[AyunuS]

http://tinyurl.com/y87juqdw with advanced robots in IMPs.

I know GIB sometimes doesn't even lead AK against 6NT but this is a different problem, although I feel its leading in general against 6NT can be improved.

So what's the problem with this lead? Well it's leading away from a very high probability offsides for declarer ♠K. In fact, if you count the tricks, it looks like the only way declarer can make is if west leads away from the ♠K as there are only 6 heart tricks, 1 diamond tricks, 3 club tricks, and 1 spade trick, for a total of 11 tricks, that declarer can win before having lost 2 tricks.

I think maybe it thought the keycard my team was missing was ♠A. But I think it's incorrect to suspect that. If south is willing to bid 6NT over the obvious 6♥, I'd put south at having:

• 19-21 HCP. West has 4, east presumably has at least one keycard worth at least 3, and north has at least 12, so there's only at most 21 left south could have. 19 is the minimum worth going for 6NT with given the bidding.
  
• 3334 or 2344 distribution. With 4+♥ or any non-super balanced hand, south would've opted for 6♥.
  
• At least an A or K in each suit except hearts, knowing partner has that covered. Again, if not, south would've opted for 6♥.
  
• Likely broken sequences of honors in suits, in hopes of picking up an extra trick on the opening lead. I don't know if GIB can think this way very well or not, but good bridge players will think about who gets to lead in the bidding. South has a very good hand to get to play last with on the opening trick. Meaning a good player could probably figure out it's good not to lead away from a K with how south bid.

Also I'd put east at having about 0-1 points other than the keycard that north/south don't have, meaning between bullet point 3 and this, it's virtually guaranteed that north/south have at least ♠AQ, meaning that leading away from the K is very likely to not end well.

Anyway, I'd be curious to see the evidence in favor of a spade lead here. I feel like all points point against it.

[johnu]

AyunuS, on 2017-December-23, 14:22, said:

I feel its leading in general against 6NT can be improved.

A few more contract types where leads could be improved on:

1. Part scores
2. Game contracts
3. Other slam contracts

At least the lead wasn't from a singleton ♠K.

I think your request, while reasonable, is the type of request that can't be accommodated by BBO. That would require much more complicated/sophisticated simulations (and inferences) than GIB is programmed for.

[steve2005]

Yes lol at least not singleton king. Have been getting at lot of those at NT.

I don't think ♠ lead that bad. ♠ was the unbid suit. If partner has Q♠ may be only lead to beat. 6N making surely bad so may be time.
But your right South surely has A♠ so may be right to be passive.

Declarer can make on non-♠ by leading towards south and picking up AJ doubleton ♦

[AyunuS]

Hmm, maybe I jumped to a conclusion too quickly about the ♦. I was thinking they could duck and only let me take with 1 of the K and Q, so that there'd only be 1 winner there. Maybe if with a low lead from North and then East either just playing the A or with some crazy good guess that East would have to play the A on the second diamond trick and saving the K, then 2 diamond winners were possible.

And yes, I know most of my bullet points are probably more like how a human player thinks than how a robot thinks. But it could still probably simulate better in order to usually avoid leading away from a K on a 6NT.