Q or 9, is there a difference

[jdonn]

Aw don't stop, this has certainly been instructive for all of us!

[foo]

Aw don't stop, this has certainly been instructive for all of us!

OK, one last salvo then :D

 

p212 Roudinesco, combination 26b.

The proper play is exactly as I have stated and for exactly the reasons I have stated.

Restricted Choice doesn't enter into it.

 

I should probably write up a post on Restricted Choice just clean up cobwebs some might have on that topic...

[jdonn]

That's probably a good idea.

[Stephen Tu]

OK, finally you are starting to put up some actual stuff. Unfortunately, as I will show below, it still doesn't help you. Keep trying if you want, I'll keep knocking them down.

 

bottom p382 "a mistake to avoid" to top p383 Ex 4 and 5

(ex 4 is KJ9+xxx; ex 5 is AQTxxx+xxx)

 

Yes, I agree that in these particular combinations, restricted choice does not apply. That is because all the relevant honors are non-touching, there are no relevant touching/equal honors where a defender has a choice of which to play & can randomize.

 

However, these combinations do not bear close resemblance to the suit combo being discussed. In the previous discussion, you have a non-touching honor, the K, and two touching honors, the JT. Restricted choice absolutely does apply to the probability of someone holding both the J & T. You seem to think that just because one of the honors is non-touching, restricted choice goes completely out the window. It doesn't, if there are still relevant touching honors. If you are looking for an example from the encyclopedia, look at the combination immediately preceding this section, under "other card combinations"

KT9 opposite 432.

Note that the best play is to hook low twice. Here, you have the ace missing, which is non-touching, and the QJ, which are touching. If restricted choice did NOT apply, it would be a tossup whether to play K or T 2nd round. But it does apply, because if RHO has both QJ he should play one or the other half the time, while with just one he will play it all the time.

This combo is analogous to the one under discussion, with the ace taking the role of the K, the QJ taking the role of the JT.

 

Note also that this example clearly breaks your previously posted bogus rules about restricted choice:

=you are missing 3 important cards, not 2 @ T1 in this suit.

=the 2 honors that remain missing are not neighbors in the suit (AK, KQ, QJ, JT)

 

Yet somehow the encyclopedia still calls it a restricted choice situation ...

 

The odds do not change enough based on the play of T1 for it to affect your plans

Please answer this. Set aside the line of playing Ace 2nd round temporarily. Do you, or do you not believe, that playing low 2nd round is better than the Q? And if so, how do you reconcile with the 3.4% vs. 2.8% chances you respectively assigned JT tight, stiff T, as I pointed out above?

 

Had the cards been 54+KJT, or x+KJTx would RHO =ever= win T1 with anything but the T?

No. He has no reason to randomize his choice or to falsecard.

These holdings are irrelevant as you always lose 2-3 tricks. The only relevant one here is JT doubleton. With the JT doubleton, the defender DOES have reason to randomize.

 

If Restricted Choice held, you would be able to pick which card to play on T2 based on the play of T1 to maximize your chances. You can not. Therefore it is not an issue.

For the question of whether hooking low or the Q is better, you CAN pick. Deciding vs. playing the ace is somewhat different, just you have to assign probabilities to the falsecarding possibilities, and can't use 50% since there is no reason for the defender to use that unlike the equal honors case.

 

..and using impossiblities as "props" for your arguments is such a logical fallacy as to make logical discussion of any topic that you bring them to impossible.

I'm not using any impossibilities. Just there are multiple ways to frame the argument. I am fine with ignoring J stiff since the problem specified RHO played the T. Just answer my questions above, explain why you get 3.4% for the frequency of JT doubleton.

 

212 Roudinesco, combination 26b.

The proper play is exactly as I have stated and for exactly the reasons I have stated.

I agree completely with Roudinesco here, and it's the answer I & others like Justin have stated all along. Against LHO who can't falsecard, play ace, against ones that can, hook low. The point of the matter, is that if you agree with this as I do, that hooking low is better than hooking Q, you are using restricted choice whether you know it or want to admit to it or not. If restricted choice was bogus, then hooking the Q would be arguably better than hooking low, since probability of JT is > than probability of stiff T. But it's not bogus, and you have to multiply that probability of JT by half before comparing to be correct. That's applying restricted choice.

