When god gives you QJT...

[gnasher]

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]

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.

[Guest Jlall]

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.

I didn't want to do double dummy analysis, I'm willing to just take hands and look at them. 20-21, 5cM possible if 5322.

[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]

Edited May 3, 2008 by gnasher

[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

[Guest Jlall]

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

 

Thanks! Glad to plug a leak in my game. Weird it was kinda split when I asked pros at the tournament. Maybe this is just trying to be too fancy, or maybe it's because people are less likely to lead the DQ when posed as a problem (since they think something else is right). FWIW Migry led the Q at the other table.

 

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.

[gnasher]

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]

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]

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.

Edited May 4, 2008 by gnasher

[TimG]

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]

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]

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?

[cherdano]

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?

If the holding is 4th best not from QJT, then it usually doesn't make a difference what to play. Your example (or J8xxx) seems to be the only where it does, even there it depends on the whole hand (and declarer may have been able to rule out this holding from the spot card led).

So "While if the lead is 4th best from a non-sequence it would be better not to play the nine?" is incorrect.

[Free]

God never gives me any cards, it's the dealer who does it...

 

When I have such holding I usually lead small. There are 2 ways you can win from it:

1. partner has Ax, Kx or 9x and the suit splits 4-2

2. partner has 8x, dummy holds the 9 and the suit splits 4-2

 

In the second case, declarer usually plays low :)

[pclayton]

Blame this whole underleading the QJT on Kantar :)

 

I think its right if RHO naturally bids the suit however and we choose to lead it.

[whereagles]

I think its right if RHO naturally bids the suit however and we choose to lead it.

Yeah, most textbooks say to underlead from 3 or 4 top honors only if RHO bid the suit.

[gnasher]

Another factor to consider is how many entries you have: having two entries would make a low diamond lead more attractive, because now you could hope to gain opposite 9x or singleton king.

[cherdano]

Another factor to consider is how many entries you have: having two entries would make a low diamond lead more attractive, because now you could hope to gain opposite 9x or singleton king.

I disagree, if you have two entries you can lead the Q and still establish the suit when partner has Kx and it splits 4-2. I think having one entry is a stronger indicator for leading low.

[gnasher]

I disagree, if you have two entries you can lead the Q and still establish the suit when partner has Kx and it splits 4-2. I think having one entry is a stronger indicator for leading low.

With one entry, leading low stands to gain when he has Kx, but not when he has 9x or worse. If he has a singleton king, leading low avoids blowing a trick but doesn't allow you to set up the suit.

 

With two entries, leading low stands to gain when partner has 9x and possibly 8x, but not when he has Kx. If he has a singleton king, leading low not only avoids blowing a trick but also allows you to set up the suit.

 

I could understand your saying that there's not much in it, but why do you say that having only one entry makes you more inclined to lead low?

[Winstonm]

What would be of benefit would be to find out how often low would blow a trick as to how often it was necessary to lead low to defeat the contract.