A paradox

[dburn]

A deal from the European Championships in Pau revolved around the play of this suit for no loser:

 

♥QJ93 ♥A642

 

North actually held ♥K8, and both defenders knew that East-West had a 4-4 fit. It seemed to me that if West were declarer the contract ought to make, while if East were declarer the contract ought to fail. How does it seem to you?

[pclayton]

So you are implying that North would cover dummy's last honor giving East a losing option of pinning 10x, but with the ♥QJ93 hidden that North would cover the Q?

 

I don't think its always so automatic for ♥K8 to cover with the A642 in dummy. With QJTx it doesn't matter, and declarer is playing the suit strangely if he holds Q-J-x-x. With QT9x it doesn't matter if we cover or not, so the relevant case is when Declarer holds QJ9x.

[Sambolino]

this is very interesting position, quite common too. i saw steve weinstein (not actually sure if that was him) on a recent vugraph dropping K10 sec.

[dburn]

declarer is playing the suit strangely if he holds QJxx

How would you suggest that declarer (West) play the suit for no loser holding QJxx in his hand? Of course, legitimately he cannot, but we do not always arrive in the best contract, and at least against you he has a chance...

[Sambolino]

i like this. everything is questionable except K10 sec

 

if W is declarer, N should (although this is questionable too) cover with Kx and K10. so that's something. that would imply 2/3 chance for finessing 10

 

if vice versa, N has a choice to duck. sooo... it's late and i'll be enjoying this tomorrow.

[pclayton]

declarer is playing the suit strangely if he holds QJxx

How would you suggest that declarer (West) play the suit for no loser holding QJxx in his hand? Of course, legitimately he cannot, but we do not always arrive in the best contract, and at least against you he has a chance...

Well you haven't shown the whole hand. Is it completely evident for the defense that declarer has three other sure losers?

 

I would think if declarer is really desperate to play QJxx opposite Axxx for no losers that declarer starts with the Jack, not the queen. Declarer doesn't want a cover. Most defenders won't cover the J for fear declarer has JTxx.

[dburn]

I don't recall mentioning what the contract was, so I am not sure whence this talk of "three other sure losers" comes. You may assume for the purposes of the exercise that declarer is in a grand slam (this was not actually the case - he was in a small slam with this suit as a side suit, but it was evident that he could not afford a loser in it since he had already lost a trick).

[pclayton]

I don't recall mentioning what the contract was, so I am not sure whence this talk of "three other sure losers" comes. You may assume for the purposes of the exercise that declarer is in a grand slam (this was not actually the case - he was in a small slam with this suit as a side suit, but it was evident that he could not afford a loser in it since he had already lost a trick).

For some reason I assumed the contract was 4♥. Maybe I saw the deal.

 

What was the contract, by the way?

[benlessard]

These are complex problems.

 

 

There is 3 relevant holdings.Depending if declarer play the suit for 0 or 1 loser.

 

QJxx

QJ9x

JTxx.

 

With JTxx declarer can play the J to induce a cover from Kx or can play A hoping to pick a stiff honor in LHO or pick up KQ. Playing the A is better. So JTxx is out (and if declarer cannot afford a loser than this case is eliminated right from the start)

 

So what remain is QJxx & QJ9x and both have the same probability.

 

QJ97,QJ95,QJ93

QJ75,QJ73,QJ53

 

There is some psychology here since with QJ9x some declarer will be tempted to lead the Q while with QJxx he might try the jack. But this doesnt hold in a strong level since all four player should assume declarer has QJ.

 

So now the question is declarer lead an honor (Q or J doesnt matter). Should LHO cover with Kx.

 

Obviously with KT & K9 he will always cover. But with K7,K5,K3 ? If declarer always cover with KT and never cover with Kx declarer will always play to drop the KT. But at the same time declarer will have QJxx as often then QJ9x. And with QJxx he will run for an honnor 100% of the times. So LHO should always cover since QJxx is as likely then QJ9x. And if LHO always cover then declarer should always finesse the ten. This is assuming the 4-4 fit is known.

[Guest Jlall]

With QJ9x in dummy against best defense you are supposed to pick up Kxx or KT onside, and always lose to Kx. That means if they cover you play to drop the ten, and if they don't cover you play to pin the ten. It is surprising how many people do not understand that combination.

[gnasher]

It is surprising how many people do not understand that combination.

Personally I'm surprised at how many people don't understand the meaning of "play of this suit for no loser".

[benlessard]

With QJ9x in dummy against best defense you are supposed to pick up Kxx or KT onside, and always lose to Kx. That means if they cover you play to drop the ten, and if they don't cover you play to pin the ten. It is surprising how many people do not understand that combination.

Dummy is Axxx and you dont know if declarer has QJ9x or QJxx.

[gnasher]

Dummy is Axxx and you dont know if declarer has QJ9x or QJxx.

The original problem said that dummy could be in either hand. You and Justin seem to have answered half of the question each.

[dburn]

With JTxx declarer can play the J to induce a cover from Kx or can play A hoping to pick a stiff honor in LHO or pick up KQ. Playing the A is better.

No, it isn't. With J1065 facing A432, you have two equally plausible lines for one loser: lead the jack, hoping that North has a doubleton honour; or lead low towards the J10, hoping that South has a doubleton honour. When you lead the jack you are indeed hoping that North has Kx (or Qx), but you're not "hoping to induce a cover" - you don't care whether he covers or not. Cashing the ace first is hopeless - that gains only when North has a singleton honour or either opponent has KQ doubleton, a worse chance than to find Hx in the "right" hand.

