[cherdano]

jdonn, on May 3 2008, 10:56 AM, said:

cherdano, on May 3 2008, 11:51 AM, said:

jdonn, on May 3 2008, 09:49 AM, said:

Wow, I totally look forward to hearing why queen of diamonds might not be right. I can't think of any possible reason, but the mind remains open.

If I didn't have ♣Q, I might actually lead a low diamond. The reasoning is more or less:

If declarer has ♦A,K,9 between him and dummy, we will hardly ever set up diamonds anyway, basically the only case being where the suit is 5332 and parter has 3 (or 5422 obviously but in this case either lead will work). On the other hand, when partner has Hx and the suit is 5422, we need to lead low to unblock.

Here with the ♣Q being a possible second entry, and with both their hands sounding balanced the risk of leading low and giving up the 9th trick seem a little too big.

I hear reasoning like this a lot. Aside from that I don't see why 3-3-2 with partner having 3 is at all unlikely, I think this starts with the faulty premise that if we don't set up diamonds we can't set them. We have good defense and they could just be down always, why can't a low diamond lead be giving up trick 9 in a silly fashion? And all to cater to partner having specific doubletons with the suit 4-2 or 2-4 in the other hand, it has always seemed like a big reach to me.

In the situation with partner having xxx, he also needs to have an entry, and declarer has to misguess to give partner his entry before I get the ♠A.

[jdonn]

Jlall, on May 3 2008, 01:54 PM, said:

jdonn, on May 3 2008, 11:56 AM, said:

cherdano, on May 3 2008, 11:51 AM, said:

jdonn, on May 3 2008, 09:49 AM, said:

Wow, I totally look forward to hearing why queen of diamonds might not be right. I can't think of any possible reason, but the mind remains open.

If I didn't have ♣Q, I might actually lead a low diamond. The reasoning is more or less:

If declarer has ♦A,K,9 between him and dummy, we will hardly ever set up diamonds anyway, basically the only case being where the suit is 5332 and parter has 3 (or 5422 obviously but in this case either lead will work). On the other hand, when partner has Hx and the suit is 5422, we need to lead low to unblock.

Here with the ♣Q being a possible second entry, and with both their hands sounding balanced the risk of leading low and giving up the 9th trick seem a little too big.

I hear reasoning like this a lot. Aside from that I don't see why 3-3-2 with partner having 3 is at all unlikely, I think this starts with the faulty premise that if we don't set up diamonds we can't set them. We have good defense and they could just be down always, why can't a low diamond lead be giving up trick 9 in a silly fashion? And all to cater to partner having specific doubletons with the suit 4-2 or 2-4 in the other hand, it has always seemed like a big reach to me.

Don't you think it's true that if we can set up diamonds we are very likely to set them, and we are not that likely to set them if we cant? Sure you put all your eggs in the diamond basket when it might not be necessary, but the diamond basket looks by far like the best shot.

But the diamond queen lead sets up diamonds many/most of the times you need to set up diamonds anyway, so it's not like you are anywhere near giving up on that chance.

I'm not sure how to quantify it other than my gut, but I think they will be down a fair amount of the time even if we can't set up and run diamonds. Maybe 1 in 5 or so?

I actually think this would be a good hand for a sim. Despite my arguing I wouldn't be altogether surprised to be proven wrong by one. It would be eye-opening for me anyway.

[gnasher]

Jlall, on May 3 2008, 07:30 PM, said:

Doesn't this analysis totally ignore that by far the most likely case of beating this contract is when partner does in fact have a diamond honor? In other words the specific case where partner has Hx and one opp has 4 diamonds or pard has 9x+ and we lead the Q is far more likely to cost the contract than any other case that you mentioned?

You might be right if you were correct about the times when leading a low diamond gains. In fact, however, leading low when partner has 9x rarely works - declarer just ducks trick one, then if partner has one of our entries he won't be able to do anything useful with it.

The only time that you gain from leading low to partner's 9x is if you have two entries all of your own. The only situations I can think of where that would occur is if partner has ♣Kxx/Axx and declarer ♣AJ10/KJ10. On any other club layout, declarer can avoid giving you the lead.

[gnasher]

Jlall, on May 3 2008, 08:16 PM, said:

I've been trying to find a simulator for like the last 30 minutes and the few I did find that didnt require me to compile the code myself (sorry if this is really easy but it sounds really hard) etc were not compatible with vista (or they were and I messed up).

I don't think simulating the play will work, because that will involve double-dummy analysis. One of the ways that we might survive a diamond blockage is by declarer's misguessing which suit to attack first, so a double-dummy analysis will bias the results in your favour.

If you want a load of hands to look at and analyse by hand, I'll generate some for you. Let me know the range of 2NT, and whether opener is allowed to have five hearts, 5422s, etc.

