It seems like the GIB refers to "the Law" sometimes, but then doesn't act on it when it's obvious, such as in this example: http://tinyurl.com/y4vlrrtm
Especially in IMPs/vulnerable, as it was, I don't know a human alive that wouldn't automatically bid 4S as North automatically. What happened here? Any thoughts?
Page 1 of 1
GIB and law of total tricks
#2
Posted 2019-February-23, 18:30
Hi
I'm sure others may have a more "expert" analysis than me but I agree that any normal human or bot would bid 4S under almost any criteria. I'm not a big fan of the law unless it is applied well but it should know it has 9-10+ trumps. A human law player would then compete but maybe not a bot with relatively few hcps. Maybe you could have doubled again but its risky? I also dont know how well it takes account of equal vulnerability and aspects of penalties vs contract in a competitive auction etc. Maybe its analysis showed a better chance of setting 4H than making 4S although that seems surprising. You would need a GiB simulation expert to clarify that
I would also bid to 4 using losing trick count
Ive been playing a lot with GiB over 8 months and it can be frustrating. However as a program it clearly cannot think flexibly like a human or (dare I say) a more advanced bot
Just put it down as one of those annoying things and have an occasional rant like the rest of us
regards P
I'm sure others may have a more "expert" analysis than me but I agree that any normal human or bot would bid 4S under almost any criteria. I'm not a big fan of the law unless it is applied well but it should know it has 9-10+ trumps. A human law player would then compete but maybe not a bot with relatively few hcps. Maybe you could have doubled again but its risky? I also dont know how well it takes account of equal vulnerability and aspects of penalties vs contract in a competitive auction etc. Maybe its analysis showed a better chance of setting 4H than making 4S although that seems surprising. You would need a GiB simulation expert to clarify that
I would also bid to 4 using losing trick count
Ive been playing a lot with GiB over 8 months and it can be frustrating. However as a program it clearly cannot think flexibly like a human or (dare I say) a more advanced bot
Just put it down as one of those annoying things and have an occasional rant like the rest of us
regards P
Page 1 of 1