Free Weekly Instant Tournament - December 13 - Board 6

Board 6
Opponents vulnerable

♠ 6 5 2  ♥ A 9 7  ♦ A K 9 8 4  ♣ K 4

RHO opens with one diamond in first seat. Three and a half honor tricks and a five-card suit qualifies as a strong notrump even if I am missing a high-card point, so I could overcall with one notrump. But, since the opponents are vulnerable and have opened in my best suit, it might be better to pass and see if they get into trouble.

I pass. LHO tries to stay out of trouble by passing also. Partner balances with one heart, and RHO doubles for takeout.

One could argue that two diamonds here should be natural. The primary reason to play that a cue-bid shows support is to allow you to play a jump raise as pre-emptive. But you don't need a pre-emptive raise when LHO has passed. You can simply raise hearts to whatever level you think is appropriate. 

But the robots do play two diamonds as a cue-bid showing heart support. So I might as well take advantage of that fact to offer a choice of strains. I can bid two diamonds to show my heart support, then bid two notrump to show just short of a one-notrump overcall.

Over two diamonds, partner bids two hearts. I bid two notrump as planned, and partner bids four hearts. RHO leads the heart ten.

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

West	North	East	South
Robot	Phillip	Robot	Robot
1 ♦	Pass	Pass	1 ♥
Double	2 ♦	Pass	2 ♥
Pass	2 NT	Pass	4 ♥
(All pass)

One of the first bridge books I ever read was Bob Ewen's Opening Leads. Ewen states that, when leading a trump, one should lead second highest. The idea is to retain your high spot for a possible overruff if you have a doubleton and to start a trump echo, to show a third trump, if you have three. So far, his suggestion is mainstream. But Ewen continues to say that "second highest" applies even when you hold an honor sequence. Thus one should lead the ten from jack-ten and the nine from ten-nine.

This method makes sense. If you lead second highest from non-touching cards (which is a common agreement), what do you gain by switching to highest when you hold a sequence? That simply introduces an ambiguity. I can't think of any reason not to play Ewen's method.

Ewen presents this as if it were expert practice. But, as I discovered later, it isn't. Not only have I never found anyone who plays this way, I've never found anyone who has even heard of this approach. That surprises me, since Ewen's book is hardly obscure.

That's not to say I've never seen anyone lead the ten from jack-ten of trumps. But they don't do so by agreement. They do so either for deceptive purposes or as some sort of mis-guided attempt to give suit preference. So, against a human, I would never assume from this opening lead that the heart jack is on my right. But against robots, I think that's a safe assumption.

Still, there is no reason not to guard against the possibility that I'm wrong. Rather than win in dummy and take an immediate finesse, I can unblock the nine, win this trick in my hand, then return to dummy with the heart ace for a finesse.

Accordingly, I unblock the nine. East plays the five, and I win with the king.

West probably has the diamond queen and presumably at least four of them for his opening bid. Let's say I draw trump now. If East has four of them, I will discard a spade from dummy on the fourth round. Then I will lead the diamond jack. West will cover, and I'll win in dummy.

If East doesn't play the ten, I know West has it, so I'll play a spade to the queen. West will take his ace and l return a spade. I'll win and play a diamond to the eight, reaching this position with the lead in dummy:

	NORTH Phillip ♠ -- ♥ -- ♦ K 9 4 ♣ K 4
	SOUTH Robot ♠ 3 ♥ 3 ♦ 3 ♣ 3 2

If I were in my hand, I could repeat the diamond finesse, take two pitches, and lose only the club ace. making five. But I'm in dummy. So I must lose the club ace and another spade and make only four.

No, wait. It's worse than that. I have no way to get off dummy without losing control. If I play ace and ruff a diamond to lead up to the club king, they can hop and cash two spades for down one.

Let's back up. On the fourth round of trumps, I must pitch a diamond from dummy. Now, after I take the second diamond finesse, this will be the position, again with the lead in dummy:

	NORTH Phillip ♠ 6 ♥ -- ♦ K 9 ♣ K 4
	SOUTH Robot ♠ 3 ♥ 3 ♦ 3 ♣ 3 2

Now I can exit with a spade. If they win and tap me, I can repeat the diamond finesse and pitch one of my clubs to make four.

