Zenith Daylong Reward - April 13 - Board 14

Board 14
Neither vulnerable

Last week, after 382 posts, I broke my commitment to choose deals for this blog at random. Now that there's a precedent, I feel emboldened to do so again. This is how slippery slopes work. But there are so many fascinating aspects to this deal that I had to post it. It features an interesting bidding problem. It features a challenging play problem. It illustrates the biggest flaw in the way bridge robots are designed. And it finishes with a Shyamalanian twist ending.

First, some preliminary remarks. Gabrielle Kamil, who has managed some impressively high scores in robot individuals, told me that in these events you should routinely open one notrump with 14-point balanced hands. I wasn't convinced. But to show how open-minded I am, I decided to keep a log. I sometimes evaluate a 14-point hand as a strong notrump. But if don't on a particular deal, I will record my result along with the average result of everyone who opened one notrump. If, over time, their results are better than mine, I'll change my strategy. 

Unfortunately, if their results are worse than mine, the test will be inconclusive. What I'm really interested in is what would happen if I had opened with notrump. Since there is no way to know, so I'm using the field average as a proxy. In practice, I suspect my results would be better.

I began this project recently, so today's deal is only the second entry in my log. On the first entry, I scored 61% opening one club with

♠ K 9 8 7  ♥ A J 5  ♦ K 10  ♣ K 9 6 3.

Those who opened with one notrump averaged 21%. (Partner has a bad hand. If you open with one notrump, you go minus. If you open with one club, you let the opponents into the auction. They reach a good game in their four-four spade fit, but it goes down on the bad trump break.)

This time, I pick up:

♠ A K J 5  ♥ A J 10 6  ♦ J 7  ♣ 6 5 4.

This hand looks a lot more like a strong notrump opening than the previous one. It has three and a half-plus honor tricks, so it qualifies by pre-HCP standards. But I still don't like one notrump, because I hold both majors. If partner has less than nine HCP, opening with one club may enable us to find a four-four major-suit fit that those who open with one notrump will miss.

I bid one club, and partner bids two clubs, showing a limit raise or better in clubs.

Now I wish I had opened with one notrump. I want to bid game. But two notrump by me is non-forcing, and three notrump shows 18 or 19 HCP. The standard way to solve this problem is to temporize with two of a major, then bid three notrump. But I'm disinclined to do that. If I bid two hearts and partner bids two notrump, we are probably playing this from the wrong side. I'd like to make sure to get the notrump bid in first.

Maybe I can. Since this is a best-hand tournament, it is unlikely partner will have a bid over three notrump anyway. So I doubt it will hurt to let him believe I have 18 or 19 HCP. In fact, it may help, since it may induce the opponents to misdefend. I once saw Pete Hollands make this bid for similar reasons. Accordingly, I bid three notrump. Partner raises to six notrump.

Well. That didn't work out. Maybe this was a bad idea. Now that I think of it, when Pete did this, his partner was a passed hand. The one-notrump openers will say I got exactly what I deserved. I underbid with one club, then grossly overbid with three notrump, trying to solve the problem I had created for myself. 

There's nothing I can do about it now. I pass, and West leads the spade three.

	NORTH Robot ♠ Q 10 4 ♥ 8 7 ♦ K 10 2 ♣ A Q J 10 9
	SOUTH Phillip ♠ A K J 5 ♥ A J 10 6 ♦ J 7 ♣ 6 5 4

West	North	East	South
Robot	Robot	Robot	Phillip
		Pass	1 ♣
Pass	2 ♣	Pass	3 NT
Pass	6 NT	(All pass)

Six notrump seems a bit aggressive with partner's hand. But partner is closer to his bid than I am to mine, so I can hardly complain. What do I need to make this? I need the club finesse for starters. That brings me up to ten tricks. If the ace and queen of diamonds are onside, maybe I can score two diamond tricks to come to twelve.

The spade three is hard to read. It could be lowest from three or four, fourth best from five, or high from a doubleton.

I play low from dummy. East plays the eight, and I win with the jack. I play the four of clubs--three--ten--eight. So far, so good. I need to repeat the club finesse and take two diamond finesses, so I need three entries to my hand. If spades are three-three, I have two spade entries, but using the heart entry is problematic. If I lead a heart to the ace, I set up heart tricks for the defense.

Perhaps I can squeeze West down to a stiff heart. If I play a spade to my hand, repeat the club finesse, run clubs, and play another spade to my hand, I come down to this position:

	NORTH Robot ♠ -- ♥ 8 7 ♦ K 10 2 ♣ --
	SOUTH Phillip ♠ 5 ♥ A J ♦ J 7 ♣ --

Now the last spade squeezes West. If he comes down to ace-queen tight of diamonds, one diamond play from my hand suffices. If he comes down to three diamonds and one heart, I can play a diamond to dummy and return to my hand with the heart ace and impunity.

Another possibility is simply to cash the club ace, hoping the king drops. If it does, I need only two entries to my hand, so I don't need to play for three-three spades.