[Guest Jlall]

At some point you gotta ask yourself what is the point stephen :D

[kgr]

Suitplay says the best play is to finesse twice small.

=> What is the use for LHO to falsecard with Jxx on the second round? If he play the J then you best play is still to play the Q loosing to the K, but you would have lost to the K anyway if you finesse twice small?

[cherdano]

Suitplay says the best play is to finesse twice small.

=> What is the use for LHO to falsecard with Jxx on the second round? If he play the J then you best play is still to play the Q loosing to the K, but you would have lost to the K anyway if you finesse twice small?

But if you know that LHO never falsecards with Jxx, and RHO doesn't often falsecard with KT, then you would play the A, dropping the king.

[jdonn]

Suitplay says the best play is to finesse twice small.

=> What is the use for LHO to falsecard with Jxx on the second round? If he play the J then you best play is still to play the Q loosing to the K, but you would have lost to the K anyway if you finesse twice small?

But if you know that LHO never falsecards with Jxx, and RHO doesn't often falsecard with KT, then you would play the A, dropping the king.

I bet if I falsecarded king declarer would be the doofus in the room who holds Jxx :P

[kenberg]

I have thought a bit about this Jxx falsecard. I think the basic reason is this: Sometimes lho is dealt JTx or JTxx. If he plays low first round the 8 or 9 will force the king after which declarer will pick up his JT (by force in the 3-2 case and by running the 9 or 8 in the 4-1 case). Therefore lho wishes to split with this JTx(x) holding. But if he only splits when holding both J and T, it won't work since the remaining honor is picked up on a finesse. So he "splits" when he holds Jxx or Txx as well.

 

Thus, the J or T from Jxx or Txx allows a split from JTx(x) w/o giving the show away. Whether it will work is another question. That is, if the first round goes spot JQK, how should declarer continue given that lho may or may not have the ten to go with his J?

 

If lho can be trusted to hop up with the J with or without the ten, it does give something away. After spot-spot-9-T and then spot-spot on the second round, declarer can rule out lho holding Jxx. This means there is no point in playing the ace second round, and the Q everyone seems to agree is clearly wrong, so lho is helpless when he is dealt KJxx. Paradoxically, perhaps, declarer has a tougher choice playing against, say, me. Or at least against me before I read this thread. I can't be trusted to hop up with the J first round, and so on the second round declarer has a real choice between playing low and playing the ace.

 

Of course no real paradox exists. Good defense makes it harder for declarer when lho holds one of Jxx, Txx, JTx(x) and these are the more frequent situations. He has to pay off when he holds KJxx or KTxx but it's not a huge deal since declarer had a good chance of getting it right anyway.

 

I hope this is not garbage or brain dead or whatever. It's an interesting combo.

[Stephen Tu]

No kenberg, the 2nd round falsecard is not at all about protecting JTxx/JTx. Declarer is supposed to pick up those combinations anyway; he will finesse 2nd round after it goes TQK o JQK. If he doesn't, he is giving up too much since those combos outnumber Jxx/Txx.

 

The falsecard is solely about protecting partner's KT, stopping declarer from dropping the K under the ace. If you falsecard, declarer still must finesse, since otherwise he'll lose to KJx which outnumbers Jxx. So it's a mandatory falsecard if you know the combo. The falsecard means that the best declarer can do is pick up KJxx. By falsecarding, defender is giving up KJxx to guarantee success on the more numerous Jxx. Declarer has no counter because KJx is still possible.

 

Although you could falsecard first round if you knew declarer held this combo, it's dangerous. Sometimes declarer has a 9 cd trump suit. He was planning to take MP line for max tricks, hook first round Q, then try to drop your J, losing 2 tricks on this layout with partner's stiff K. Falsecarding 2nd round doesn't have this problem.

 

Paradoxically, perhaps, declarer has a tougher choice playing against, say, me

.

Yes, if the goal is max exploitation. Declarer could always just decline to play ace 2nd round and guarantee the normal rate of success against good players, giving up the chance to gain 1.1% vs. ones that don't know to falsecard.

[kenberg]

I meant first round.

 

But anyway, even if defender is sure that declarer has exactly three cards (often he will be) I guess, as you say, he cannot save his JTx. If defender is known to throw the J/T on the first round from Jxx or Txx, then declarer might have a tense moment on the second round worrying whether the J was from Jxx or JTx but he will presumably count out the possibilities and then, I think, finesse.

