Join Bridge Winners
All comments by Robert Thorstad
You are ignoring the author of this comment. Click to temporarily show the comment.
I would consider involving a dean. Can you talk to the student life or other relevant dean and explain the situation? This sounds like a lack of knowledge from the people making the decision.
Sept. 27, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Just describing their system i trust people can think for themselves
Sept. 17, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Can be 4432 else 4 or more
Sept. 16, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Love the way you described these options. So obviously 2S is out? :P
Sept. 6, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Fixed thanks Richard
Aug. 31, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Crap. Typo. Meant 2 (but 5 people bought it!)
Aug. 31, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Sorry 3D was GF minors . 3S just spades
Aug. 3, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Fair enough, Michael. Yeah, I meant simulate 1x/real play against yourself. Fair enough there's a lot of hidden complexity in that step.
June 25, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Michal can you unpack that ? Not familiar with DD solving but seems I need to check all permutations . Self play can’t be worse than that in complexity (its linear).
June 25, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
How about , for a simulated hand, score it using actual result of self-play (play against yourself) with the appropriate hands closed?
June 24, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Serious question, has any effort been put in to convince someone working in the machine learning community to take on this problem? Feels like an interesting scientific problem, and, of course, would be good for us.
June 24, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Does anyone have a reading list put together here? Would be curious to see what has been done. both old and new (for example there's a research paper on GIB, isn't there?).

To state the obvious, I would expect a machine learning algorithm that succeeds at bridge will be based on reinforcement learning: taking actions (like playing cards) and then observing whether you win/lose IMPs. Bidding can of course also be tackled as a supervised learning problem: learn to make the bids that experts made in real auctions.

I also think an algorithm will need to get past double dummy analysis. Learning to take the action that leads to the best double dummy result is not the same as learning to take the action that leads to the best single dummy result (e.g. if you lead to a stiff king in the closed hand, opponents don't always get it right, also, opponents can't always read an end position, etc).

Think there should be much more discussion about this topic. Bridge would be an interesting AI challenge. With a lot of effort, I bet we could get a machine learning algorithm that beats humans at bridge.
June 24, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Strongly agree, nice article Michael. I especially agree with your point about starting with mini bridge.

One more thing we can all consider: is there a club game that juniors can play in your town? Juniors can usually only play nights and (even better) weekends. Do you have a non-weekday-during-the-day game?
June 17, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
How do you preference hearts with this partner ?
June 4, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Yeah I opened 5D and LHO made the same decision. Polls this week are really not coming in for me, I guess I was just too aggressive. Still 8-4 with a super pure preempt. And bidding is fun.
May 15, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Man no one taking me up on 3NT ? :P cowards
May 12, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Yeah fair enough . Assume any two suiter. At the table was not discussed with a strong partner but let's simplify by saying he has 2 suits but not necessarily minors
April 26, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Lavazza
Nickell
March 15, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Congratulations!
March 12, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Correct you have no special agreements here
Jan. 23, 2018
.

Bottom Home Top