[Ant590]

Ok, so I've been going through some suit combinations and I have a follow-up question I'm hoping someone can help me with.

Say we have the standard combination

AQTxx

xxxx

for 5 tricks. Now I know the best line is small to Q then cash Ace, hoping for Kx onside or stiff jack offside. But say RHO has shown 12 outside cards, then it must be right to finesse the ten.

So my question is: Presumably then there is some dividing line here when it is better odds to play the double finesse? How about a 3-level or 2-level opening by righty? Is there a good general rule here or is it all gut instinct?

[Hanoi5]

Are there many sequences or plays where you'd know 12 cards from opponents before committing to play this suit?

[Ant590]

Hanoi5, on Jul 25 2010, 05:00 PM, said:

Are there many sequences or plays where you'd know 12 cards from opponents before committing to play this suit?

Sure. I only put this to demonstrate there was a line where it was demonstrably correct to play the ten, in order to phrase the question where the "probably right to" line may be.

[Cascade]

Hanoi5, on Jul 26 2010, 04:00 AM, said:

Are there many sequences or plays where you'd know 12 cards from opponents before committing to play this suit?

Its reasonably common.

One common situation in which the count is partially inferential is where RHO is on lead and has shown a five-five hand and then does not the remaining suit. If the lead is not from a strong sequence you might deduce that RHO has at most a singleton trump.

[Cascade]

You can only get five tricks when RHO has two or one card in this suit. Let the probabilities of these two events P2 and P1 respectively.

Small to the Queen

This wins 0.5 * P2 + 0.25 * P1 (half the time the king is onside with a 2=2 break and 1/4 of the time the jack is offside with a 3=1 break).

Small to the Ten

This wins (1/6) * P2 + 0.5 * P1 (1/6 of the time there is KJ only onside and you obviously modify your plan and 1/2 of the time KJx is onside - half of the singletons offside are small)

Break Even

(1/2) * P2 + (1/4) * P1 = (1/6) * P2 + (1/2) * P1

(1/2 - 1/6) * P2 = (1/2 - 1/4) * P1

1/3 * P2 = 1/4 * P1

4/3 * P2 = P1

The probability of a singleton has to be more than 4/3 times the probability of a doubleton with RHO.

[Cascade]

With 6 known cards

Always play to the queen

With 7 known cards

Play to the ten only if you have no information about LHO.

With one known card in LHO the queen is the better play.

With 8 known cards

Play to the ten with 2 or fewer known cards in LHO

Play to the queen with 3 or more known cards in LHO

With 9 known cards

Play to the ten with 4 or fewer known cards in LHO

Play to the queen with 5 or more known cards in LHO

With 10 known cards

Play to the ten with 6 or fewer known cards in LHO

Play to the queen with 7 or more known cards in LHO

With 11 known cards

Play to the ten with 8 or fewer known cards in LHO

Play to the queen with 9 or more known cards in LHO

[OleBerg]

Hanoi5, on Jul 25 2010, 06:00 PM, said:

Are there many sequences or plays where you'd know 12 cards from opponents before committing to play this suit?

I wanted to win the "Most silly comment of the year" award, but how am I going to top this???

[Ant590]

Thanks for the awesome answer Wayne!

[MFA]

If this is trumps at matchpoints and everything else is dull and standard and we have the 9 also (so we can handle 4-0), then the percentage play is small to the ten, since that will gain more often than small to the Q. It loses 2 tricks to the singleton J offside but that is still just one matchpoint. If we instead want to get the maximum number of tricks on average, then small to the Q is right. Cute combination.

[655321]

With no information about the opponent's shapes, the reason you don't take a second finesse is just the simple vacant places calculation - i.e. if you play low to Q and K, then low toward the AT, RHO has 12 spaces, LHO has 11, so play for the drop.

Cascade, on Jul 25 2010, 12:02 PM, said:

With 10 known cards

Play to the queen with 7 or more known cards in LHO

Not sure about any of these numbers, so I randomly picked this one.

We are saying that we know 10 of RHO's cards (e.g. he has guaranteed exactly 5-5 in 2 suits) and we know 7 of LHO's cards? Then after low to the Queen and King, we would be playing low to the ten on the next round (LHO having 4 spaces compared to RHO's 2 spaces), so it would make sense to me to play low to ten, then low to queen (instead of low to queen).

I haven't done the maths here so perhaps I am wrong, but on an initial glance this recommendation of low to queen doesn't look right to me.

