[JLOGIC]

Say we are in a suit contract at imps, RHO gets in and will obviously play dummys 3 small suit as the other 2 suits have no future, and they need to take 3 tricks immediately since we have a side suit set up or w/e. If they shift to low and we have KTx, do any of you ever play the king in that spot? The point obviously being most people shift to the jack or the queen if they have it without the ace as a technical play to guard against us having Kxx.

This happens a lot, and I notice I always shift to an honor as defender, if others do as well then the king might be percentage. We gain against AQxxx and AJxxx and Axxxx I guess, losing to Qxxxx and AQJ on our left. That's probably pretty close, but we also frequently gain 2 overtrick imps which seems like a lot. Also, even though they should not card honestly many people will shift to high from nothing and low from an honor without thinking about it too much, so we can re-evaluate if they shift to a high one and they are run of the mill good players.

How about if we don't have the ten? Kxx opp xxx in dummy or something. People seem to pop king here on a low shift, since assuming they're not dumb low only caters to AQJ/AQJT/AJ on our left. And does it even cater to those? If RHO shifted to say the ten, what would we do? If the ten is honest, I guess we would duck hoping for AJ or AQJ rather than QJ doubleton on our left, but that just means the ten is the correct play from AJTxx, AQTxx etc. Likewise, if we are ducking the 9 the 9 is the right play with Asomething9 fifth. If we are going to pop king on the 9 or ten shift, then low is never a good play unless they have the ace. If they are playing somewhat randomly and we are always popping king from Kxx then that means...

With Jxxx or Qxxx they COULD comfortably shift to low. It does not cost even when partner has AJT or AQT since declarer will go up king. And you get the added benefit that with KTx they may go up king lol. But my view of the real world is that experts shift to the jack or queen with J or Q empty there, and they often play the ten with KTx on the low shift, so my main point is that I think T from KTx is probably wrong. I don't think shifting to low from Q or J empty has any real world implications since it's more likely they go low from Kxx than K from KTx.

Thoughts? Theoretical thoughts welcome ofc but personally I am more interested in exploiting real life play since I wanna win more tricks and stuff

[JLOGIC]

Here is an example hand that I am making up if my post was too rambly to understand:

xx
KQJT
9xx
AKxx

AKQJxxx
xx
KTx
x

3N p 4H p 4S AP. 3N opener shows a solid major and a side ace or king. 4H was p/c (north wanted his partner to play 4S to protect the hypothetical DK).

Heart lead to east. Low diamond shift. Imps. My theory is that the king is the correct play.

This post has been edited by JLOGIC: 2013-November-02, 17:18

[ggwhiz]

I just don't see it at imps. Even if rho has the Ace it could be 3 or 4 times (with 2 they should lead it). We only go down to double Q or J with lho or like ALWAYS when they have the Ace and the shift was coming from rho regardless of what they hold. An example where the contract is not otherwise cold and/or rho doesn't know it would be very legit but I don't have any.

At Mp's I fly with the King but only against opponents named Marty.

[JLOGIC]

Qx and Jx are 9 combos (including QJ doubleton). AJ doubleton and AQJ and AQJx are 6 combos. So if those are the relevant holdings we go down more often by playing low than playing the king.

So I guess you are saying west could have AJx(x) or AQx(x)?

It is interesting you bring up matchpoints though, I would argue that it's more likely low is right at MP since with AQ they will probably cash the ace (if you have Kxx you will play the king, and if you have KJx you cannot misguess, if you have Kx you will obv play the king, so low from AQ is playing you for specifically KTxx). I guess it matters whether they know you have the king, if their partner might have it they would still probably play the ace though even though it gives partner a problem of whether to unblock for a ruff or not. Similar arguments apply to cashing the ace with AJ at MP.

Going with that, they might well shift to low from the queen in MP to give you a guess rather than cater to an unlikely AJT tripleton with partner where you still might go up king anyways. Matchpoints is really a mind ***** in this regard lol.

[cherdano]

I agree people always shift to an honor from Hxx. But what about Hxxx?

• Technically: now honor is catering for a single layout, AHT with partner.
  
• Psychologically: shifting to the Q from Qxxx just doesn't feel that automatic to me. You are also giving away half an overtrick any time declarer has KJx.

[JLOGIC]

 cherdano, on 2013-November-03, 04:39, said:

I agree people always shift to an honor from Hxx. But what about Hxxx?

• Technically: now honor is catering for a single layout, AHT with partner.
  
• Psychologically: shifting to the Q from Qxxx just doesn't feel that automatic to me. You are also giving away half an overtrick any time declarer has KJx.

Thats a great point, you actually give up more than half an overtrick since they should play the king from KJ(x) every time even if they know you shift to low from Qxxx and high from Qxx since you are more likely to have the ace than the queen. And in situations where they have two pitches if they pop king, that will gain them 2 imps vs 1 if they play the jack and that's right.

I think shifting to low from Qxxx is unexploitable, they should still pop king from Kxx even if they know you do so since low now caters to AJT, AJ, AQJ, and AQJx (7 combos) while losing to AQxxx/AJxxx/Axxxx (9 combos). So if they adjust by playing low from Kxx they're making an error, and they will still play the K from KJ(x) so you just pick up an imp there.

Also, they still cannot start ducking the Q shift with KTx picking up Ax with partner since Qxx is still more likely, and you might choose to randomize with QJxxx and shift to the jack sometimes.

So shifting to low from Qxxx is definitely right in theory. Wish me luck next time I have this and I shift to low and declarer plays low and I am explaining to my client that I'm not a noob and why this is a good play as I'm getting fired Glad I've learned at least something from this thread already.

