[gwnn]

ok I accept that my bidding is not very intelligent here but:

Axxx
Axx
Txxxxx
void

1♣-p-3♠*-p
3NT-x-p-p
4♣-x-p-4♥!!!

I mean OK all doubles are takeout it's just a problem with the rules but seriously? 4♥? what am I supposed to hold here?

[Bbradley62]

Yes, partner's bid Is bizarre, but did you check to see how partner would read your bids before you made them?

THank you for the clarification, Helene... I don't know how I read the original post.

This post has been edited by Bbradley62: 2010-March-28, 10:56

[helene_t]

I think gib held the hand posted and bid 4♥, Bradley.

[arigreen]

Interesting. I tried several simulations and most of the time GIB bid 4♠. Occasionally it bid 5♦, but never 4♥. I'm not sure what to tell you here, except that GIB does not have a good definition for the double of 3N. (16- total points)

[junyi_zhu]

arigreen, on Mar 29 2010, 06:45 AM, said:

Interesting. I tried several simulations and most of the time GIB bid 4♠. Occasionally it bid 5♦, but never 4♥. I'm not sure what to tell you here, except that GIB does not have a good definition for the double of 3N. (16- total points)

Is it possible to define some pure penalty double situations so that gib would never pull? Gib may not know when to make the penalty doubles, which is fine, cause human players can make them and take profit as long as gib doesn't pull the double.

[helene_t]

I think GIB would be quite good at deciding when to make such pure penalty doubles. Easy to evaluate the result since partner's action is easy to simulate

The problem is, I suppose, how to define such situations.

[arigreen]

junyi_zhu, on Mar 29 2010, 09:18 AM, said:

arigreen, on Mar 29 2010, 06:45 AM, said:

Interesting.  I tried several simulations and most of the time GIB bid 4♠. Occasionally it bid 5♦, but never 4♥.  I'm not sure what to tell you here, except that GIB does not have a good definition for the double of 3N. (16- total points)

Is it possible to define some pure penalty double situations so that gib would never pull? Gib may not know when to make the penalty doubles, which is fine, cause human players can make them and take profit as long as gib doesn't pull the double.

Yes, this is possible, and in fact we have done this in some auctions. However, 1♣-P-3♠-P-3N-X is not one of them.

[junyi_zhu]

helene_t, on Mar 29 2010, 04:51 PM, said:

I think GIB would be quite good at deciding when to make such pure penalty doubles. Easy to evaluate the result since partner's action is easy to simulate

The problem is, I suppose, how to define such situations.

One simple situation is if a double is made after two or more passes after opp opens the bidding, it has to be a penalty double, because you can't wait two rounds to make a takeout double. For example: 1S p 1N p 2C p 2H p 2S x, after two passes, this double has to be a penalty double.

Another very simple situation is that a double against any natural NT bids above 2NT (including 2NT) should be penalty.

Another very simple situation is that a double against any slam contracts should be penalty double.

Another simple situation is that when you show gameforcing value and found a fit, a double against opps sacrifice is penalty and doesn't require any extra. (Also, gib does extremely poorly in this area, it often doubles with some very distributional hands when bidding or passing should be quite obvious, so advanced rules should be determined.)

Another situation is that a double against opp's game contracts after opp has set up a gameforcing sequence is for penalty, for example: 1S p 2N p 4S x.