Can I make five if I manage my entries better? Let's say I lead the diamond jack at trick two. West covers. I win in dummy, draw trump, and lead a diamond to the eight, reaching this position:

	NORTH Phillip ♠ 6 5 ♥ -- ♦ K 9 4 ♣ K 4
	SOUTH Robot ♠ K Q 3 ♥ 3 ♦ 3 ♣ 3 2

Now when I play a spade to my queen and West wins and returns a spade, I'm in my hand for another diamond finesse. I can pitch my spade and a club and lose only the club ace to make five. If I'm going to take three diamond finesses, I need to take one of them now.

The danger in taking a finesse now is that diamonds might be 5--0. If so, I'm going down, losing two aces and at least two diamond ruffs. If everyone is in game, the fact that I risk going down doesn't matter. I stand to gain as many matchpoints as I stand to lose, so I should just take the percentage play. But if some pairs don't reach game, the risk-reward ratio is different. When my play works, I gain half a matchpoint against every pair in game. When it loses, I lose a half a matchpoint to every pair in game and a full matchpoint to every pair not in game. 

Is it worth the risk even so? There are three ways East can have a small singleton diamond and only one way for him to have a void, so an immediate finesse gains three times out of four. Let's say the field consists of four more pairs: two are in game; two are not. The immediate finesse will gain 1 matchpoint (.5 for each pair in game) 75% of the time. It will lose 3 matchpoints (.5 for each pair in game and 1 for each pair not in game) 25% of the time. So the net expectation is zero when half the field is in game. That means if I think more than half the field will be in game, I should lead the diamond jack now. If I think less than half will be in game, I should play safe and draw trump.

I could be wrong, but I think at least half the field will be in game. So I play the diamond jack--seven--four--.

The good news is East doesn't ruff. The bad news is he follows with the ten. That means I don't need three finesses to pick up the diamond suit. Two finesses suffice, so my trick-two finesse was unnecessary. Everyone in game will be making five.

I play a heart to dummy's ace. West follows with the four; East, with the six. I play a heart to my hand, drawing East's last trump, and West pitches the spade four.

Now I play a diamond to the eight and run the diamonds, pitching two clubs. East pitches the spade seven and the nine, five, and seven of clubs, carefully holding all his even spotcards. On the last diamond, West pitches the club ten. We've reached this position:

	NORTH Phillip ♠ 6 5 2 ♥ -- ♦ -- ♣ K 4
	SOUTH Robot ♠ K Q 3 ♥ 8 3 ♦ -- ♣ --

I play a spade--eight--queen-ace. West returns the jack of spades to my king. I cash the last two trumps and West "misguesses" and holds the club ace instead of the spade ten, so my spade three wins the last trick. Making six.

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

Plus 480 is worth 100%. It turns out the last overtrick was only a small bonus. Plus 450 would have been worth 93%. That surprises me. With the diamond ten dropping, making five should be routine.

West made a clear error in the end position. When he was in with the spade ace, he should try to cash the club ace instead of exiting with a spade. If it cashes, great. If I ruff, he knows he has a spade trick coming. Trying to cash the club ace eliminates the need to guess which card to hold at trick thirteen.

Robots make this particular mistake often. It's a flaw in their algorithm. They assume double-dummy play by everyone, including themselves, so they see no difference between a play that is always 100% and a play is 100% provided you do the right think later on. To be fair, we humans sometimes make this mistake as well. But at least we are embarrassed when we do.

The most dramatic example I've seen of this mistake is when a robot went down in seven notrump with thirteen top tricks. I led dummy's void, and declarer pitched from dummy's long suit. Now he had only twelve tricks. In the robot's "mind," he could afford this discard because he had a two-way guess for a queen for his thirteenth trick. So, double-dummy, the discard made no difference. If only robots were allowed to claim when the dummy comes down! Then he would have made it.