I passed the 2D limit raise of hearts by GIB over robot partner's neg dbl. Then I back into the auction with a 2S call. That should be the end of it. Instead, robot decides to tell the same story twice with a limit raise of spades. That turned what would have been a decent board into a cold bottom. (I dropped a trick out of annoyance.)
Page 1 of 1
What am I expected to do? Robot bids crazy forcing raise
#1
Posted 2024-April-12, 13:01
I passed the 2D limit raise of hearts by GIB over robot partner's neg dbl. Then I back into the auction with a 2S call. That should be the end of it. Instead, robot decides to tell the same story twice with a limit raise of spades. That turned what would have been a decent board into a cold bottom. (I dropped a trick out of annoyance.)
#2
Posted 2024-April-12, 16:21
Pick a question / answer..
Why is GIB programmed to bid 3♥ here? It isn't; it's told to pass; a simulation told it not to make the usual bid.
Does it always bid 3♥? No; giving GIB this hand 100 times, it passes 98 times, and bids 3♥ twice. You were one of the unlucky 2%.
Why would GIB ever think 3♥ is right, knowing you can't have a fit or would have raised earlier? Because GIB doesn't know this; in most cases it ignores any implications of passes entirely. In fact, despite the fact the description says 3+ spades, it assumes you have 4 spades 100% of the time.
Even assuming you had 4 spades, why would it ever bid 3♥ with just a 7 count? Because it doesn't know it is allowed to simulate later. The same database that tells it to pass now also tells it to pass if the opponents compete to 3♥, so in simulated deals where 2♠ isn't passed out and 3♠ beats 3♥, it sees bidding now as a winner. It's rare, but plausible there are enough of these cases to outweigh the rest, thus the 2%.
Even if that were all true, why would it not then bid 3♠ now in those 2% of cases? Probably because it doesn't see any difference; in those 2 cases, it didn't think you'd ever bid game anyway. The logic thus could potentially be that 3♥ is "alphabetically" before 3♠ (in the sense of listed earlier in the database).
What would GIB have done? Doubled to show 15+ total points, instead of bidding 2♠. Despite the fact the description also shows 2+ hearts, implying it may be a penalty double, GIB North always takes it out, and bids 2♠.
Does that mean I should have doubled? Who knows; I would probably have taken the 98% option too.
Can/will this be fixed?
Why is GIB programmed to bid 3♥ here? It isn't; it's told to pass; a simulation told it not to make the usual bid.
Does it always bid 3♥? No; giving GIB this hand 100 times, it passes 98 times, and bids 3♥ twice. You were one of the unlucky 2%.
Why would GIB ever think 3♥ is right, knowing you can't have a fit or would have raised earlier? Because GIB doesn't know this; in most cases it ignores any implications of passes entirely. In fact, despite the fact the description says 3+ spades, it assumes you have 4 spades 100% of the time.
Even assuming you had 4 spades, why would it ever bid 3♥ with just a 7 count? Because it doesn't know it is allowed to simulate later. The same database that tells it to pass now also tells it to pass if the opponents compete to 3♥, so in simulated deals where 2♠ isn't passed out and 3♠ beats 3♥, it sees bidding now as a winner. It's rare, but plausible there are enough of these cases to outweigh the rest, thus the 2%.
Even if that were all true, why would it not then bid 3♠ now in those 2% of cases? Probably because it doesn't see any difference; in those 2 cases, it didn't think you'd ever bid game anyway. The logic thus could potentially be that 3♥ is "alphabetically" before 3♠ (in the sense of listed earlier in the database).
What would GIB have done? Doubled to show 15+ total points, instead of bidding 2♠. Despite the fact the description also shows 2+ hearts, implying it may be a penalty double, GIB North always takes it out, and bids 2♠.
Does that mean I should have doubled? Who knows; I would probably have taken the 98% option too.
Can/will this be fixed?
#5
Posted 2024-April-13, 15:16
pescetom, on 2024-April-13, 12:42, said:
I appreciate yours, but have another question.
Would South never pass 2♦ with 4 card spades?
If not, then what is 2♠ explained as showing 3+card about?
Would South never pass 2♦ with 4 card spades?
If not, then what is 2♠ explained as showing 3+card about?
There's a low-level default rule saying that bidding a suit naturally that partner has shown some length in (3+ cards) shows a minimum combined length of (level+5) - i.e., a 7 card fit at the 2 level.
This is overridden by specifically defined rules in most other sequences that show a proper fit, like if you raised on the previous round.
There's a higher-level default rule saying that if you haven't shown a fit yet but have one, show it. GIB will never pass, then bid 2♠, but if a human passed the first time with 4 spades, then GIB subbed in, it would bid 2♠ based on this rule.
GIB doesn't use any negation logic at all, so it can't figure out that an initial pass should result it in throwing out other rules that would only have made sense if it didn't pass.
Thus the situation where it's described as an impossible 3, but simulated as an impossible 4.
#6
Posted 2024-April-13, 15:20
smerriman, on 2024-April-13, 15:16, said:
There's a low-level default rule saying that bidding a suit naturally that partner has shown some length in (3+ cards) shows a minimum combined length of (level+5) - i.e., a 7 card fit at the 2 level.
This is overridden by specifically defined rules in most other sequences that show a proper fit, like if you raised on the previous round.
There's a higher-level default rule saying that if you haven't shown a fit yet but have one, show it. GIB will never pass, then bid 2♠, but if a human passed the first time with 4 spades, then GIB subbed in, it would bid 2♠ based on this rule.
GIB doesn't use any negation logic at all, so it can't figure out that an initial pass should result it in throwing out other rules that would only have made sense if it didn't pass.
Thus the situation where it's described as an impossible 3, but simulated as an impossible 4.
This is overridden by specifically defined rules in most other sequences that show a proper fit, like if you raised on the previous round.
There's a higher-level default rule saying that if you haven't shown a fit yet but have one, show it. GIB will never pass, then bid 2♠, but if a human passed the first time with 4 spades, then GIB subbed in, it would bid 2♠ based on this rule.
GIB doesn't use any negation logic at all, so it can't figure out that an initial pass should result it in throwing out other rules that would only have made sense if it didn't pass.
Thus the situation where it's described as an impossible 3, but simulated as an impossible 4.
And I thought human partners were perverse about negation
Thanks.
Page 1 of 1