GIB fell for the oldest trick in the book

[Wackojack]

[hv=d=e&v=n&n=sak4h42dj9732ck63&e=sj102h10873dk8c9742]266|200|Scoring: Chicago

Playing money bridge with unapposed bidding:

S N

1♦ 2♦

2♠ 3♦

6♦ [/hv]

 

I was sitting West and led Q♣ to South's (my human opponent) Ace, who then led 5♠ to dummy's Ace. Next J♦ and East (my GIB partner) covered. I am sure you can guess what happened next.

 

I already knew that GIB could make wild and incomprehensible bids from time to time, but always thought that its play of the cards was top notch. I can't see any card combinations where covering pays over not covering. Am I missing something? My human opponent must have been laughing himself silly.

 

Even if North had led Q from ♦Q9xxx I think it would still be wrong for east to cover.

[Free]

It depends on what 2♠ and 3♦ mean imo. Because it might suggest that opener has only a 3 card ♦, in which case covering might be better. But if opener suggests 4+♦, there's indeed NO reason at all to cover :)

[Wackojack]

GIB prompt on 6d bid says 4+ diamonds, stop in spades

[Stephen Tu]

My guess is that bidding generator couldn't conceive of hand where opener jump shift then blast 6♦ not holding DAQ. If all the hands it deals out for a representative sample hold DAQ, then it doesn't matter to cover. Also note that it's right to cover if declarer made this play (a technical mistake without additional clues, though not *huge* mistake) opposite AQ8x, partner with Tx, you are pretending to have KTx which declarer catering to by lead high from dummy. Maybe hope partner has Tx plus a heart trick & declarer goes wrong.

 

Perhaps bidding generator should give more lax parameters when defending to account for crazy overbidding opponents :)

 

BTW it's GIB's *declarer* play that is top notch, world class really, if it understands what it is going on in the auction & is making the right inferences about the bids. It uses true single dummy techniques there (though perhaps not first couple tricks depending on speed settings & flags, don't know about BBO specifics). Defense - not so much, doesn't really pick up inferences from signalling (other than count).

[EricK]

If you had the singleton ♦A then it serves you right for not leading it.

 

I have read a fair number of defensive books, and whenever the player on lead has the singleton Ace of trumps it is correct to lead it to avoid a throw in.

[Stephen Tu]

Certainly it's possible to construct hands where singleton DA lead is necessary to avoid a situation as you say, but it's silly to make a blanket statement that it's always right to lead it.

 

There are hands where you have to set up a side suit trick then use DA to cash it; if you cash DA first declarer has time to pitch loser in that suit. Or you might find that the opponents have reached slam with suit like Qxxx opposite K9xxx, partner with JTx, and declarer was going to misguess. or Qtxx opposite K9xxx.

 

I think those situations are much more common than one where declarer has to go for an endplay line; after all singleton A is rare (11-12 cds fit rare, 10 cd & 9 cd fits quite antipercentage to find stiff A), and these lines often may risk opp ruffing in with small trump, when alternate line available.

 

In those books your hand + bidding may suggest lead singleton trump ace is right, but I doubt they said it's right to do it routinely no matter what.

[cherdano]

If you had the singleton ♦A then it serves you right for not leading it.

 

I have read a fair number of defensive books, and whenever the player on lead has the singleton Ace of trumps it is correct to lead it to avoid a throw in.

I believe you 100% this rule holds in defensive books. I am not so sure it holds in real life :o

[cherdano]

Btw, I would play low here not because partner might have singleton A (that didn't occor to me, really...) but because declarer might have AQ-6th.

 

Arend

[pclayton]

Also note that it's right to cover if declarer made this play (a technical mistake without additional clues, though not *huge* mistake) opposite AQ8x, partner with Tx, you are pretending to have KTx which declarer catering to by lead high from dummy.

Except we hold the ♦8, so this isn't the case.

[Trinidad]

My guess is that bidding generator couldn't conceive of hand where opener jump shift then blast 6♦ not holding DAQ.  If all the hands it deals out for a representative sample hold DAQ, then it doesn't matter to cover.

That is also my guess. The bridge program Jack makes sure that there are some "foul deals" in the sample. That method is used to break a tie between technically 100% plays (as in really 100% unless someone revoked) and statistically 100% plays (as in 99.9 % since there may be an idiot [or a human] at the table).

 

Without the foul deals the program would consider both plays equally good and would pick one of the two. It looks like that is what GIB did here.

 

Rik