[shyams]

Here's a Matchpoints hand for your consideration:

You are South (Vul vs. NV) and are playing with a robot (i.e. GIB 2/1 system).

Part 1:

---

Do you plan to pass 4♠ or explore slam? If you intend to bid on, what is your action plan?


Part 2:

---

Let's say you pass partner's 4♠ bid. Now, West (robot) leads ♥4, dummy plays the singleton ♥2, East (robot) play ♥J. How do you proceed with the play?

Useful notes:
a. This is ROBOT MPs. And everyone who bid as above will get the same lead as you (i.e. you aren't the only lucky one to not get a club lead).
b. Opponents trumps split 2-1. How many sure tricks do you have in 4♠?

In a nutshell, your objective is to make one or two more tricks than what you already have. What is your best approach to making the extra overtrick?
Finally, did you do something good when drawing trumps? Describe what you did in the trump play.

[Kaitlyn S]

shyams, on 2016-September-05, 03:12, said:

Here's a Matchpoints hand for your consideration:

You are South (Vul vs. NV) and are playing with a robot (i.e. GIB 2/1 system).

Part 1:

---

Do you plan to pass 4♠ or explore slam? If you intend to bid on, what is your action plan?


Part 2:

---

Let's say you pass partner's 4♠ bid. Now, West (robot) leads ♥4, dummy plays the singleton ♥2, East (robot) play ♥J. How do you proceed with the play?

Useful notes:
a. This is ROBOT MPs. And everyone who bid as above will get the same lead as you (i.e. you aren't the only lucky one to not get a club lead).
b. Opponents trumps split 2-1. How many sure tricks do you have in 4♠?

In a nutshell, your objective is to make one or two more tricks than what you already have. What is your best approach to making the extra overtrick?
Finally, did you do something good when drawing trumps? Describe what you did in the trump play.

With 2-1 trumps I should be able to trump all my hearts and lose only a club and a diamond. Leading a diamond hoping to get rid of a club loser shouldn't work. I think your only legitimate chance to make 6 is to hope that RHO started with H-QJ or H-QJx so that you can discard all three of dummy's clubs on your hearts and trump your club loser.

[Kaitlyn S]

shyams, on 2016-September-05, 03:12, said:

Here's a Matchpoints hand for your consideration:

You are South (Vul vs. NV) and are playing with a robot (i.e. GIB 2/1 system).

Part 1:

---

Do you plan to pass 4♠ or explore slam? If you intend to bid on, what is your action plan?


Part 2:

---

Let's say you pass partner's 4♠ bid. Now, West (robot) leads ♥4, dummy plays the singleton ♥2, East (robot) play ♥J. How do you proceed with the play?

Useful notes:
a. This is ROBOT MPs. And everyone who bid as above will get the same lead as you (i.e. you aren't the only lucky one to not get a club lead).
b. Opponents trumps split 2-1. How many sure tricks do you have in 4♠?

In a nutshell, your objective is to make one or two more tricks than what you already have. What is your best approach to making the extra overtrick?
Finally, did you do something good when drawing trumps? Describe what you did in the trump play.

With 2-1 trumps I should be able to trump all my hearts and lose only a club and a diamond. Leading a diamond hoping to get rid of a club loser shouldn't work. I think your only legitimate chance to make 6 is to hope that RHO started with H-QJ or H-QJx so that you can discard all three of dummy's clubs on your hearts and trump your club loser.

[shyams]

I was hoping for a few more responses. In any case, here are some interesting points relating to this problem:

Part 1:
North's jump shows 5+ card ♠ suit and fewer than 8 HCP. You really need a "perfect" arrangement for a slam (e.g. trumps to split kindly + North to hold ♣A & ♦K, and West to hold ♦A so that when you lead up to dummy's King it scores a trick). In short, at MPs, it makes little sense to explore slam.

What is the risk in trying? Well, this deal was from the daily free MP tournament. Everyone who explored slam received a club lead and had to settle for 11 tricks {either 5♠= or 6♠ -1}.

Part 2:

As Kaitlyn S pointed out, you need the ♥Q to drop on the 3rd round of hearts (it need not be RHO who began with QJx, it can also be LHO deciding to lead a low ♥ from Qxx). This allows us to draw trumps in 2 rounds, discard all the clubs from dummy on declarer's hearts and then ruff the ♣5 in dummy. You only lose one diamond.

