August 26, 2013 – When Should We Double Partscores?

The short answer to my title question, at IMPs, is never.  That is an exaggeration, of course, but … IMP odds make doubling a part-score quite dangerous.  As a simple example, let’s suppose the opponents have bid up to three spades, and you are contemplating pouncing.  We’ll say that they are not vulnerable, and your teammates are allowed to play there undoubled, and we will also say that the contract will usually be set.  Here are some numbers:  Suppose, out of four hands, three spades will make once, go down two once, and down one the rest of the time.  So, 75% of the time they are going down, which should make your double pretty good.  What would happen?

You gain 5 IMPs when they go two off, 2 IMPs twice when they are down one, and lose 9 on the unlucky layout where three spades makes.  Net IMPs – a complete wash.  Think about that.  They are going down most of the time, yet your double gains zero IMPs, on average.  So, next time you feel tempted to double some part-score, think again, and then pass. 

There are two exceptions to this, and these exceptions occur because the IMP scale rewards plus scores.  These are:

 

1.  Bidding on rates to be poor, or
2.  Doubling will help partner on defense.

The first case arises frequently when partner has made a take-out double at a high level.  You may not beat their contract, but bidding to your own contract looks really hopeless.  When your last chance at a plus score is defending, then defend.  It is better to defend undoubled, but, if partner has already doubled, you have no choice.  Consider a hand like

The bidding, with both sides vulnerable, goes 3S Dbl P to you, and?  You have no reason to think they are going down, but what choices do you have?  If, by some miracle, you could make, say four clubs, partner will bid five.  If you pass, you might actually go plus.  And, in practice, -730 could win IMPs against the 800 or 1100 awaiting you if you bid. 

For 2), again let’s suppose the contract is three spades, which may go down if partner can find the right opening lead (maybe you can get a ruff).  Doubling this might work out well, particularly if you weren’t going to set the contract without your double.  Now the double, when it works, wins six IMPs, and loses 9 when it fails.  For this type of double to show a profit, you need to be right a bit more than 60%. 

Our crew doubled two part-scores on Sunday, and set one of them.  They gained 8 IMPs on that hand, and lost 9 on the one that made, for a net loss of only 1 IMP.  But, had they not doubled either contract, they would have gained 6 IMPs on the successful double anyway, and pushed the other hand.  So, this aggressive doubling was really minus seven IMPs on these boards, not just one IMP. 
Here was the unsuccessful double:

North held  

and heard this auction, with nobody vulnerable:

This seems like a very poor penalty double to me, but I am pretty stodgy, and never double part-scores.  I can think of three reasons why North doubled:

 

1.  I have ten points and partner opened the bidding, so …
2. My redouble created a forcing auction, so I had to double, or bid four hearts. 
3.  I have clubs locked up, which is likely to be West’s source of tricks.  This contract rates to go down on a trump lead, and partner is much more likely to find a trump lead if I double.

Reason 1 is pretty lame.  East rates to have a lot of shape for the three diamond call, and West invited a game, so what makes my ten points enough to stop eleven tricks?  The second reason is more interesting, and something well worth discussing.  In my book, power auctions which show at least game invitational auctions, like cue-bids or these redoubles, create forcing auctions at low levels and at high levels, but not in between.  Once you redouble, partner’s pass over a one diamond or two diamond bid is forcing.  But the pass over three diamonds is not forcing.  Likewise, had East jumped to five diamonds, partner’s pass would be forcing. 

     The redouble creates a forcing auction for us except at the three and the four level.

Those are the rules in my partnership, but you may have different guidelines.  Discuss these with your partner and with your mentors. 

Reason 3 is quite powerful, and, as the full hand shows, quite prescient:

 
Only a trump lead would set four diamonds.  Partner would certainly lead the spade king without a double.  Would he find the trump lead if you double?  Not this time.

The two hands that most appealed to me were basically pushes – only one IMP exchanged hands, but both were quite interesting play problems.  First up, try to land this four spade contract:

   
With no one vulnerable, the bidding went:

First question:  Do you agree with North’s four spade call?

I don’t.  The club honors won’t be worth much on offense.  Either partner is very short in clubs, or East can ruff a club.  In a spade contract, the North hand is closer to xxx KQx J10xx xxx, hardly a four spade call.  But the clubs are two sure winners on defense, so I would rather defend.  I would pass North’s hand, and pass again if partner reopens with a double.

Now to the play.  West leads the diamond two (using third-fifth leads).  East wins the ace and returns the four.  Do you play the nine or the king?