[Fluffy]

655321, on Jul 25 2010, 11:02 PM, said:

With no information about the opponent's shapes, the reason you don't take a second finesse is just the simple vacant places calculation - i.e. if you play low to Q and K, then low toward the AT, RHO has 12 spaces, LHO has 11, so play for the drop.

Cascade, on Jul 25 2010, 12:02 PM, said:

With 10 known cards

Play to the queen with 7 or more known cards in LHO

Not sure about any of these numbers, so I randomly picked this one.

We are saying that we know 10 of RHO's cards (e.g. he has guaranteed exactly 5-5 in 2 suits) and we know 7 of LHO's cards? Then after low to the Queen and King, we would be playing low to the ten on the next round (LHO having 4 spaces compared to RHO's 2 spaces), so it would make sense to me to play low to ten, then low to queen (instead of low to queen).

I haven't done the maths here so perhaps I am wrong, but on an initial glance this recommendation of low to queen doesn't look right to me.

this thread is all about making all the tricks. Not about what to do at second round when we are already 1 down .

[655321]

Fluffy, on Jul 25 2010, 06:41 PM, said:

this thread is all about making all the tricks. Not about what to do at second round when we are already 1 down .

Ah yes, I missed that.

[dake50]

Are we edging on "Roudi's rules"?
How about 6-suit RHO, how about NT overcall LHO?
Either of these sway "best" play --not just on this combo?

This combo must need very strong evidence to sway off to Q, then to A.

[Cascade]

MFA, on Jul 26 2010, 09:12 AM, said:

If this is trumps at matchpoints and everything else is dull and standard and we have the 9 also (so we can handle 4-0), then the percentage play is small to the ten, since that will gain more often than small to the Q. It loses 2 tricks to the singleton J offside but that is still just one matchpoint. If we instead want to get the maximum number of tricks on average, then small to the Q is right. Cute combination.

How does this work?

Can't you finesse for the king on the second round?

Also the problem with small to the ten is it loses one trick unnecessarily to the common holding of Jx offside.

[MFA]

Cascade, on Jul 26 2010, 02:52 AM, said:

MFA, on Jul 26 2010, 09:12 AM, said:

If this is trumps at matchpoints and everything else is dull and standard and we have the 9 also (so we can handle 4-0), then the percentage play is small to the ten, since that will gain more often than small to the Q. It loses 2 tricks to the singleton J offside but that is still just one matchpoint. If we instead want to get the maximum number of tricks on average, then small to the Q is right. Cute combination.

How does this work?

Can't you finesse for the king on the second round?

Also the problem with small to the ten is it loses one trick unnecessarily to the common holding of Jx offside.

Low to the T handles 4 cases (KJx, KJx, KJxx, Jxx onside) while low to the Q handles only 3 cases (Kx, Kx, Kxx onside). And since my assumption is matchpoints with everything else dull, we only care if it gains or loses, not if we win or lose 1 or 2 tricks.

If we take a finesse that loses, we should play for the drop on the second round. That is best in isolation and particularly crucial if we went deep on the first round, since everything is already lost anyway if it was the stiff J that took the trick.

[Fluffy]

low to the queen also handles KJxx onside michael. Besides, if this is amtchpoints but we are in slam it won't matter hwo many downs that much

[MFA]

Fluffy, on Jul 26 2010, 10:27 AM, said:

low to the queen also handles KJxx onside michael. Besides, if this is amtchpoints but we are in slam it won't matter hwo many downs that much

Not for 0 losers.

[Fluffy]

MFA, on Jul 26 2010, 09:02 AM, said:

Fluffy, on Jul 26 2010, 10:27 AM, said:

low to the queen also handles KJxx onside michael. Besides, if this is amtchpoints but we are in slam it won't matter hwo many downs that much

Not for 0 losers.

well sorry I mean, KJ98 onside actually since the suit is

AQ10xx

xxxx

it will make the same number of tricks.

[MFA]

Fluffy, on Jul 26 2010, 11:19 AM, said:

MFA, on Jul 26 2010, 09:02 AM, said:

Fluffy, on Jul 26 2010, 10:27 AM, said:

low to the queen also handles KJxx onside michael. Besides, if this is amtchpoints but we are in slam it won't matter hwo many downs that much

Not for 0 losers.

well sorry I mean, KJ98 onside actually since the suit is

AQ10xx

xxxx

it will make the same number of tricks.

Yes. But I carefully assumed the 9 also in my post to make my point, since I thought it would be an interesting addition to OP's question.