Leaderboard

Thanks for your thoughts! 1. I am also interested! 2. I think it will not develop transfers at all. Maybe I am wrong. Application of a deep learning to bridge bidding is certainly a good idea! Recently (2016) there was made a great breakthrough in this area. See this paper: https://arxiv.org/pdf/1607.03290.pdf However, the method (DRL from the paper) has its limits. I will probably not be applicable to competitive bidding. Since the algorithm learns each sequence separately by Q-learning, it can't learn more general and abstract rules. This is a major obstacle. In competitive bidding, there are way too many possible sequences (they are there inherently; and what's worse, the algorithm cannot choose to limit the length of the bidding as it did with 4 bid limit in pure NS bidding). There will never be enough training data to learn every sequence from them. For comparison, real bridge players (people) don't learn each sequence separately. They use analogies. My work focuses on creating generalized rules. I hope that my program will be able to perform in a competitive bidding, at least to some degree. I decided to use evolutionary algorithms. However, it is highly possible that deep learning is also useful for this purpose. For example, I could use evolution for exploration and generalization of bidding rules and then I could use deep learning for fine tuning of parameters (without creating nor destroying existing rules).

Hello! I am a student of Computer Science at the Charles University in Prague. I am also an enthusiastic bridge player. You can probably guess where this combination of interests leads to ... My master thesis deals with an artificial intelligence for bridge. Specifically, my goal is to let my computer design its own bidding system without any prior knowledge. I perform experiments with evolutionary algorithms (involving genetic programming and learning classifier systems) for this purpose. Would you be interested in watching my progress? I think that posting information about my research could increase my motivation to progress faster.

First, I would like to share one funny moment with you. About 5 months ago, when I finally implemented the learning classifier based genetic programming, I decided to train it on about 200 deals (which is not enough of course) and then test it on 1 deal, only to see the opening bid. I didn't expect any high-quality bidding, I just wanted to see whether it can do something. pass I started it again. pass I wanted to see, whether it can bid anything else (because always-pass system would be a symptom of either underfitting or implementation bug). So I started it again. pass It seemed that it can bid only pass. I started to feel disappointed. However, I started it again. 7NT Well, it doesn't look like a reasonable opening!

Before I share information about methods that I use, I will add a brief comment on your question. My algorithm will never explore relay bids. I don't inject any human-created knowledge into the system. I let my algorithms create everything from scratch, using a form of reinforcement learning. And because I decided to use a model-free approach to reinforcement learning, "pure ask bids" cannot emerge. PS: With a very, very, very low probability, a "pure ask bid" can occur. However, the evolution will not be able to select them by their fitness anytime soon after they appear, so they will most likely die.

I don't think double then 3S is forcing. It is a truism of bridge theory that one does not design two methods to show the same hand. To me, if I reopened with a double and then wanted to create a unilateral gameforce in a new suit, I would double, then cuebid, and then show my suit. Such would announce to partner that he can pass me, below game, only if holding only 12 cards or fewer. Since I have that route to establish an unequivocal force, I don't need and would not use double then 3S to show the same hand. To me double then non-jump spades shows a hand unable to balance with 1S or 2S (the latter showing an opening hand with 6 spades). So double then 2S would be, say, 16 high and 5 spades. Double then 3S shows a better hand, but one can have a hand much stronger than 16 high with 5 spades, and still be from gf. AKQxxx x AJx KQx This is a double then 3S hand. Were partner to hold something like xx xxx xxxx xxxx I'm too high already and I surely don't want to bid a game. Whereas if I double then cuebid then bid spades, partner, even with a misfitting yarb, must bid As for what approach I would take: I'd bid a simple 4S. While the hand is strong enough to double and cue, the odds that we have a slam that both makes and can be bid with some measure of confidence is vanishingly low. Meanwhile, double, and even more so double then cuebid begs to wrongside the contract, often costing us an imp. Say it costs us an imp, compared to 4S from my side, 25% of the time, we'd need to have better than 40-1 odds of being able to bid and make a slam. I don't see it. 40-1 is not a high hurdle, but to place partner with an Ace means LHO opened 1st seat with at most 11 hcp, one Ace and no Kings. If partner has an Ace, even it being the heart Ace doesn't mean we are making. And in any event, even if we can hold our red suits to one loser, we still have clubs to worry about. So much so that I doubt that we can bid a making slam with any confidence, and gambling a slam in this situation, merely because partner has an Ace (less so if it the unlikely heart Ace) seems foolish to me. As for double then 3N....no doubt that it can be a big winner. However, I don't like gambling that the opps won't do something brilliant, either because they are brilliant or because they stumble into brilliancy. LHO, on the auction double then 3N, will likely guess that his partner has no entry. So a low heart lead is unlikely to prevail. He may well lead the diamond Ace and then have them run some diamonds and then play a heart through. Down 6 in 3N cold for 4S is likely to cause some amusement. Or the heart Ace and switch to the diamond Ace, although we are probably down no more than 1 or 2 on that defence, assuming partner has no stopper. Sure, 4S could fail. But I think it more likely than not to make, especially if I just up and bid it. It may be difficult for an opponent to hold onto enough clubs to beat me even if, double-dummy, it would be easy and, of course, 4S will often be cold or cold on the lead.

[hv=d=w&v=0&b=8&a=1hppdp2dp3n]133|100[/hv]