However, there was a small but significant precaution that must be taken. You cannot afford to throw both ♠4 and ♠6 on the top two trumps.
For example,
#1 ---- ♥4 - ♥2 - ♥J - ♥A
#2 ---- ♠A - ♠5 - ♠4 - ♠3
#3 ---- ♠K - ♠Q - ♠6 - ♣3
#4 ---- ♥K - ♥6 - ♣2 - ♥7
#5 ---- ♥5 - ♥3 - ♠9 - ♥Q
Oops! You have two winning hearts in the South hand but no way to reach it without allowing opponents to cash a club and a diamond. Instead, you must play a high spade on trick 3, ruff the low heart with a high spade on trick 5 and then overtake ♠6 in hand and discard two clubs on the winning hearts.

The reward for playing carefully? You score 95% of the Matchpoints! Over half of the players in 4♠ led a diamond at trick 4. Of the handful who went about trying to establish hearts, some found too late that they had no way to return to the South hand without conceding two tricks.

[Cyberyeti]

shyams, on 2016-September-08, 10:46, said:

I was hoping for a few more responses. In any case, here are some interesting points relating to this problem:

Part 1:
North's jump shows 5+ card ♠ suit and fewer than 8 HCP. You really need a "perfect" arrangement for a slam (e.g. trumps to split kindly + North to hold ♣A & ♦K, and West to hold ♦A so that when you lead up to dummy's King it scores a trick). In short, at MPs, it makes little sense to explore slam.

What is the risk in trying? Well, this deal was from the daily free MP tournament. Everyone who explored slam received a club lead and had to settle for 11 tricks {either 5♠= or 6♠ -1}.

Nope, xxxxx, Qxx, Ax, xxx is sufficient for slam even if the hearts misbehave as long as trumps are 2-1

[shyams]

Cyberyeti, on 2016-September-08, 14:25, said:

Nope, xxxxx, Qxx, Ax, xxx is sufficient for slam even if the hearts misbehave as long as trumps are 2-1

Why is there a fascination to literally disprove every statement in a generic argument with specific examples? How does it add to the generic argument OR the gist of what was being conveyed?

[Cyberyeti]

shyams, on 2016-September-08, 14:48, said:

Why is there a fascination to literally disprove every statement in a generic argument with specific examples? How does it add to the generic argument OR the gist of what was being conveyed?

Purely because the example/analysis you gave was wrong, the key is the heart/diamond holdings. The holding you give still goes down with ♥ xxx, most holdings with Q♥/A♦ will make, a decent number of holdings with ♥xx/♦Ax are good particularly if holding some top spades. The actual main danger of investigation is that partner has a quacky hand that gives you 3 losers in the minors, so that 5♠ is not safe.

There is also not a good way to explore for most people, 4N doesn't tell you which ace, if you cue 1st/2nd and cue 5♣ you don't know if partner's 5♦ is 1st or second round.

[quiddity]

The play is not so simple; there is an alternate line which succeeds when LHO led from Qxxx. It's not obvious to me which is better

Sorry, it is obvious which is better, but IMO still worth discussing both lines and how to choose between them.

[quiddity]

No takers? Here's a stab at some analysis, though I'm sure I'm missing something. Please point out any mistakes.

The alternate line is a ruffing finesse against West. Draw trumps, unblocking as described, cash the second high heart pitching a club, then run the ♥T pitching another club. If LHO started with Qxxx you can now ruff out the last heart, return to hand with a trump, and pitch the last club on the fifth heart. So which line is better?

The ruffing finesse gains when East started with Jxx. Initially there were 5 small heart spots outstanding, and there are 10 ways to choose 2 from those 5.
It loses when East started with QJx or QJxx. There are 5 combinations of QJx and 10 combinations of QJxx. So absent any other information the ruffing finesse is a 2:3 underdog.

However, there is the question of restricted choice. When East holds Jxx he must play the J at trick one, but when he holds QJx or QJxx he could theoretically play the queen instead. Perhaps the bots are programmed to always play the lower of touching honors here, in which case it doesn't apply, but if it plays an honor randomly then we should discount half of the QJx(x) holdings and the ruffing finesse now becomes a favorite.

Finally there is the question of opening leads. If E-W are leading low from three or four we don't have any extra information, BUT if they are leading "3rd from even and low from 3 small" then we can do better. We lead the ♥T on the third round and check West's spot. If West's opening lead was the lowest spot then he can't possibly have Qxxx, so we would ruff. In that case we would never lose to East's QJxx, only to QJx, so the comparison would be 10 combinations of Jxx vs 5 combinations of QJx and the ruffing finesse would be a 2:1 favorite (not counting restricted choice). Similar reasoning applies if E-W are leading MUD from three small and low from Qxxx.

[kenberg]

There is, I think another important point regarding the ruffing finesse versus trying to drop the honors.

