[gwnn]

Jxxx
Ax
Axxx
Qxx

(x means 7 or lower)

Not a great collection to lead from but it goes
1NT-2D (hearts)
2NT-3NT
all pass

2NT showed "a max without a heart fit" so you can imagine that this is not the bermuda bowl final, but you still have to lead something.

[helene_t]

Spade

[Free]

Yep, ♠

[mikeh]

Years ago a very strong player analyzed opening leads against 3nt, and the results were cited in the Bridge World. Iirc, Jxxx, where the highest x was an 8 or lower was about the worst possible lead, which makes sense to me given how that lead will often blow a trick and/or a tempo. I think it even more obviously wrong when partner rates to hold a weak hand. We need partner to hold something, but if he has, say, a K, then he won't have much else. I opt for a low diamond. It isn't that I like the diamond....it is that I dislike it less than all the other choices.

[masonbarge]

gwnn, on 2015-February-17, 02:24, said:

Jxxx
Ax
Axxx
Qxx

(x means 7 or lower)

Not a great collection to lead from but it goes
1NT-2D (hearts)
2NT-3NT
all pass

2NT showed "a max without a heart fit" so you can imagine that this is not the bermuda bowl final, but you still have to lead something.

At IMPs, you have to assume partner has a card, since if he doesn't, declarer will make 4 or 5 depending on the heart situation.

Assume partner has a heart trick, as he rates to have 4 of them. It seems to me your big chance is that dummy has KQxxx or KJ10xx. That's still only three tricks, and he's making 3NT even if you get your ♣Q.

Low from Axxx in an unbid suit is worse (or should I say even worse) than Jxxx in terms of possibly giving away a trick. But it seems less likely to give away the suit - you get a major fail only with four diamonds to two honors in dummy, possibly giving away a trick you would have gotten with the club Q.

Even with the slight preference for a major suit - unlikely that dummy has 4 spades! - I'm leading a low diamond at IMPs. It looks like the best chance to set the contract, although not a high percentage shot. You also have the possibility of killing dummy's only outside entry.

At matchpoints the story changes. I want to lead passively, and a low spade is somewhat less terrible than a low diamond in terms of giving away a trick.

I really wouldn't think about it much. I'd lead a low ♦ at IMPs, which just looks like the best attacking lead, and a low ♠ at matchpoints, which looks like the least dangerous of the four suits.

[BillPatch]

I simulated the hand 200 times setting the leading hand as J732 A2 A732 Q32, the opener hand as balanced or semi-balanced with suit lengths from 2 to 5 and 16 to 17 hcp, the responder as 0-3 spades. 5 hearts. and 0-4 in the minors. no limits on the other defending hand. J dealing out the hands to a pbn file. Lead Captain was used to analyze the leads double dummy using Hoagland's Double Dummy Analyzer and gather stats.
Results:
S 2,3 sets 10.5% defensive tricks 3.19
S 7 sets 10.5% defensive tricks 3.18
S J sets 8.5% defensive tricks 3.07
D A sets 8.0% defensive tricks 3.05
H A sets 6.5% defensive tricks 3.04
D 2,3 sets 6.0% defensive tricks 3.05
D 7 sets 6.0% defensive tricks 3.04
H 2 sets 6.0% defensive tricks 2.83
C 2,3 sets 5.5% defensive tricks 2.84
C Q sets 5.5% defensive tricks 2.80

Analysis: Low spade best at both IMPS sets contract most often, and at MPs has the greatest average defensive tricks. Next are J of S and A of D at IMPS. The other cards bunch at 6.5% to 5.5% sets at iMPS with the low H and all C lagging at MPS.

[masonbarge]

Nice job Bill. Trusting your methodology, hehe, the unbid major rules again!

[kenberg]

Before reading all of this I was going to lead a diamond. I can be very stubborn so maybe I still lead a diamond. But I am impressed with the analysis. My thinnking was that I could get a long diamond and the ace of diamonds. With the hearts not splitting for declarer, maybe we get three others before he gets nine. Computers are really smart, but Gib has also been known to make some really weird plays.

[gwnn]

Actually, declarer had 5 spades and a singleton heart (!). Since they had a 5-3 spade fit, I gave away the suit with my spade lead. Partner, who happened to have KJT8x in hearts but was obviously not affected by the actual deal, said that the lead of the ace of hearts is "automatic" because "declarer has at most 2 hearts" and "probably spade length" and did not budge. We ended up both annoyed as is often the case. A few deals later, I made a takeout double of 1♣-p-2♣ and when they ended up in 3NT, my partner led an "automatic" club from xxx xxx xxx J8xx, giving away the whole suit. Anyway thanks for the replies! All I wanted to know is if anyone at all would consider the ace of hearts.

[Mbodell]

I'm actually pretty shocked that the double dummy solver doesn't prefer one of the two A. Usually there is a strong A bias.

[BillPatch]

Yes, the double dummy analyzer often selects A's without K's for leads against 3NT, and often partner's signal and the sight of dummy still provide nearly as difficult a problem single dummy at trick two as the initial lead problem. Leading the A of ♦ allows the opening leader the chance to attack either long suit on next round at the cost of possibly establishing possible strength in the opening NT bidder's probable diamond suit. The A of ♥ also allows opening leader to chose the best continuation according to dummy and signal, but it gives up a tempo in establishing dummy's known long suit, so it rates to be weaker than the ♦ A. Since, assuming a rational bidder, dummy does not rate to have a 4 card ♠ because he avoided Stayman, and opener might prefer to open with one of a suit to show his spades at the one level even with the balanced hand(the tendency would be more pronounced with less balanced distributions.) If the opponents bid as badly as the actual opponents your superior talent will beat them over 95% of the time without changing your tactics as long as you avoid unforced errors. Assuming that partner, who made a TO double emphasizing the three unbid suits, has instead the super strong hand, after the opponents show great strength by bidding to 3NT, is surely an unforced error. That client had better have a big wallet!

[kuhchung]

Mbodell, on 2015-February-20, 02:30, said:

I'm actually pretty shocked that the double dummy solver doesn't prefer one of the two A. Usually there is a strong A bias.

Maybe that's how bad it is!!

I can't believe it's still better than a diamond bleh

[BillPatch]

Finally back with better stats due to larger sample size: 2001. Now using DealMasterPro to generate deals, using Lead Captain as before to run Bo Hoagland's DDA. Much better automation with DealMasterPro.
S 2,3 17.94% set 3.52 average defensive tricks
D 2,3 17.69 3.52 tricks
S 7 17.59 3.51 tricks
D 7 16.89 3.49 tricks
D A 15.79 3.45 tricks
H A 13.89 3.30 tricks
S J 13.79 3.33 tricks
H 2 9.95 3.21 tricks
C 2,3 9.60 3.21 tricks
C Q 8.15 3.12 tricks
Now low spade and low diamond are the two best leads, the low spade continuing to be slightly preferable at IMPs, where setting the contract percentage is our proxy. As the original problem was IMPS, still a victory for the unbid major! (Significant at the .05 level at IMPS) The two Aces fall behind the low pointed suit leads.
Note: continued to allow opening NT on all 5422 hands. Sample of 325 with purely balanced NT openers showed similar results.