Which layout is more likely? Three-three spades or a doubleton club king? Before I decide, there is a third line to consider. I can play a heart to my jack. If West doesn't hold the diamond ace, he will expect me to have it for my three notrump bid. So, after he wins the heart, he will see no reason to break diamonds. He will probably exit passively with a spade. Then I can cash the spades and clubs, coming down to this position:

	NORTH Robot ♠ -- ♥ 8 ♦ K 10 ♣ 9
	SOUTH Phillip ♠ -- ♥ A 10 6 ♦ J ♣ --

If East began with four hearts, the last club squeezes him.

East can thwart this plan by hopping with his heart honor. But I can't imagine anyone's finding that play from honor fourth. This looks like a better plan to me than either of the previous two lines I considered.

I play the heart seven; East, the deuce. Does it matter whether I play the ten or the jack? Probably not. But the ten might make my heart suit look too threatening. The stronger West thinks my heart suit is, the more likely he is to play his partner for the diamond ace in desperation.

I play the heart jack, and West takes the queen. He doesn't find the diamond shift. But, unfortunately, he continues hearts. That's another way to thwart my plan. The heart continuation destroys my entry for the squeeze. I play dummy's eight, East plays the nine, and I win with the jack. The good news is I may not need the squeeze. Dummy's hearts were eight-seven. So if East began with king-nine third, my heart six is good.

Do I have any chance if East began with four hearts? Maybe I can execute a pseudo-squeeze. I can cash the clubs, pitching two diamonds, then cash spades, finishing in my hand. If East thinks I'm 4-3-3-3, he will hold the diamond ace instead of the long heart at trick thirteen. I could cash the heart ace first to see if the king drops, but I see no reason to do that. If it drops now, it will drop later. And cashing the heart ace may make it easier for East to count my hand. If there is only one small heart outstanding, East may be suspicious when West doesn't pitch it.

I play the club five--seven--jack--diamond five. Oops. Clubs are four-one. Now I have a new problem. I need to repeat the club finesse. Since I have to win trick twelve in my hand for the pseudo-squeeze to work, I need two hand entries. I'm back to overtaking the spade and hoping for three-three spades.

Or am I? If the heart king is dropping, I don't need the pseudo-squeeze. I have the rest in top tricks. Maybe there was a reason to cash the heart ace after all. Then I would know for sure whether I had to overtake the spade.

Can I infer how spades split? East has a stiff club, so it's unlikely he has a doubleton spade. That means West's spade three was probably lowest from three or high from three-deuce doubleton. I'll play the spade ten and see what East follows with.

On the spade ten, East plays the deuce. It appears, then, that West's lead from lowest from three. I overtake the ten with the king and repeat the club finesse. Then I run clubs and overtake the spade queen. Luckily, both opponents follow. I have now reached this position:

	NORTH Robot ♠ -- ♥ -- ♦ K 10 2 ♣ --
	SOUTH Phillip ♠ 5 ♥ A 6 ♦ -- ♣ --

On the last spade, my opponent pitches a heart, clutching his diamond ace. My heart ace drops his king, and my heart six takes the last trick. Making six.

So where is the Shyamalanian twist ending I promised you? It turns out it is West, not East, who is pseudo-squeezed. This is the full deal:

	NORTH Robot ♠ Q 10 4 ♥ 8 7 ♦ K 10 2 ♣ A Q J 10 9
WEST Robot ♠ 7 6 3 ♥ K Q 5 3 ♦ A 6 ♣ K 7 3 2		EAST Robot ♠ 9 8 2 ♥ 9 4 2 ♦ Q 9 8 5 4 3 ♣ 8
	SOUTH Phillip ♠ A K J 5 ♥ A J 10 6 ♦ J 7 ♣ 6 5 4

Before you are too critical of West's defense, let me point out that he defended correctly according to his programming. When he was in with the heart queen, he knew I had at most ten tricks. He could cash the diamond ace for down one. But if his partner had the heart ten, he could beat me two by continuing hearts.

Double-dummy, continuing hearts can't cost. At worst, it gives me the heart ten for my eleventh trick. So playing a heart either gains or breaks even versus cashing the diamond ace. But while the play entails no risk double-dummy, it does entail a risk single-dummy. It requires West to read my shape and discard accordingly.

Since plus fifty is probably a 100% board anyway, taking any risk at all of letting me make this can't be worthwhile. But the robot doesn't even see that there is a risk. As far as he's concerned, a heart can't possibly cost. This is the biggest flaw in the robots' design. In analyzing the play, they assume everyone is double-dummy. Usually in making this assumption you give your opponent too much credit. Sometimes, as here, you give yourself too much credit. 

Of course, while West's defense may be understandable given his limitations, I can't say the same for his opening lead. Or his final pass. I don't understand either one.

I wish I could replay the board to see if my choice of the heart jack rather than the ten made a difference. If I played the heart ten, would West think it was impossible for his partner to hold the jack and just cash the diamond ace? He already knew I didn't have the 18 HCP I promised. Who knows what assumption he would make about the heart jack?

Anyway, I now have entry number two in my log. Opening with one club scores 100%; opening with one notrump averages 54%. 

And, by the way, I was wrong about this being a best-hand tournament. While most robot individuals are, the Zenith Daily Reward tournaments are not. It's fortunate I didn't remember that. Had I done so, I never would have bid three notrump.