[uva72uva72]

My link

Match points, ACBL robot individual.

In every exposition of negative doubles I have encountered (including the most recent edition of the ACBL Encyclopedia) a negative double is used to describe hands for which no other good bid is available. The Encyclopedia, in fact, states that "Players should avoid a negative double when a good natural bid is available." So, over West's 3♥ bid, North might double holding a game-going hand with only 4♠, or with 5♠ and a hand too weak to force with 3♠. However, with a direct 3♠ bid available to show a game going hand with 5+♠, North should not double with the hand it holds. In addition to the fact that double represents an inaccurate description of North's hand, note that, because it does not establish a game force, it forces South to choose between 4♣ (12-21 total points, an impossibly wide range and, on the low end, a huge underbid), risking a missed game, and and the slight overbid of 5♣. Looking at the results, nearly half of the 4♣ bidders, having underbid and gotten away with it, raised North's 5♣ to 6; while South's 5♣ bid brought out the hangman in North. Is it not possible to stop the robot from bidding a grand slam when an Ace could be missing and it cannot check? Substitute the ♠Q for the 5 in South's hand and it would fully comply with the system definition for the 5♣ bid, but the grand would still fail against the ♥A lead.

I'd also be interested to know why the robot will lead an Ace against a grand slam bid in a suit but (in my experience) will almost never do so against 7NT, even when it has doubled that contract.

[Bbradley62]

So, we're talking about Negative Doubles, not Double Negatives

[steve2005]

What a silly 7♣ bid. Gib doesn't have 1st round control in ♥ a suit bid by the opposition.

But even if Gib had A♥ instead of KQ♥ its a huge gamble to bid 7♣ when 6♣ will likely give a good score. Partner is bidding on shape and Gib is assuming a monster hand and simulating its way to a bad contract.

[Bbradley62]

Of course, GIB doesn't understand the concept of "a good score"; he can only calculate the probability (in a small number of simulations) of making each potential contract and factor that with the points scored for each.

The problem here is the definition of 5♣. It includes "strong rebiddable clubs", which means at least six clubs, and at least as long diamonds. Presumably, this makes the probability of making 7♣ sufficiently high to warrant placing the contract there.

[steve2005]

 Bbradley62, on 2014-July-20, 09:45, said:

The problem here is the definition of 5♣. It includes "strong rebiddable clubs", which means at least six clubs, and at least as long diamonds. Presumably, this makes the probability of making 7♣ sufficiently high to warrant placing the contract there.

but only if p has A♥ or void in ♥ which seems low. Don't know if Gib has way to ask for ♥ control? 5 ♥ asking?. but to bid 7♣ missing A♥ when there is still lots of bidding room not good

plus if p has a void KQ ♥ wasted on your bidding grand with 9 working hcp

[Bbradley62]

 steve2005, on 2014-July-20, 12:29, said:

but only if p has A♥ or void in ♥ which seems low. Don't know if Gib has way to ask for ♥ control? simple cuebidding perhaps. but to bid 7♣ missing A♥ when there is still lots of bidding room not good

The probability of South having a heart void is not low since North thinks South has shown at least 12 cards in the minors. So, it's more than 50%, since there are more available slots in spades than in hearts for South's 1 card in the majors, and South could even be 0076. Of course, North doesn't directly calculate probabilities, he simulates hands, so who knows what his sample size and/or resulting percentage might be.

[uva72uva72]

If the jump to 5♣ shows 6-6 in the minors, I'd suggest that the system notes should so state and NOT state that the bid shows "5+♦s." My bridge training was to the effect that, if I opened in one suit, a decent second suit of 4 cards was "biddable" and a second suit of 5 cards was "rebiddable." I'd suggest that KQJxx is a very strong rebiddable suit by that definition.

[Bbradley62]

 uva72uva72, on 2014-July-21, 20:19, said:

My bridge training was to the effect that, if I opened in one suit, a decent second suit of 4 cards was "biddable" and a second suit of 5 cards was "rebiddable." I'd suggest that KQJxx is a very strong rebiddable suit by that definition.

Apparently, your bridge teacher and GIB's programmer did not use the same dictionary. Barmar has posted explanations of GIB's terminology for definition suit length/strength: http://doc.bridgebas...criptions.xhtml
As you can see from that post, "strong rebiddable" is better than "twice rebiddable" and is defined as six cards including either AKQ or 4 of the 5 honors. "Very strong rebiddable" is not a GIB-accepted term, although "solid 6card" means AKQTxx or AKQJxx.