[matmat]

deleted. suggestion wouldn't have worked.

[gnasher]

http://www.smohandes.pwp.blueyonder.co.uk/...ds20080503a.txt

Generated with Thomas Andrews's dealing program. Conditions:

west is "A32 T92 QJT32 Q2"

main {
# South is 20-21
if {[hcp south] < 20 || [hcp south] > 21} {reject}

# North is 3-11
if {[hcp north] < 3 || [hcp north] > 11} {reject}

# South has 4+ H, <4 S
if {[hearts south] < 4 || [spades south] > 3} {reject}

# North has 4 S, <4 H
if {[spades north] != 4 || [hearts north] > 3} {reject}

# South is balanced or 5H332
if {[balanced south]} {accept}
if {[semibalanced south] && [hearts south] == 5 && [spades south] < 4 && [diamonds south] < 4 && [clubs south] < 4} {accept}

reject
}

[Edit: changed the URL]

This post has been edited by gnasher: 2008-May-03, 14:59

[Halo]

Queen of diamonds.

[ArcLight]

I did a small simulation using Deal Master Pro (by Ed Marzo)

The ♦QJT was the stand out lead, resulting in the best score for the defense more than twice as much as any other lead.
The ♦x was next.

The worst lead?

♠ Ace

[gnasher]

Jlall, on May 4 2008, 05:11 AM, said:

I looked at some of the hands gnasher posted, and the DQ was indeed clearly right.

You were more diligent than me then. I looked at the first 15 or so, but didn't find a hand where it mattered, except in terms of overtricks.

[gnasher]

ArcLight, on May 4 2008, 05:08 AM, said:

I did a small simulation using Deal Master Pro (by Ed Marzo)

The ♦QJT was the stand out lead, resulting in the best score for the defense more than twice as much as any other lead.

That is to be expected, because a low diamond is likely to concede an overtrick more often than it defeats the contract. This was IMPs, however, where the size of the score matters more than whether it's the best score. I doubt that anyone would advocate a low diamond at matchpoints.

[gnasher]

Jlall, on May 4 2008, 05:11 AM, said:

Oh and a side note, declarer had AKx opp 97x in dummy and remembered to play the 9. Good careful play that I'm sure everyone thinks they would get right but I think few would always get that right.

I'm not sure that you should get it right. If LHO has Q8xxx or J8xxx and two entries, the suit is blocked unless you unblock it by playing the nine. The spot cards might or might not make that believable, but at best you'd have a nasty guess as to whether he has led fourth from QJ10xx or a non-standard card from H8xxx.

Edit: originally I mentioned only Q8xxx, but J8xxx is the same.

This post has been edited by gnasher: 2008-May-04, 12:34

[TimG]

Jlall, on May 3 2008, 11:11 PM, said:

Oh and a side note, declarer had AKx opp 97x in dummy and remembered to play the 9. Good careful play that I'm sure everyone thinks they would get right but I think few would always get that right.

Isn't the nine equivalent to guessing that you've led small from QJTxx? While if the lead is 4th best from a non-sequence it would be better not to play the nine?

[cherdano]

TimG, on May 4 2008, 09:51 AM, said:

Jlall, on May 3 2008, 11:11 PM, said:

Oh and a side note, declarer had AKx opp 97x in dummy and remembered to play the 9. Good careful play that I'm sure everyone thinks they would get right but I think few would always get that right.

Isn't the nine equivalent to guessing that you've led small from QJTxx? While if the lead is 4th best from a non-sequence it would be better not to play the nine?

No.

[jdonn]

I think I would have missed the 9 if I was declarer hehe, it's good to get this reminder about making plays like that. They are easy to miss at the table, I can think of two examples. Yesterday after the auction 2NT on my right, 3♥ transfer on left, 3NT (screw you and your spades) on right, all pass, I led 9 of spades from KQJ98, and declarer who had 10 fast tricks to run anyway forgot to play the ten with Txxxx in dummy and Ax in hand. I also remember a hand a couple years ago where I opened 1NT and played there, lho led something to RHO, and RHO returned a low card to dummy's 9xxx through my Jxx. I forgot to play the jack and indeed it was low from AKQxx to the singleton ten

[gnasher]

cherdano, on May 4 2008, 05:41 PM, said:

TimG, on May 4 2008, 09:51 AM, said:

Jlall, on May 3 2008, 11:11 PM, said:

Oh and a side note, declarer had AKx opp 97x in dummy and remembered to play the 9. Good careful play that I'm sure everyone thinks they would get right but I think few would always get that right.

Isn't the nine equivalent to guessing that you've led small from QJTxx? While if the lead is 4th best from a non-sequence it would be better not to play the nine?

No.

Why?