[uva72uva72]

My lihttp://www.bridgebase.com/tools/handviewer.html?bbo=y&lin=pn|uva72uva72,~~M25261,~~M25259,~~M25260|st||md|3SQAH69TD2JAC389KA%2CS47JH234JD4KC246Q%2CS6KH8KAD56TQC57TJ%2C|rh||ah|Board%205|sv|n|mb|1D|an|Minor%20suit%20opening%20--%203%2B%20!D%3B%2011-21%20HCP%3B%2012-22%20total%20points%20|mb|p|mb|2C|an|Forcing%20two%20over%20one%20--%2011%2B%20HCP%3B%20biddabl|mb|p|mb|3C|an|Raise%20of%20minor%20--%204%2B%20!C%3B%204%2B%20!D%3B%2011-21%20HCP%3B%203-card%20!D%3B%2012-22%20total%20points%20|mb|p|mb|4N|an|Blackwood%20%28C%29%20--%20biddable%20!C%3B%2020%2B%20total%20|mb|p|mb|5D|an|One%20or%20four%20key%20cards%20--%204%2B%20!C%3B%204%2B%20!D%3B%2011-21%20HCP%3B%203-card%20!D%3B%2012-22%20total%20points%20|mb|p|mb|6N|an|Signoff%20--%20biddable%20!C%3B%2020%2B%20total%20points|mb|p|mb|p|mb|p|pc|H2|pc|HA|pc|H7|pc|H6|pc|CJ|pc|S3|pc|C3|pc|CQ|pc|H4|pc|HK|pc|H5|pc|H9|pc|C5|pc|D9|pc|CA|pc|C6|pc|CK|pc|C4|pc|C7|pc|S2|pc|C8|pc|C2|pc|CT|pc|D3|pc|DT|pc|D7|pc|D2|pc|DK|pc|H3|pc|H8|pc|HQ|pc|HT|pc|ST|pc|SA|pc|S4|pc|S6|mc|10|nk

Match points, ACBL robot individual

If my analysis is correct, this slam will make if:

Clubs are 2-2 (~40.7%)
Clubs are 3-1, Queen is singleton (~12.43%)
Clubs are 4-0 onside (~4.78%)
Diamonds are 3-3, King onside (~17.76%)
Diamonds are 4-2, doubleton King onside (~8.07%)
Diamonds are 5-1, singleton King onside (~1.16%)

Adding 0.1% for the very slight possibility that East holding Qxx will cover the Jack when it is lead from dummy, I estimate that this slam had a better than 85% chance of making.

Going down in this slam didn't make me a conspiracy theorist (although another hand in the same set did exhibit both a 5-0 and a 4-1 break); but it did remind me of all the 0.0% slams that continue to make because the opening leader (sometimes having doubled the final contract) will not cash the setting tricks - and even underleads aces against 7NT. I've seen this issue discussed but I've never seen any indication that it is being addressed. If this is impossible to fix, is there an explanation?

[iandayre]

I do not believe that GIB hands are more likely to have bad splits and unfavorable lies of the cards that defeat good contracts than any other bridge hands. There is nothing at all out of the ordinary about this hand, and there are plenty of hands where suits break and finesses work, allowing marginal contracts to make.

Of course your point about GIB not cashing an Ace vs 7NT (or an AK vs 6NT) is valid, but in no way related.

[uva72uva72]

Clearly I didn't make my point very well. It's that BBO provides in most respects a very realistic experience of playing bridge, to include 5-0 splits and failure in high-percentage contracts. However, in a very few areas it falls short of realism, and in the case of the robot's failure to cash setting tricks on opening lead against slams, it approaches the surreal. The only correlation between my example hand and the failure to beat nullo slams is that, as frustrating as it is to go down in a great contract, that is a normal part of the game. 6NT making off two aces (both in the hand of the opening leader) is not, and is, therefore, even harder to take. I'd hope that some emphasis would be put on correcting that problem.

[Bbradley62]

uva72uva72, on 2014-August-14, 10:50, said:

If my analysis is correct, this slam will make if:
<snip>,
I estimate that this slam had a better than 85% chance of making.

It is not relevant to your actual point, but (for future reference) you cannot simple add all these probabilities and get your chance of making the contract. You have a 58% chance of picking up the missing queen... If your other chance was another missing queen with the same probability of success, and you needed either one of the to chances to work in order to make your contract, you wouldn't simply add 58% and 58% and say that you had a 116% chance of making... You must discount the chance that both plays work, so you don't double-count those occurrences. So:
58% of the time you pick up ♣Q and succeed.
42% of the time you do not pick up ♣Q and must get 4 diamonds. This succeeds 26% (I don't think you get four diamonds if singleton K is onside) of the 42%, for a net of 11%, which can be added to the 58% for a total of 69%.

The pessimist's (and simpler) view is: In order to go down, both your club play and your diamond play must fail: (100%-58%, or 42%) * (100%-26%, or 74%) = 31%.

Of course, this ignores the influence of the club distribution on the diamond distribution, which should be negligible.

[chasetb]

Bbradley is almost 100% right. He nailed the double-counting, and how to get the Diamond percentage from what o.p. gave as the percentages to win 4 diamond tricks. SuitPlay says you can pick up a singleton King, HOWEVER if East reads the situation properly, you can't get 4 tricks from Kx onsides (duck the first round when North leads low, then when the Queen or Ten is led next, cover it). SP has it as ~19.0% that you can get 4 diamond tricks, so 65.9% of the time you should make the slam.

You are unlucky that it fails and have a few wasted HCP in Spades, but are lucky in that North showed up with 4 controls, nice spot cards, and 8 AKQ points (A = 3 / K = 2 / Q = 1 ; better than controls or HCP for NT contracts).

[iandayre]

To address uva's clarified point, it is quite easy to write a program to deal cards randomly. Obviously teaching a computer to play a complex game such as bridge is infinitely more difficult. That said I would like to see a lot more effort put towards increasing GIB's skill level.

[georgi]

Also while considering slam chances, somehow hands are not really random in robot tourneys assigned with "Best hand South".

So by having 14HCP you should have in mind partner has maximum 13HCP.

Apart from these facts you could assume there is no conspiracy for putting queens and kings not on side.

So the 6NT is in possession of maximum 27HCP. Best assets are ♥AK and ♦K. Wishing ♣Qxxx makes the best shape in front.

Nevertheless you need ♦Q and spade to finesse as you could have heart/diamond loser. 5♣+2♦+2♥+1♠=10 tricks. you still need two more to develop. Partner didn't promise K109x in diamonds to provide the valuable 2 more tricks.

But with xxx in hearts and lead from unfavorable side slam is rather optimistic. Even 5♣ might be too much while 3NT would be safer once your dummy's hearts holdings are not set well.

[Bbradley62]

georgi, on 2014-August-17, 14:31, said:

So by having 14HCP you should have in mind partner has maximum 13HCP.

Nitpicky, but... my understanding is that "best hand South" means that no other player has more HCP than South; it's entirely possible that someone (or more than one someone) has the same number of points as South.

Also, South has 18HCP on this hand, so North could certainly have enough for slam.

[georgi]

Correction about 14HCP. Might have looked at another link at the moment.

18HCP are in south's.