[hanp]

Double dummy results are awful for determining the best opening lead imo.

[bluecalm]

Quote

Double dummy results are awful for determining the best opening lead imo.

Why do you think so ?
I find them very accurate. I spent a lot of time comparing best dd leads to actual leads made by world class players (mainly Meckstroth and Versace) and what I found is this:
a)in most situations they are the same leads
b)in hands where they are not the same dd leads would be better on average.

Of course those comparisons can be made only for hands with simple, well defined auctions and maybe I still suffer for too little sample size (only about 200 hands analyzed with simuls) but my experience with simuls taught me to trust dd results more than judgment of any player.

[cherdanno]

hanp, on Jun 10 2010, 10:48 AM, said:

Double dummy results are awful for determining the best opening lead imo.

But usually they don't lead to such huge differences.

[Rodney26]

I think the heart is obvious from jump street; lead the shorter major on this auction with a couple stray jacks and where partner could not bid. Lead toward partner's length, not yours, when you have garbage. The books don't stress this enough.

The more interesting question on this auction is what to lead if you switch the clubs and spades; I don't know how double dummy analysis is going to validate the inference that Stayman wasn't used, or the opponents style in using Stayman when 3433 or holding 4 low in a major. I expect the heart is still superior especially since a club keeps coming up as the worst choice (for a beat) in all these scenarios.

One other interesting question; are we barred from leading a heart if partner takes a little more than 10 seconds to pass 3NT? Does a committee tell us we have to lead the obvious club?

[hanp]

By leading a heart here we can easily take away one of declarer's guesses. But double dummy, declarer would guess right anyway so it doesn't cost.

By leading a strong suit or an honor sequence we often give information that may help partner with the defense. Of course double dummy that is irrelevant.

Double dummy tends to warn against leads away from honors. Better to lead another suit, our double dummy partner knows to switch to our strong suit anyway.

Look for example at the xxx A109 xxxx xxx hand where Phil suggested a double dummy analysis. It is quite likely that double dummy analysis favors a diamond lead. But in real life, if we lead a diamond, partner often won't expect that we have such a strong heart suit!

[mikeh]

I did my own rough simulation. I don't like DD analysis, since nobody leads or plays dd: I have never seen deep finesse make the losing but percentage play: it always makes the winning play: thus it never loses to 5-0 trump splits against silent opps, and never gets a 2-way hook wrong when LHO counts out to hold 4 cards in the suit and RHO only 2, and so on.

So I ran 70 hands with the constraints being dealer was balanced and 12-14, and was allowed a 5 card major but not 5422, and responder was 13-17, no 4 card major. This tilts the deal in favour of a major suit lead since I didn't cater to responder being 4=3=3=3 or 3=4=3=3, either of which would tend to minimize the effectiveness of a lead in the 4 card suit.

I also used 13-17, since many 12 counts will prefer to invite. But a quick check with 12-17 seemed to suggest no real change.

Anyway, not surprisingly (to me) the contract was usually cold on any lead. I didn't analyze overtricks..if it was clear that declarer, through normal play, would score 9 tricks, I moved on.

I found 2 hands on which the contract simply failed on any lead.

1 on which a diamond lead was best

1 on which a spade was the winning lead

1 on which either red suit would win

1 on which either clubds or hearts would win

5 on which a heart was needed and 5 on which a club was needed

So the hand made 50 out of 66 times.

This is a small sample, but one example of how dd can go right where no declarer would was on a hand on which to beat the contract, the defence needed to get a club trick and then be able to switch to hearts. Declarer held Axxx opposite Qx in dummy. Unless one was playing coded honour leads (where the J denies a higher honour....which I played for several years a long time ago and stopped precisely because of this type of issue), declarer will always play the Q, and then has to duck the Ace. Double dummy, he wins the A on round one and then establishes the diamonds, depriving us of the tempo. An immediate heart lead left the defence a trick short.

[andy_h]

Some past threads here and here

[phil_20686]

The most relevant issue about leading a heart is that it will probably remove the guess in every suit once its revealed that you made a short suit lead. While removing guesses is an issue, a lot of the time declarer simply doesnt have much choice on how to play a suit. Partner may just have enough entries to establish even quite a poor heart suit. QJTxx hearts and a couple of entries could easily be the case. Especially when your hearts are 98. Those are big pips. Even when declarer has 4 hearts it could easilly be that partner has KT7xx over Qx or Jx in dummy say. Its true that partner could just have KQxx clubs, but the chances are that when partner does have club values he has enough minor suit cards to be able to hold up and kill your suit.


