board Round-Robin, first two qualify to the final consisting of 64 boards. This year eternal rivals Jack and WBridge5 will face again each other in the final. These two engines pretty much shared the "Champion" title in the last 10 editions. WBridge5 is free and can be downloaded here:
http://www.wbridge5.com/
Bridge Baron is the most well known and sold bridge software, and it ended the last, 6th place.

Maybe an example for ICGA dinosaurs on how to hold WCCC.

[hgm]

What makes this any different from the WCCC then?

The links you give do not seem to list any rules for participation, but since you mention a location I suppose this was not an on-line event.

It also seems to be a multi-day event.

Bridge is a game of incomplete information where luck plays a far greater role than in Chess, so the fact that there are many games per pairing is just a game-specific adaptation. Playing 64 hands of bridge probably gives you as much information on the relative strength of the players as playing a single game of Chess.

The number of particpants is lower than the typical WCCC attendance.

On the face of it there isn't anything here the ICGA could learn from.

[Laskos]

Playing 64 hands of bridge probably gives you as much information on the relative strength of the players as playing a single game of Chess.

You are a bit confused or you don't know what you are talking about, right? Playing IMP reversed hands, it's 2 bridge boards = 2 chess games more or less. This very thing, 64 boards final, 48 boards RR to qualify ICGA has to learn.

[hgm]

It seems ridiculous to equate two reversed Bridge hands with two Chess games. The number of decisions that have to be taken to play a Bridge hand is not even at the level of what you need for a single move of Chess.

[Laskos]

It seems ridiculous to equate two reversed Bridge hands with two Chess games. The number of decisions that have to be taken to play a Bridge hand is not even at the level of what you need for a single move of Chess.

I mean statistically they are of the same order, only that probably more bridge hands have a pre-determined bidding outcome than chess opening positions have. You stated an absolute nonsense about the statistical significance of 64 boards. About the complexity of the game itself, I wouldn't digress, but the simulations do take time, and CPU time is an important factor in computer bridge too.

[hgm]

The chances that a stupid player wins against a perfect player are far greater in a hand of Bridge than in a game of Chess (where this is a virtual impossibility). You don't seem to take account of that at all. Skill in Bridge means exploiting all tiny (or at least significantly smaller than 1) probabilities that could work in your favor. If you cut instead of leading the Ace when Kx is missing, you have 50% in stead of 52% chance. If you don't know the principle of restricted choice, you might go for a cut in the 33% direction rather than in the 66%, Still means there is a 33% (or 48%) probability the stupid player that doesn't know these rules that even beginners know will win that reversed pair of hands from the player that does it right.

If you would rate Bridge players by the same method as Chess players, based on the results from single hands, there might only be 300 Elo difference between the World Champion and an absolute beginner.

[Laskos]

The chances that a stupid player wins against a perfect player are far greater in a hand of Bridge than in a game of Chess (where this is a virtual impossibility). You don't seem to take account of that at all. Skill in Bridge means exploiting all tiny (or at least significantly smaller than 1) probabilities that could work in your favor. If you cut instead of leading the Ace when Kx is missing, you have 50% in stead of 52% chance. If you don't know the principle of restricted choice, you might go for a cut in the 33% direction rather than in the 66%, Still means there is a 33% (or 48%) probability the stupid player that doesn't know these rules that even beginners know will win that reversed pair of hands from the player that does it right.

If you would rate Bridge players by the same method as Chess players, based on the results from single hands, there might only be 300 Elo difference between the World Champion and an absolute beginner.

The chances beginners draw Jack in 1 hand reversed (duplicate) are about 20%. No IMP points for a draw in duplicate. The chances beginners beat Jack on 1 hand duplicate are very tiny, maybe 0.1%, So the "Elo" span from beginner to Jack, based on IMP, if any IMP points were given, is maybe 1200 "Elo" points, not that far from chess. All this amounts to having 20% less boards than the nominal, so take 48 boards as an "effective" 40 boards.

[hgm]

I don't know where you get these numbers, but they do not sound right. Hands where it depends on a single decision whether you will make an extra trick or not are quite common. And the difference between the correct and the wrong decision giving you the extra trick is often not all that large, sometimes even close to 50%. And there are also cases where you simply cannot know at all, that it is a blind guess whether you place the Queen left or right from you. And if the beginner takes one decision, and the expert with the reversed hand takes the other, the one who is lucky gets the extra trick and wins the duplicate.

Also, bidding too high or too low because you cannot bid accurately will bring you many wins, in cases where the opponent card distribution happens to favor the 'wrong' (i.e. on average lowest scoring) bid.

[Laskos]

I don't know where you get these numbers, but they do not sound right. Hands where it depends on a single decision whether you will make an extra trick or not are quite common. And the difference between the correct and the wrong decision giving you the extra trick is often not all that large, sometimes even close to 50%. And there are also cases where you simply cannot know at all, that it is a blind guess whether you place the Queen left or right from you. And if the beginner takes one decision, and the expert with the reversed hand takes the other, the one who is lucky gets the extra trick and wins the duplicate.

Also, bidding too high or too low because you cannot bid accurately will bring you many wins, in cases where the opponent card distribution happens to favor the 'wrong' (i.e. on average lowest scoring) bid.

I think you didn't play a lot of bridge, do you? The beginners first don't know how to bid, then they don't know how to play the hand. Hell, they can't remember the cards! In 60-70% of hands a technique or a combination of techniques must be applied when playing, like signaling, holding up, blocking and unblocking, managing entries, maintaining tempo, drawing trumps, not drawing trumps, card reading, endplay, squeezes, etc. It may affect only few tricks, but these tricks are not ordered at random. Not talking of bidding systems, where novices lose from the start. Overall, from seeing novices play Jack, about 20% of the hands could be drawn by the beginners. Jack may lose to beginners in a very weird distribution of cards, like 5% of 6-0 distribution of trumps (itself a 3% event) of a defender and such rare events, not detected during the bidding. And the beginners have to fructify the over-bidding of the opponents in a duplicate.

In any case, 48 or 64 boards of Duplicate Bridge are statistically equivalent to as much or just a bit less than 48 or 64 games of chess. And the current format of Bridge WCC can be taken as an example by the silly ICGA.

[hgm]

In any case, 48 or 64 boards of Duplicate Bridge are statistically equivalent to as much or just a bit less than 48 or 64 games of chess.

That statement is completely meaningless if it is not know what the Elo difference is that has to be measured. It could very well be that the rating differences of Bridge players are 10 Elo where they would be 100 Elo between similarly ranked Chess players. In which case you would need 100 times as many games to achieve the same LOS. Your arguments are less than convincing, ans seems to stem mainly from gut feeling. There must be a reason why Bridge matches always contain a large number of hands, while Chess tournaments typically play only one game per pairing. Most Bridgers I know would consider the result of a single duplicate hand pretty meaningless. You can talk all the way you want, but you will never lull us into believing that Bridge is not a game of chance, while Chess is a game of complete information.

People play Bridge all the time, and it must be known from national competitions what the winning frequencies of individual duplicate hands is between players of various ranks.If you cannot substantiate your unlikely claims by such facts, I don't consider them credible.