Help
On Tricks 10-11, why does East keep his useless ♥7 instead of protecting his ♣Q?
Page 1 of 1
Throw out losers; protect potential winners
#1
- Group: Advanced Members
- Posts: 6,542
- Joined: 2010-February-01
- Gender:Male
- Location:Brooklyn, NY, USA
Posted 2011-March-09, 20:38
On Tricks 10-11, why does East keep his useless ♥7 instead of protecting his ♣Q?
#2
- Group: Advanced Members
- Posts: 7,786
- Joined: 2004-September-05
- Gender:Male
- Location:Galloway NJ USA
-
Interests:Bridge, Poker, participatory and spectator sports.
Occupation - Tax Attorney in Atlantic City, NJ.
Posted 2011-March-10, 09:46
Anyone who plays with or against the robots sees this kind of thing all the time.
#3
- Group: Admin
- Posts: 22,136
- Joined: 2004-August-21
- Gender:Male
Posted 2011-March-10, 14:21
My guess is that it assumed you had to have either [HQ] or ♣K for your bidding, and in those cases it didn't matter what it pitched.
I don't think GIB knows the strategy "If there's only one layout that allows us to succeed, assume it's true even if it's inconsistent with the bidding/play so far." Its strategy is based on generating hands consistent with what it knows, and then performing double dummy analysis of them.
I don't think GIB knows the strategy "If there's only one layout that allows us to succeed, assume it's true even if it's inconsistent with the bidding/play so far." Its strategy is based on generating hands consistent with what it knows, and then performing double dummy analysis of them.
#4
- Lamentable Bunny
-
- Group: Advanced Members
- Posts: 1,505
- Joined: 2008-March-01
- Gender:Male
- Location:Portland, ME
Posted 2011-March-10, 14:37
barmar, on 2011-March-10, 14:21, said:
My guess is that it assumed you had to have either [HQ] or ♣K for your bidding, and in those cases it didn't matter what it pitched.
I don't think GIB knows the strategy "If there's only one layout that allows us to succeed, assume it's true even if it's inconsistent with the bidding/play so far." Its strategy is based on generating hands consistent with what it knows, and then performing double dummy analysis of them.
I don't think GIB knows the strategy "If there's only one layout that allows us to succeed, assume it's true even if it's inconsistent with the bidding/play so far." Its strategy is based on generating hands consistent with what it knows, and then performing double dummy analysis of them.
I think that is how it currently works. Out of curiosity, would the following improvement be possible? When there are few enough remaining cards, the computer simply exhaust all possible hands. This would certainly prevent trick 9 and 10 difficulties. I dunno how much slower the computer would play.
Bridge Personality: 44 44 43 34
Never tell the same lie twice. - Elim Garek on the real moral of "The boy who cried wolf"
Never tell the same lie twice. - Elim Garek on the real moral of "The boy who cried wolf"
#5
- Group: Admin
- Posts: 22,136
- Joined: 2004-August-21
- Gender:Male
Posted 2011-March-10, 15:34
It's not that simple. Suppose half the hands say that pitching clubs is right, the other half say that pitching hearts is right. To decide which one to go with, it has to figure out which of them are more consistent with the bidding and play.
It could be even worse. Maybe 75% of the possible card layouts suggest pitching clubs, but inferences from bidding and play should tell you that most of them are unlikely.
It could be even worse. Maybe 75% of the possible card layouts suggest pitching clubs, but inferences from bidding and play should tell you that most of them are unlikely.
#6
- Lamentable Bunny
-
- Group: Advanced Members
- Posts: 1,505
- Joined: 2008-March-01
- Gender:Male
- Location:Portland, ME
Posted 2011-March-10, 16:05
barmar, on 2011-March-10, 15:34, said:
It's not that simple. Suppose half the hands say that pitching clubs is right, the other half say that pitching hearts is right. To decide which one to go with, it has to figure out which of them are more consistent with the bidding and play.
It could be even worse. Maybe 75% of the possible card layouts suggest pitching clubs, but inferences from bidding and play should tell you that most of them are unlikely.
It could be even worse. Maybe 75% of the possible card layouts suggest pitching clubs, but inferences from bidding and play should tell you that most of them are unlikely.
Ah yes, of course...it's only helpful in cases that humans can clearly see are 100%, but those are not easy to program.
Bridge Personality: 44 44 43 34
Never tell the same lie twice. - Elim Garek on the real moral of "The boy who cried wolf"
Never tell the same lie twice. - Elim Garek on the real moral of "The boy who cried wolf"
Page 1 of 1