It would be interesting to see how peoples opinions chance if this was an invitational auction. I think i would be much more likely to go passive with j of clubs now.

[Phil]

hanp, on Jun 10 2010, 11:06 AM, said:

Look for example at the xxx A109 xxxx xxx hand where Phil suggested a double dummy analysis. It is quite likely that double dummy analysis favors a diamond lead. But in real life, if we lead a diamond, partner often won't expect that we have such a strong heart suit!

http://www.bridgehan.../Smith_Echo.htm

[Rodney26]

I think "removing the guess for declarer" is not a good reason to avoid leading a heart.

1. There may not be a guess in the suit.
2. If there is a guess, declarer is slightly more likely to finesse into this hand to protect his honors from attack anyway.
3. The lead might take partner off an endplay even when you strike silver rather than gold (QT7x).
4. If we're going to beat this, partner is going to need both strength and probably a source of tricks (a 5+ card suit). Partner is probably going to get endplayed a couple times anyway with a balanced or semi-balanced 11-12 count so the beat is unlikely.
5. While the club lead is termed "safe," it is also likely to be partner's shortest suit and is most likely to put partner under exiting and discard pressure immediately.

[foo]

Phil,

First, thanks for the welcome back.

The reason I explicitly broke the situation into 3 possible catagories is that, at least to me, it would be interesting to know ~ when the transition occurs between the 3 defense strategies.

Just how much weaker than partner should you be before it rates to be better to a) help them establish their hand rather than trying for a mutual defense or c) to establish your own?

The extreme cases are clear. If you have no entries, or partner rates to have none, then what you should do is obvious.

But things get far more complex in a hurry when the HCP are not so unevenly split.

The OL is by far the most difficult card to "get right" in Bridge.
So it seems worthwhile to spend some "skull sweat" looking for rules to improve their likelihood of being right (or at least of not being disasterous!).

[mikeh]

I did another simulation, using responder hands with no 4 card major and 12-16 hcp this time.

I ran 200 hands and had deep finesse do a dd analysis. 154 hands were cold double dummy and I didn't look deeper into those. I am sure that some of them would have failed at the table, played by the world's best real life declarer, but I only have so much time.

Of the 46 that failed, the most effective lead was a heart, which surprised me a little. It was the only winning lead 17 times, while a club was the only winning lead 7 times, and a spade 6 times.

However, when I looked at some, but not all, of these hands, the flaws of this approach became obvious.

On one hand the only reason that the heart was the only working lead was that deep finesse dropped partner's stiff diamond King when they hold a 10 card fit: in the real world, few declarers would get this right.

On another, the spade was the only dd lead but in real life, a club would also work: would declarer really fly with the A from Axxxx in dummy and Qx in his hand? He only had 7 red winners and was wide open in spades.

Another 'win' for the heart lead found partner with Axx AQJ10xxx Kx x; and I suspect that many players would find some non-pass unless red v white.

Having said that, this simulation means that I have come around to the view that the heart lead is probably best. This is true, it seems, only in partnerships where partner cannot make a lead-directing double. On quite a few hands, it seemed to me that partner had a good double.

But: if the double simply says lead your shortest major, it would backfire on a few hands. On one, where a spade lead was required, partner held AKQJx in spades...and might well make a speculative double.

There is a lead-directing double over this auction that specifies a heart lead (I think it is called a Fisher double) and that would help a lot on the heart lead hands. Playing that sort of double makes the club/heart decision too close to call based on what is now about 260 hands.

[peachy]

Rodney26, on Jun 10 2010, 11:06 AM, said:

One other interesting question; are we barred from leading a heart if partner takes a little more than 10 seconds to pass 3NT? Does a committee tell us we have to lead the obvious club?

