[wank]

---

(I thought this was the bidding, but maybe partner wasn't a passed hand after all. When I played the board it was also 1NT-3NT and my partner never passes 12s, even filthy ones)

Anyway, I got into a debate over this hand.

The world class declarer got the 4♥ lead (opps play 4th highest) which went to the 9, 10 and Jack. Seemingly, LHO has underled AK to a number. Taking the diamond finesse to establish a sure 9th trick is perforce dangerous as they may cash 4 hearts and a diamond. You can try to drop the Q♠ first so you don't need the diamond finesse, but then you set up spade tricks and risk going off when hearts were 4-3. Returning a heart to see how they break is problematic as you will need to discard a spade off Jxx anyway if they're 5-2.

As we all know leading a high one from AKxx is theoretically superior in isolation. Declarer was of the opinion that's is so superior that one should never underlead on an auction such as this and that no strong player would ever do so. The implication being then that if it is a low lead from a strong player [this wasn't a strong player, but I'm more interested in the theory than what you do on this hand], you know the suit's 5-2 so you would play for the ♠ drop.

On the other hand, I said the information leak which would allow declarer to get this hand right was too great and that one should therefore underlead AKxx on occasion. He dismissed this saying you would't gain enough by disguising your hand to offset situations where the suit lies AKxx, Qx, 10xxxx, Jx and so on. It seems to me that you're bound to underestimate the effect of disguising your hand as the results are very intangible, whereas the success from the honour lead is obvious.

Comments?

[FrancesHinden]

I'm not sure it's really addressing your primary point, but I have seen (at least) two of the top English players leading low from AKxx on this sort of auction in the last year. I don't think they were doing it for 'information leak' purposes either.

[chasetb]

I don't see any "information leaks" by leading low from AKxx, I have done it more than a few times and I'm not WC. However, underleading AKxx at NT has been the best thing for me more often than not. You also don't necessarily have to throw off a Spade. You decide either that the ♦K is onsides and you will finesse it, or that it is off and so you are trying to drop the ♠Q. There's even a chance that someone has ♣Txx, in which case there might be a squeeze on.

[JLOGIC]

Agree with your WC friend, leading low from AKxx would be so strange/bad from a good player I would assume AKxxx.

Leading low from AKxx sometimes in order to do what in poker is known as "balancing your range" is pretty absurd. If you only do it something like 1 in a thousand times, you are not gaining anything, since declarer could just ignore it as it is not frequent enough for it to change the odds for declarer. On the other hand, if you are doing it often enough to affect declarers play on this hand, you are now making an inferior lead a significant amount of the time simply so that they don't know that you have AKxxx when you lead low. Do you really think that a situation like this where it matters comes up often enough to recover your losses? I mean, in this case you gain when you catch this exact layout, and you have AKxxx, and there is Qx of spades. There are too many variables in bridge to ever make an inferior lead for the sake of balance.

Even in poker, balance is extremely overrated. Do you think you should play 74 offsuit under the gun even though it's a losing hand so that you can represent the nuts/2 pair on 653 or 742 boards? Of course not.

[mich-b]

What you are asking is equivalent to the suggestion that one should sometimes pass (as dealer) a 13 hcp hand, because otherwise if you do PASS as dealer , and later end up defending , the declarer may draw the inference from your Pass that you don't have an opening and place a critical honor in your partner's hand.
I think we all agree that this will not produce very good results. We try to make our best choices during the bidding (and the lead) and if an inference from this works against us - we live with it.

[hatchett]

FrancesHinden, on 2011-November-05, 12:03, said:

I'm not sure it's really addressing your primary point, but I have seen (at least) two of the top English players leading low from AKxx on this sort of auction in the last year. I don't think they were doing it for 'information leak' purposes either.

Maybe this explains why England weren't playing in the Bermuda Bowl.

[FrancesHinden]

One of them has won a world championship.

[FrancesHinden]

FrancesHinden, on 2011-November-06, 15:54, said:

One of them has won a world championship (and I am not counting the venice cup here).

[benlessard]

Agree with your WC friend and Justin. There is no need for mixed strategies in bridge since there is a huge range of possible hands. Only when the possible range of the hand is severly limited you need to avoid pure strategies and apply mixed startegies.

Even in holdem mixed startegies are in very specific situation. Most of them are AK offsuit.

[rhm]

mich-b, on 2011-November-06, 02:43, said:

What you are asking is equivalent to the suggestion that one should sometimes pass (as dealer) a 13 hcp hand, because otherwise if you do PASS as dealer , and later end up defending , the declarer may draw the inference from your Pass that you don't have an opening and place a critical honor in your partner's hand.
I think we all agree that this will not produce very good results. We try to make our best choices during the bidding (and the lead) and if an inference from this works against us - we live with it.

I think this is the tendency but too simplistic. Of course nobody passes say a 20 HCP hand to avoid information leak.
Nevertheless this remains a contentious area.
For example Fred believes mini notrump is not a sensible way to play because if you pass you give too much inferences to a competent declarer.
Other actions you may take in the bidding may give you little chance of buying the contract profitable but pass a lot of information to a competent declarer.
An example might be preempting in fourth position after both opponents have already bid.

To take an example from recent high level play

Zia had to play 6♠ against Zimmermann on the following deal:

---

He got the lead of ♦K and a ♦ continuation.
Zia ruffed came to hand with the ♥Q and ran the ♠9
After 3 rounds of spades, he ruffed a ♥ back to hand and drew the last trump with the ♠K discarding the ♣7 from dummy
Well played by Zia.
But I wonder whether Zia would have found this line if East would not have bid 4♦ at his second turn.

Rainer Herrmann

[benlessard]

These are unrelated cases.

a quick and simplistic definition is

-- a mixed strategy is when you need to randomize otherwise the opponent with a specific counter strategy can profit from your lack of randomizing.

-- pure strategy is when you dont need to randomize and its possible for you to have a specific strategy where even if the opponent know your strategy he cannot gain from it//


If you play 10-12 NT and all and only the hands that fit in that range you open 1NT, its a pure strategy. The information you give to the opps is a loss that you occur (wich mean that its a possibly bad pure strategy) but it has nothing to do with randomizing/ balancing your range/ nor mixed strategies.

If you say that with with a specific case you need to lead high a % of times and in the same case you also should low a % of times than its a mixed strategies. Because if with that specific case you play high 100% and play low 0% the opponenet can figure it and adapt and put you at a disadvantage.


I believe there are case that leading low is better but there are quite rare. So there are case where its better to lead high there are case where its better to lead low but there is no case where you must randomize to decide to lead low or high. This is beacause your "universe" of case is immense and unpredicatable. But in holdem after 3 raise your universe of hand can easily be AK/AA/KK/QQ at this point using the correct pure strategy EX folding QQ + AKo// calling KK AKs and raising AA is easily exploitable.


So since your "ensemble" of possible hand is limited you need to randomize the correct strategy will look like

fold QQ fold AKo,
call with KK with some AKs and with some AA (slowplaying a few AA preflop to protect KK is questionnable)
raise with some AKs & with some AA