The queen of diamonds may well be off-side, but, if so, there is nothing you can do.  You’ll lose two diamonds, a club, and likely a trump or two.  So, you must finesse.  It can’t hurt.  One declarer played the diamond nine (at the other table, on a different auction, South tried the diamond king).  West trumps the diamond nine with the spade queen – interesting, lays down the club ace, and continues clubs.  You win in dummy as East follows.  Two more questions:

 

1.  What do you discard on the club?
2.  Is that spade also a singleton?

The discard is easy.  We have a possible heart loser, so throw that.  For 2), restricted choice tells us that Q alone is more likely than QJ- doubleton.  There are more possibilities here, which I will come back to, but the singleton spade seems likely.  More importantly, West bid four clubs on a poor suit, suggesting lots of distribution.  Which is more likely – Q J109x x AJ9xxxx or QJ J109 x AJ9xxxx? 
So, East has the J982 of trumps left, but we have three dummy entries to take three finesses.  Lead the trump planning on finessing with the six or seven next.  If East splits, use our heart entries to pick up trumps.  The full hand was, as expected:

   
Could the defenders have done any better?

Yes.  You lacked the third entry to the table, and could not pick up East’s trumps, until West continued clubs.  Without this help, the hand couldn’t be made.  Is West at fault?

Not entirely.  The diamond six looked like a suit-preference card, but it was hard to read.  East might well play the six back from AQ64 or AQ86, or, as here, from AQ864.  Was it high?  Low?  Middle?  East does not want partner trying to give a club ruff, and should suit-preference unambiguously high, returning the eight, not the six. 

 Here is a fantasy hand.  Imagine this rose in a knock-out match:

Both tables reached four spades after a one spade opening and a three club overcall.  The defense got their diamond ruff and ace of clubs, but declarer simply cashed a high spade, felling the jack, crossed to dummy, and played a trump to the ten.  +420 at both tables, with no thoughts at all about the hand. 

Think for a moment, though, about what might have happened had West trumped the diamond with the jack rather than the nine.  Trumping with an unnecessarily high card has gained a surprising number of tricks for me over the years.  West could even try this with Q9 or Q8 in spades, trumping with the queen.  That might set a cold contract on a layout like this, or hand declarer a hopeless contract when partner held, say, Kx in trumps rather than J8x.  Being tricky is risky, but often profitable. 

For my other hand, try playing in six clubs on this layout, with the heart four lead (fourth best leads):

  
Suppose, first, that there was no opposing bidding.  The contract looks quite secure, with only one possible snag.  What could go wrong?

Only a four-zero trump split will cause problems.  If East has all four trumps, we will make the hand easily by finessing East for the ten eventually.  So long as we don’t play the ace of clubs early, or a trump to the jack, the hand is simple.  Can we make the hand if West has all four trumps?

Yes, if the distribution is friendly, and we read West’s shape.  Either way, we should win the heart in hand and lead a low trump to the queen.  If West shows out, easy.  If East shows out …

How can the hand be made if West was originally 3-3-3-4?

After the club queen holds, heart ace, throwing a diamond, heart ruff, spade ace, spade ruff, diamond to dummy, spade ruff, and two more diamonds.  At that point, we will have AJ9 in clubs left, and West will be down to K108.  Exit with the club nine or jack.  You can follow the play via the next button.

Can the hand be made if West was originally 3-4-2-4?

Yes, but we will have to discard two diamonds on winners.  That means we will need a spade trick, and could finesse either way.  If we judge that West has the spade king, we play heart ace, heart ruff, spade to the queen, spade ruff, diamond to dummy, spade ace, heart ruff, diamond. 

If, instead, we think East has the spade king,

heart ace, heart ruff, spade to the ace, spade queen, throwing a high diamond, spade ruff, diamond to dummy, heart ruff, diamond. 

At both tables, North opened 1NT.  At one table, East overcalled 2D, showing both majors.  South bid 2H, a game force with clubs, and drove to six clubs.  South won the heart lead and correctly played a low club to the queen.  Nice play.  We’ve already seen how the hand can be made if East was 5-5-3-0.  Could the hand be made if East was 6-5-2-0? 

Now we will have to cash all four diamonds, so we don’t want to throw a diamond on the heart ace.  A line that will work is:  Trump queen, heart ruff, spade ace, spade ruff, diamond queen, heart ruff, diamonds. 

What is the best way to combine all of our chances?

My choice is this:  Win the heart and play a club to the queen, as East discards.  Throw my spade on the heart ace, and trump a spade.  Then diamond king, diamond to dummy.  At this point, I have seen West’s heart play (and a discard from East), so should have a good idea whether West started with two hearts or more.   Trump another heart, if that looks safe, or, if you think hearts were 2-6, trump a spade.  Then another diamond to dummy.  This will leave either:

  
or

 
At this point, you will know the diamond count.  If West held four diamonds, he has left a diamond, three trumps, and one other card.  Guess which suit that is in, trump that suit, and cash your diamond.  If, instead, West held three diamonds, cash the spade ace and trump the right suit back. 

At the other table, East overcalled a natural two hearts, and South drove to six clubs.  Again the lead was the heart four.  How should this declarer play?

Here, there is a real danger that hearts are 1-7.  Indeed, hearts will be 2-6 only when West has exactly the four-three doubleton.   Leading a low club to the queen could easily cost the contract if East wins and plays a heart.  With this information, it is substantially better to play the club ace, and another club.  This is how our declarer played.  Again, well done!  Since the full hand was:

 
the club ace snared an overtrick IMP.  Well-bid, and well-played at both tables.