[Poky]

Let's say the opponents play a specific system, with your RHO opening:
1♥ = 5+♥, 11-22
.... 3♦ = exactly 3 hearts, invitational
3♠ = Asking bid: "Do you have a shortness?"
.... 3NT = "I do have an unspecified singleton, but not a void"
6♥

You are on the lead with:
♠Jxxx
♥xxx
♦Axx
♣Qxx

What would you lead here and why? Think a bit about it before reading further.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.


The board that made me think about this lead problem was the following declarer problem:

♠x
♥J9x
♦KQT9x
♣Kxxx

♠ATx
♥AKQ8x
♦xx
♣AJT

Same bidding, 6♥ contract, ♣Q lead.

My reasoning was: since nowadays in practice the ♣Q lead is almost exclusively a singleton (meaning: LHO is trying to find RHO with the ♣A, which suggests he doesn't hold the ♦A), I would take trumps and finesse the ♦J right away, hoping for two discards on diamonds.

If LHO started with something similar to:
♠Qxxxx
♥xxx
♦Jxxx
♣Q
this line is going to win.

This pointed out the following eventual flaw - declarer can always exploit leads of that kind if the range of such "Q without J" leads is composed exclusively (or close) of singleton queens. That fact suggests the defender could have a better expectation if he used a mixed strategy instead:
n% - singleton queen,
(1-n)% - non singleton queen
when choosing to lead from an unsupported queen.

Now the big question arises - in this same scenario, can LHO afford to have such high (1-n) frequency of leading non-singleton queens to make declarer indifferent between the:
- finesse ♦J, and
- ruff spades using club as entry
lines, minimizing that way declarer's expectation of winning the contract?



Let's return to the main lead problem. When the leader looks at:
♠Jxxx
♥xxx
♦Axx
♣Qxx

he can make some sound conclusions from the bidding:
- declarer didn't ask 4NT because he's got aces.
- declarer doesn't have ♦KQx+ or ♦Kxx+ because a singleton in diamond in dummy would be a bad feature for jumping to 6♥ and he wouldn't do that with these holdings,
- declarer holds similar holdings in both spades and clubs, which are suitable both when a singleton or some values are across.

From that we can deduce his hand should be something like:
- 3613 with both black aces, or
- 4612/2614 with both black aces, or
- 5S(332) with both black aces and nothing in diamonds (rarely).

Since the dummy might hold ♦KQx+ (providing two discards via Morton's fork coup when declarer is singleton), it appears like the leader has to attack the weak opponent's fragment to cash a trick while retaining the ♦A. Since the leader holds the Q in clubs but just the J in spades, this fragment will more often be in clubs (if anywhere). Therefore, declarer will lead clubs hoping to score a club trick (when partner has the ♣K) and the ♦A. Once he decides to lead clubs, he might as well lead the Q from ♣Qxx since:
- that might make declarer misplay when partner holds the ♣J (important when clubs are 3-3),
- that might win against some "laydown" boards like this one that I presented (when leader's partner held nothing but a piece of trash like ♠KQxxx♥xx♦Jxx♣xxx).

Of course, since the ♣Q lead might occasionally give the contract away, we should know the exact frequencies, equities and expected values to be sure if such mixed strategy is the game-theory optimal play in some specific spots (like this one, for instance), but we can hardly do that without having a perfect single-dummy robot who could analyze spots and play appropriately. Until that we can only base our thoughts on experience and intuition.

However, that doesn't prevent me to ask the question - what is your feeling?

If we had a super-robot (say, a sophisticated neural network) which both plays optimally and calculates optimally (taking with proper weight in consideration all the possible information - mostly opponents' ranges and all the Bayesian inferences available), would such machine pick here the unprotected queen of clubs as the best lead, or would the danger of such lead nevertheless outweigh the eventual profits arising from the mixed strategy consideration, neglecting the possibility such plays might be ever a part of the GTO strategy - making such creativity in leading generally avoidable?

[SteveMoe]

Usually when you are weak, need to hold the lead at trick 1, and do not want partner returning that suit. Also when leading to partner's suit through suspected strength in Dummy and no outside entry to your hand. Not sure why this is a "Zia lead".

[Poky]

SteveMoe, on 2015-July-23, 20:42, said:

Not sure why this is a "Zia lead".

Zia leads an unprotected honor quite often, at least compared to how other experts lead in similar spots.

[Free]

Poky, on 2015-July-24, 03:03, said:

Zia leads an unprotected honor quite often, at least compared to how other experts lead in similar spots.

Zia knows what he's doing, we don't