If any committee thought "the obvious club", then that group of people [don't want to all it a committee any more] are incompetent in bridge judgment. Club is the least effective, typically finding partner's shortness. What is the point of leading a long suit if the opening leader does not have entries, not even one, to first develop and then run it? This is IMPs, we want to find the best lead to set them. If partner has spades, he probably has only four, but if he has hearts the probability goes up that he has five - with one or more entries. I still think heart is the best and too bad the sims were introduced, with dubious parameters, before people had a chanve to think about this lead a little more. I would also lead a heart with 32 instead of 98.

[Free]

Can someone please do the same simulations but with ♥32 instead of ♥98 please?

[gnasher]

peachy, on Jun 11 2010, 06:50 AM, said:

If any committee thought "the obvious club", then that group of people [don't want to all it a committee any more] are incompetent in bridge judgment.

If any committee member thought the relevant question was "What is the obvious lead?", he would certainly be incompetent.

The right question (or one of them) would be "What are the logical alternatives?" I can't imagine anyone thinking that a club lead isn't one of those.

[bluecalm]

Quote

Can someone please do the same simulations but with ♥32 instead of ♥98 please?

This is not relevant.
I started work similar to Mikeh (reviewing hands where we can set a contract). ♥ looks the best but not that good as in simul (not twice as good as ♣) as it seems there are significant number of hands where partner is required to make very good switch or when he could have made lead directing double.
On the other hand on many hands we just need to lead heart because we won't be in again and there is an honour in dummy which requires a play from our hand.
I am not interested in spending any more time on this though so I just skimmed through like 30 hands and stopped.

[Rodney26]

gnasher, on Jun 11 2010, 05:19 AM, said:

peachy, on Jun 11 2010, 06:50 AM, said:

If any committee thought "the obvious club", then that group of people [don't want to all it a committee any more] are incompetent in bridge judgment.

If any committee member thought the relevant question was "What is the obvious lead?", he would certainly be incompetent.

The right question (or one of them) would be "What are the logical alternatives?" I can't imagine anyone thinking that a club lead isn't one of those.

Let's say though for argument's sake that as leader you hear 3NT on your left, which you recognize as the likely end of the auction. You mentally pick the heart 9 right away....and then partner trances. Maybe the most ethical thing to do is use this as an opportunity to educate partner not to trance. The opponents probably are going to feel like they were snookered no matter when you picked the lead. Still, the trance was truly irrelevant in this instance.

[gnasher]

Rodney26, on Jun 11 2010, 02:44 PM, said:

Let's say though for argument's sake that as leader you hear 3NT on your left, which you recognize as the likely end of the auction. You mentally pick the heart 9 right away....and then partner trances. Maybe the most ethical thing to do is use this as an opportunity to educate partner not to trance. The opponents probably are going to feel like they were snookered no matter when you picked the lead. Still, the trance was truly irrelevant in this instance.

The trance might be irrelevant to your opinion of what the right lead is, but it's very relevant for determining what the rules allow you to do.

If partner's trance suggests a heart lead over a club lead, it is illegal for you to lead a heart. That applies regardless of whether you already knew what you planned to lead.

[mikeh]

gnasher, on Jun 11 2010, 09:28 AM, said:

Rodney26, on Jun 11 2010, 02:44 PM, said:

Let's say though for argument's sake that as leader you hear 3NT on your left, which you recognize as the likely end of the auction. You mentally pick the heart 9 right away....and then partner trances. Maybe the most ethical thing to do is use this as an opportunity to educate partner not to trance. The opponents probably are going to feel like they were snookered no matter when you picked the lead. Still, the trance was truly irrelevant in this instance.

The trance might be irrelevant to your opinion of what the right lead is, but it's very relevant for determining what the rules allow you to do.

If partner's trance suggests a heart lead over a club lead, it is illegal for you to lead a heart. That applies regardless of whether you already knew what you planned to lead.

This is a point that many miss. They feel that a ruling against a call or play they 'would have made without the UI' is a finding that they acted unethically. The ruling is made without ANY imputation of improper motive. It can be difficult to accept that as an individual but it is central to the application of the laws.

[billw55]

Free, on Jun 10 2010, 09:06 AM, said:

It's interesting to see simulations, but I just wonder why a ♥ is so much better.  Is it because we have ♥98 and not ♥32?  Or is it because our ♠ aren't more solid like T982?

I'm stumped too. Why is a 2 card major so much better than a 4 card major ???