 

But if this is so (as I think it is) then the first round falsecard is another way to get two defensive tricks when holding Jxx or Txx opposite KT or KJ. Given confidence in declarer's length, it should work.

 

But fundamentally I now agree, you are right. It's a second round play. Actually I think I would be up for the second round falsecard. It's almost obvious, really. But maybe I'll try it first round (the play I though was being suggested) sometime just for the excitement. As a practical matter it could be worth a trick or two later as declarer fumes.

 

Thanks for the response.

[kgr]

The falsecard is solely about protecting partner's KT, stopping declarer from dropping the K under the ace. If you falsecard, declarer still must finesse, since otherwise he'll lose to KJx which outnumbers Jxx. So it's a mandatory falsecard if you know the combo. The falsecard means that the best declarer can do is pick up KJxx. By falsecarding, defender is giving up KJxx to guarantee success on the more numerous Jxx. Declarer has no counter because KJx is still possible.

You loose first finesse to the T:

- Playing the Ace is better when playing against opps that never falsecard with Jxx or Txx?

- Finessing with Q is better against opps that falsecard and play J or T second round (as he could have started from KJx or Jxx). I would think that KJx is as likely as Jxx?

- If a falsecarding opp plays x in the 2nd round, then you know he does not have Jxx. He can still have started from xx, Kxx, or KJxx. I would think that Kxx is more likely then KJxx and that finesse with the Q should be better then with the 9?

[cherdano]

The falsecard is solely about protecting partner's KT, stopping declarer from dropping the K under the ace.  If you falsecard, declarer still must finesse, since otherwise he'll lose to KJx which outnumbers Jxx.  So it's a mandatory falsecard if you know the combo.  The falsecard means that the best declarer can do is pick up KJxx.  By falsecarding, defender is giving up KJxx to guarantee success on the more numerous Jxx.  Declarer has no counter because KJx is still possible.

You loose first finesse to the T:

- Playing the Ace is better when playing against opps that never falsecard with Jxx or Txx?

Yes.

- Finessing with Q is better against opps that falsecard and play J or T second round (as he could have started from KJx or Jxx). I would think that KJx is as likely as Jxx?

Yes. No, KJx (two possible holdings) is more likely than Jxx (exactly one holding).

- If a falsecarding opp plays x in the 2nd round, then you know he does not have Jxx. He can still have started from xx, Kxx, or KJxx. I would think that Kxx is more likely then KJxx and that finesse with the Q should be better then with the 9?

Kxx is slightly more likely than KJxx a priori; but due to restricted choice, after the first round to the ten, KJxx is almost twice as likely. (I hope foo doesn't read this because I don't want to see more nonsense in this thread.)

[Stephen Tu]

You loose first finesse to the T:

 

FYI, there is only one 'o' in "lose", "loser", "losing", opposite of win/winner/winning. Loose is a different word, opposite of "tight".

 

-I would think that Kxx is more likely then KJxx and that finesse with the Q should be better then with the 9?

 

There are two valid ways you can think about this.

Remember that a priori, JT is 3.4%, stiff T is 2.8%

1:

The opponent shouldn't play T from JT all the time. So when you see the T, out of the combinations that matter here when comparing finesse lines, it is either he holds stiff T (2.8%), or he held JT AND he chose to play the T (3.4% * n=amount he doesn't falsecard). The opp is suppose to falsecard enough such that 3.4*n < 2.8, and also 3.4 * (1-n) < 2.8 (so that you don't have a tell when the J falls either). n = 0.5 = 50% is a reasonable estimate.

 

2:

Instead of just considering when the T falls on the right, also consider when the J falls on the right. If you treat both honors as the same, then finessing with the 9 if either falls picks up KJxx & KTxx (2.8 x 2 = 5.6%), while playing the Q picks up 3.4%, regardless of how often opponent is playing one card or the other from JT. 5.6 > 3.4 so ...

 

You can improve on this if the opponent from JT plays too often one card or the other AND you know what the tendency is, but if he keeps frequencies in the optimal range you can do no better than pick up the stiff honor combos.

[kgr]

Tx Cherdano and Stephen (and all others) for these interesting answers. This is the first time I try to analyse a suit combination like this. (Trying to lose :) less in the future.).