 

A case where one should perhaps not select randomly from equals is this one:

 

[hv=n=s987&w=sj1043&e=sa652&s=skq]399|300|[/hv]

Declarer leads low from dummy (East). South should always play the king, never the queen (because declarer will not believe that South would play the queen from Qx).

 

However, benlessard is right about one thing: these are complex positions. Roudinesco's Dictionary of Suit Combinations and Warmdeman's program Suitplay are, as far as I know, the best attempts so far to address the question of how to handle certain common (and uncommon) combinations. As far as I can see, though, they both operate on the premise that the defenders know what declarer's holding is, and will false-card and give other losing options where necessary. In real life, of course, this is almost never the case.

 

The actual position in Pau was:

 

[hv=n=s987&w=sj1043&e=sa652&s=skq]399|300|[/hv]

or the like, and what interested me was what North's strategy ought to be. At more than one table (with East as dummy) North did not cover the queen; declarer led the jack next (playing for North to have been forced to play low from Kxx, rather than to have chosen to play low from Kx) and went down - an interesting application of the Principle of Restricted Choice. At more than one table North did cover the queen; declarer led to the nine next (playing for North to have refused to risk letting a no-play slam through if declarer did not hold the nine) and made the contract.

 

Since it is exactly 50-50 whether declarer holds the nine or not, were the Norths who did not cover the queen defending rationally?

[Trumpace]

Since it is exactly 50-50 whether declarer holds the nine or not, were the Norths who did not cover the queen defending rationally?

If we ignore the play upto the point when declarer leads the Q, then yes, it is 50-50 in the cases which matter (QJ with declarer and T with partner).

 

So basically, an answer to your question if North's who did not cover were defending rationally, is "not enough information" :P

 

Also, I guess the scoring was IMPS...

[ceeb]

Since it is exactly 50-50 whether declarer holds the nine or not, were the Norths who did not cover the queen defending rationally?

I take your point. No.

 

Under the given (reasonable) assumptions:

1. Declarer is equally likely to have QJ9x or QJxx

2. No other holding matters:

 

By covering you win whenever partner has the 9 and possibly additional cases where declarer misguesses with QJ9x, so at least half the time.

 

By ducking you lose whenever partner has the 9 and possibly additional cases where declarer guesses correctly with QJ9x, so at least half the time.

 

Given that choice, it's not rational to lose at least half the time rather than win at least half the time. But every dog has its day.

[benlessard]

you are right

 

Hxx -----Hx is more frequent than.

 

H--------Hxxx

Hxxx----H

xxx------HH.

 

My point was that playing the A is better then leading the J from JTxx so when he lead the J you have to assume hes got the Q and we he lead the Q you have to assume hes got the J. So wich honor he lead doesnt matter. If QJxx is as likely then QJ9x (sometimes when you see the rest of the hand you can assume that QJxx isnt strong enough for the bidding) then north that didnt covering the honnor are making a mistake.

[Finch]

My point was that playing the A is better then leading the J from JTxx

And the point he made is that with J10xx opposite Axxx leading the Jack (or low towards the Jack) is better than leading the ace.

 

But as has been explained already declarer is known to need to play the suit for no loser so this is all completely irrelevant to the problem at hand.

 

and we he lead the Q you have to assume hes got the J.

 

No, that's not always true either.

Suppose the A643 are in dummy (so you don't know if declarer has the QJ or not).

We're told that we know declarer is in a 4-4 fit.

 

Now, here we have the 8 which changes matters slightly, but if we had K5 then delcarer, holding Q1098 has a genuine choice of whether to play for singleton king in your hand or singleton jack in partner's.

 

In this particular case when we have the 8 he does have to have the jack, as the with Q1097 the only holding he can pick up is singleton king in your hand.

[Finch]

At more than one table North did cover the queen; declarer led to the nine next (playing for North to have refused to risk letting a no-play slam through if declarer did not hold the nine) and made the contract.

 

Are you sure that the fit was known to be 4-4 at these tables?

 

At the table I was watching, with East as dummy and this suit as trumps, declarer could easily have had 5 hearts - he responded 1H to dummy's opening one of a minor and was raised.

 

When declarer may have 4 or 5 hearts the position is totally different, as with Q109xx declarer will also lead the queen (it's as good as any other technical line plus the chance of the non-cover). In that position the defence should always cover with Kx because of this risk, and hence declarer should always finesse against the 10 on the way back.

 

Mind you, with QJ9xx declarer should lead the Jack if it's possible on the auction that he has J10xxxx because he doesn't want a cover...

[benlessard]

and we he lead the Q you have to assume hes got the J.

 

No, that's not always true either.

Suppose the A643 are in dummy (so you don't know if declarer has the QJ or not).

We're told that we know declarer is in a 4-4 fit.

 

Now, here we have the 8 which changes matters slightly, but if we had K5 then delcarer, holding Q1098 has a genuine choice of whether to play for singleton king in your hand or singleton jack in partner's.

 

In this particular case when we have the 8 he does have to have the jack, as the with Q1097 the only holding he can pick up is singleton king in your hand

 

A6xx vs QT98 hes aways got at least 1 loser no matter if we cover or not. So in this particular example you have to assume hes got at least QJ. But the point is good if we dont know how many loser the declarer can afford.