My gut reaction would be that most experts do play the Q from Qxxx though because GOTTA PLAY FOR THE SET and that's what we've been taught to do, I might be wrong though. Anyone else have any empirical evidence on what they or other experts do or would do in this situation from Qxxx?

I think the jack from Jxxx is still right though, I don't see when it costs and obviously gains if declarer is a ducker from Kxx.

[ggwhiz]

 JLOGIC, on 2013-November-02, 19:43, said:

So I guess you are saying west could have AJx(x) or AQx(x)?

You should go down if you fly King and it's AJ tight? Interesting topic.

[jallerton]

 JLOGIC, on 2013-November-02, 17:11, said:

Here is an example hand that I am making up if my post was too rambly to understand:

xx
KQJT
9xx
AKxx

AKQJxxx
xx
KTx
x

3N p 4H p 4S AP. 3N opener shows a solid major and a side ace or king. 4H was p/c (north wanted his partner to play 4S to protect the hypothetical DK).

Heart lead to east. Low diamond shift. Imps. My theory is that the king is the correct play.

Maybe, but not if the opponents know what you are thinking. Playing against your theory, I'll switch to low card from Qxx or Jxx (beating the contract even when it was cold double dummy). I'll even switch to a low card from QJx, beating the cold contract when you hold K10x; even when my partner holds the 10 and I am letting the contract through by switching to a low one, declarer will come to the rescue by JLOGICally playing the king.

If I really want to extract the maximum from knowing your strategy, I won't even let you have the pleasure of 2 overtricks when the ace is onside.

Of course, you're a quick learner and then you'll start playing low when I switch to a low card. When I learn this, I'll revert to switching to my quacks. Then...

This reminds me of a hand posted from a recent BBFvsJEC match. It was given as a play problem, but declarer's optimal card at trick 2 depends on knowing the defender's strategy. Meanwhile, the defender's best switch at trick 2 from a certain holding depends on his assessment of what the declarer's strategy will be.

[JLOGIC]

 jallerton, on 2013-November-03, 15:01, said:

Maybe, but not if the opponents know what you are thinking. Playing against your theory, I'll switch to low card from Qxx or Jxx (beating the contract even when it was cold double dummy). I'll even switch to a low card from QJx, beating the cold contract when you hold K10x; even when my partner holds the 10 and I am letting the contract through by switching to a low one, declarer will come to the rescue by JLOGICally playing the king.

If I really want to extract the maximum from knowing your strategy, I won't even let you have the pleasure of 2 overtricks when the ace is onside.

Of course. It is an exploitative play, but do you think it would be good in real life normal situations (for instance, I was playing against someone like you who had never read my BBF posts)? I think knowing theory is good but knowing when to deviate to exploit generally normal strategies is very important in real life, and bridge is not like poker where you have to worry about playing balanced/close to optimally since generally you only see these types of situations at most once vs most people. If I did ever play the king against you and later was playing against you and had the exact same spot I would assume you remembered and be much more weary of doing so but that is very unlikely to occur.

My general opinion on situations like this is that you should actively try to exploit peoples tendencies to make the technical play, not randomize their spot cards, play in predictable patterns like pitching from a 5 card suit, etc. Those people can also exploit my tendencies to try and exploit this so I'm opening myself up to that but imo this doesn't happen often, even world class bridge players don't think about doing stuff like that, etc etc. Even situations like counterswinging are rarely discussed or thought about (I think).

This post has been edited by JLOGIC: 2013-November-03, 16:11

[nige1]

 JLOGIC, on 2013-November-02, 17:11, said:

♠ xx ♥ KQJT ♦ 9xx ♣ AKxx
♠ AKQJxxx ♥ xx ♦ KTx ♣ x
Heart lead to east. Low diamond shift. Imps. My theory is that the king is the correct play.

 jallerton, on 2013-November-03, 15:01, said:

Playing against your theory,
I'll switch to low card from Qxx or Jxx (beating the contract even when it was cold double dummy).
I'll even switch to a low card from QJx, beating the cold contract when you hold K10x; even when my partner holds the 10 and I am letting the contract through by switching to a low one, declarer will come to the rescue by JLOGICally playing the king.
If I really want to extract the maximum from knowing your strategy, I won't even let you have the pleasure of 2 overtricks when the ace is onside.
Of course, you're a quick learner and then you'll start playing low when I switch to a low card. When I learn this, I'll revert to switching to my quacks. Then...

Interesting problem, JLOGIC! IMO Jallerton is right that this is another context, where both declarer and defender should adopt a mixed strategy, assuming both are games-theory gurus.

[gszes]

This is closer to a coin flip than one might imagine. I concur that virtually everyone will switch
to H from any H(H)(xx) in order to cater to the possibility declarer has Kxx and they can set
the contract this is especially the case at IMPS where setting the contract is much way hugely
more important than engaging in mind games for an overtrick. We are reduced to the cases
where it is right for rho to play low. We are assuming the low cards are low enough that lho
will win any low spot play and continue the suit as needed.

It turns out that playing the Ten loses to LHO Qx Jx QJ which = 9 cases while playing the K wins those
------------------------------ the KING loses to LHO AQJ(x) 4 cases AQJ AJ Ax (4 cases) playing the TEN wins

to make matter worse if rho adopts a strategy of playing low from Qxxx or Jxxx playing the K
will allow them to set you while playing the T will win. Assuming the opps will play an honor to try
and set you (a high degree of probability) going up with the K is a 10 to 9 losing proposition.

This hardly makes going up with the K a poor decision but Ill keep on taking my 11% house edge
especially since my odds will only improve if rho starts to play low from Hx(xx) so I will play the TEN

If I have missed something obvious don't hammer on me too hard its been a longggggggggggg day:)))