It's true that a successful ruffing finesse against Qxxx will bring in 12 tricks. It's also true that if rho started with both heart honors then the ruffing finesse fails and, if you try it, the opponents can and no doubt will hold you to ten tricks. However, if you simply win at T1 with the ace, draw trump in two round being careful to leave an entry back to hand, and then play the heart K tossing a club and the heart Jack, ruffing whether or not it is covered, you surely take 11 tricks and you take 12 if the Q comes down on the third round of hearts.

I go conservative and try for the drop, guaranteeing 11 tricks..

Anyway proper care of transportation is an important point, often missed. We shouldn't miss it, but we do.

Note: I edited this for brevity and, I hope, for clarity.

[quiddity]

Field considerations. Let's assume conditions which make the finesse a favorite: no restricted choice but EW are leading 3rd/5th so it doesn't lose to East's QJxx. Then there are two heart distributions which matter:
A: Qxxx Jxx (10 comb)
B: xxxx QJx (5 comb)

Let's assume favorable field conditions for the ruff: one table ruffs the third round, one table takes the finesse, and all other tables play on diamonds to score 11 tricks. This means that the ruff gets a near top whenever the queen falls and the finesse gets a zero when it loses to QJx.

In case A, two-thirds of the cases that matter, the ruff will get a middle board and the finesse will get a top. In case B, one-third of the cases which matter, the ruffer will get a top and the finesse will get a zero. So the expectations are:
ruff: .66(.5)+.33(1) = .66
finesse: .66(1)+.33(0) = .66

Equally good, but that was for the optimal field condition. As more tables try to ruff their score for case B goes down without any compensating effect, so the overall expectation for the ruff is lower.

What happens if the field moves the other way and more tables try to finesse? Then the score for the finesse in case A goes down (no longer a clear top) but so does the score for the ruff (losing to more finessing tables). Also, the score for the finesse in case B goes up (no longer a clear bottom).

Thus it is right to take the finesse when it is odds-on and not worry about the occasional bad board.

[kenberg]

First an apology. This is about to get a little heavy for the NB forum. But moving on:

---

The heart 4 is led to the 2-J-A, the trump AK are cashed. Cash the heart K, lead the J. No Q has yet appeared.

If we consider tossing a club, we should, we agree. consult their cc. The OP said bots, but let us assume humans so we have to look.

3/5 leads were suggested. Let's assume that allies to spots as well so that the 4 might have been from 8,6,4. Let's analyze on the assumption that the leads is true to heir agreements.

You, the declarer, have now seen three spots from W. Suppose W has not shown, say, 8,6,4. He does not have the Q, since from Q,8,6,4 he would have led the 6. Same thing with any two spots higher than the 4.

So we have a problem only when W's played spots are 4,3, and then, on the J, a spot higher than the 4.

EW started with Q,J 8,7,6,4,3

After W leads the 4, plays the 3, and follows with one of 8,7,6 we have to decide whether W started with Qx43 or xy43, where x and y must come from 8,7,6. It seems that there are three ways for either.


Ah yes, but now maybe we go back to Restricted Choice at T1. It's not too much to ask of E to see that deception could be useful.

Now some people play that the lead of a small card promises an honor. If that is their agreement, I can work out what to do! Checking the leads on their card can be useful, granted.

[shyams]

quiddity, on 2016-September-14, 11:48, said:

No takers? Here's a stab at some analysis, though I'm sure I'm missing something. Please point out any mistakes.

In the opening post, I said it is Matchpoints.

So let's revisit your alternatives from the MP perspective:
1. You play West to hold Qxxx and therefore run the Heart 10.
....1a. You got it wrong. East wins, returns a club and you are no longer in any position to score more than 10 tricks. Your MP score is (almost) zero MPs.
....1b. You got it right. You score a top on the board.

2. I play to drop Queen on the third round.
....2a. I get it wrong. I switch back to diamonds. And score an average on the board.
....2b. I get it right and score a top.

Which of these two is better?

[quiddity]

kenberg, on 2016-September-15, 18:03, said:

After W leads the 4, plays the 3, and follows with one of 8,7,6 we have to decide whether W started with Qx43 or xy43, where x and y must come from 8,7,6. It seems that there are three ways for either.

Consider the cases:
Qxxx Jxx
xxxx QJx

In both, after you play the ♥T on the third round you will have seen 3 spots from west and one honor plus one spot from east. You could always reason as you did that there are three ways for either. But if you count up the original possibilities you'll see that the first case is twice as likely as the second so the reasoning must be somehow faulty.