Maybe this will be the start to analyse and understand others in the future.

Let me add some questions/remarks (with the help of suitplay):

1. After you played small to 9 and T, you play another small and LHO plays small. Following possible initial combinations are left:

a- xx vs KJT: 3.4 (not important, always loses 2)

b- Jxx vs KT: 3.4

c- Kxx vs JT: 3.4

d- KJxx vs T: 2.8

=>

- Play 8: 2.8 (picks d)

- Play Q: 3.4 (picks c); but this is actually only 1.7 as RHO would play the J half of the time.

- Play A: 3.4 (picks b )

==> Play the Ace is best (1b- )

2. Opps will always play J from Jxx in the 2nd trick.

Now option 1b- is no longer possible and it is better to play the 8 (==> 1d- )

3. Opps will play 50% of the time J from Jxx in the 2nd trick.

option 1d- is still better as probability of 1b- is only 50% of 3.4. (==> 1d- ). Sorry to ask this B) : Do you consider this a restricted choice?

Questions:

A. The play against opps that don't falsecard feels very strange. First you take the deep finesse and then you play the Ace. I would then think it's better to play Ace the first trick and play small to the Q in the 2nd trick.

B. How is suitplay taking (or not) taking these falsecards and restricted choices into account? Seems like suitplay also prefers line 1d- (finesse twice deep). It also considers that with Jxx you can play as well twice small, as 1st small and then J?

C. I wonder how WC players do this at the table. They will only get this right if they studied the combo in advance?

 

Thanks again!

Koen

[Stephen Tu]

==> Play the Ace is best (1b- )

(assuming never a 2nd round falsecard)

2. Opps will always play J from Jxx in the 2nd trick.

Now option 1b- is no longer possible and it is better to play the 8 (==> 1d- )

Yes.

3. Opps will play 50% of the time J from Jxx in the 2nd trick.

option 1d- is still better as probability of 1b- is only 50% of 3.4. (==> 1d- ).

Yes.

Do you consider this a restricted choice?

Not really this part of the problem. It's still a kind of Bayesian inference, but restricted choice is specifically the case when an opponent has either choice of playing one of equals, or no choice since he only has one of those cards. Restricted choice here only really affects the probability you assign to JT doubleton, since the other times RHO holds both J & T, anything you do doesn't really matter.

 

A. The play against opps that don't falsecard feels very strange. First you take the deep finesse and then you play the Ace. I would then think it's better to play Ace the first trick and play small to the Q in the 2nd trick.

 

Ace first, you lose the ability to trap certain honor combinations onside, and lose extra trick to: KJTxx, KJTx, JTx. Play low hook first exposes these.

 

B. How is suitplay taking (or not) taking these falsecards and restricted choices into account? Seems like suitplay also prefers line 1d- (finesse twice deep). It also considers that with Jxx you can play as well twice small, as 1st small and then J?

 

Very good question. I do not know how suitplay does its magic; you will have to email the author of the program for that. But it is considering optimal play by the defense, and it is assuming defenders are falsecarding appropriately.

 

Note that there is no flag you can set in the program to say "my opps don't know how to falsecard", nor does that it account for the fact in real life defense is not double-dummy, so some falsecards it assumes are impossible/improbable because declarer could hold something else & the falsecard would be too dangerous. It is great program, cuts short the busywork of enumerating all the combinations & percentages, but still one needs to think about which layouts the line it recommended is catering to, what falsecards are assumed & necessary. Especially when it counters your intuition. So here you think "why not Q, take JT", then hopefully you know principle of restricted choice, and notice that stiff J + stiff T > JT. Then you say "why not A, take care of KT", then start thinking about different orders opps can play the cards. Then you can make adjustments for real-life conditions.

 

C. I wonder how WC players do this at the table. They will only get this right if they studied the combo in advance?

Yes, I think they have read/thought about combo previously. If you are fast enough to figure this all out at table I don't want to play against you :P.

[skjaeran]

C. I wonder how WC players do this at the table. They will only get this right if they studied the combo in advance?

Yes, I think they have read/thought about combo previously. If you are fast enough to figure this all out at table I don't want to play against you :P.

Not all of them. You can give Helgemo almost any suit combination and he'll come up with the correct solution in just a couple of seconds.