[quiddity]

shyams, on 2016-September-16, 02:29, said:

In the opening post, I said it is Matchpoints.

So let's revisit your alternatives from the MP perspective:
1. You play West to hold Qxxx and therefore run the Heart 10.
....1a. You got it wrong. East wins, returns a club and you are no longer in any position to score more than 10 tricks. Your MP score is (almost) zero MPs.
....1b. You got it right. You score a top on the board.

2. I play to drop Queen on the third round.
....2a. I get it wrong. I switch back to diamonds. And score an average on the board.
....2b. I get it right and score a top.

Which of these two is better?

If the conditions are such that the finesse is a 2:1 favorite vs the ruff (either because of leads or restricted choice) then the two are equivalent. I will get a top 2/3 of the time. You will get a top 1/3 of the time and an average 2/3 of the time. Over the long run these have the same expectation.

But in a reasonable game you won't be the only one trying to ruff out the hearts so you won't get a true top when you're right. And perhaps I won't be the only one finessing so I won't get a true zero when I'm wrong. These seem symmetric but as I mentioned in a previous post they are not.

The ruffing expectation depends upon getting an average board when it's wrong. That average decays as more of the field chooses to finesse.
The finesse always loses to everyone else when it's wrong, so its expectation does not decay as more of the field chooses to ruff.

[kenberg]

quiddity, on 2016-September-16, 06:48, said:

Consider the cases:
Qxxx Jxx
xxxx QJx

In both, after you play the ♥T on the third round you will have seen 3 spots from west and one honor plus one spot from east. You could always reason as you did that there are three ways for either. But if you count up the original possibilities you'll see that the first case is twice as likely as the second so the reasoning must be somehow faulty.

I don't think so. The issue is that some of the spots are indistinguishable, some are not. That is, holding the 8643 and leading the 4 to the A, then on the K and T W must play the 3 since playing the 8 and 6 would show declarer that he, W, does not hold the Q. This is, of course, assuming that declarer after consulting the cc believes that the original lead was third best. So he must play the 3 on either the K or the T, but then he is free to choose between the 8 and the 6 for his other discard.
The point is that simply counting spots won't suffice. The spots all have equivalent value in trick taking (none) but they have different values in information that they disclose.

At first there appears to be something of a restricted choice argument here If W started with Qx43 then he must play the x, while if he started with xy43 then he had a choice of which card to play. The problem is that this is balanced by a restricted choice argument on spots for E. He played the J at T1 and then some z when the heart K was led. If he started with Jwz he had a choice, if he started with QJz he did not. So An RC argument on spots can be applied to either E or W and by counting, they balance out.

My way of counting treats the 8,7,6 as indistinguishable but not the 3. In order for a problem to develop, assuming that he led third best, W must have 3 and he must play it. So the relevant spots are the indistinguishable 8,7,6. If we go only by these, it's 50-50. I am willing to buy that there may be something in your RC argument at T1, E could play either the Q or the J if he had both. Standard play is the J to help pard work out the hand, but perhaps he would play a deceptive Q of he had both. Perhaps.

I am pretty sure all this is right, but I want to make a different, and simpler, point in a separate post.

[kenberg]

As mentioned, a simpler point.

Theorem: One can argue, as we have, about the merits of the ruffing finesse. Proof: We have been doing so. QED.

However.

OP wanted to show the necessity of some care in transportation. This is very important and very suitable for the B/N Forum. So let's not get lost in this other stuff.

Also. Even at matchpoints I am reluctant to gamble when I am in a good contract and get a favorable opening lead. I know it was stipulated in the OP that these are bots and the lead will be the same at all tables. But I prefer that my friends, my partners and my opponents be human so even if the OP said bots I think it is worth thinking a little about play against humans. I look at this and say: I am taking 11 tricks. If I am lucky maybe 12. If I can take 12 others can to if they get the same lead, but they might not get the same lead and, really, they might not be careful in planning their transportation. So I am not going to do anything to botch this up.

Yes I sometimes make plays at matchpoints that I would not dream of making at imps. For the B/N readers out there, I (Advanced not Expert) suggest going easy on such schemes.

[quiddity]

Thanks Ken, I think I understand now. The opening lead ♥4 excludes more of the original Qxxx possibilities than xxxx possibilities. That's a great point.

[kenberg]

quiddity, on 2016-September-16, 13:59, said:

Thanks Ken, I think I understand now. The opening lead ♥4 excludes more of the original Qxxx possibilities than xxxx possibilities. That's a great point.

And leads to the practical problem that if the lead had been the 6 instead of the 4 we would have to do a new calculation!