Monday, February 7, 2011

An accident waiting to happen

The trouble with rules is that sometimes they are in mutual conflict and it isn't always obvious which rule takes precedence.  Isaac Asimov gave his robots three rules and explicitly gave them a priority order.  But in bridge, it isn't always so clear.  Which rule is more binding may be "obvious" to one partner, while the other partner thinks a different rules is paramount.

Kim and I had just such a misunderstanding.  It was a bomb waiting to go off.  Fortunately for us, the bomb exploded alright but it blew up in the faces of our opponents, not us.

Here's the dichotomy in particular:
  • Rule "1" -- when one player bids notrump to play*, all subsequent doubles are for penalty (* "to play" is defined as any natural notrump bid in the face of competition or at the two-level or higher);
  • Rule "A" -- advancer of a notrump overcall bids with systems on^, that's to say as much as possible, we stick to our standard responses (^ "systems on" in our case means [after 1NT] Stayman, four-suit transfers, Texas transfers, various 3-level bids; [after 2NT] Romex Stayman, Jacoby transfers, minor-suit stayman).
In the event that the opponents intervene over 1NT, we use Lebensohl and negative doubles at the two and three levels (except that double of any 2C bid is always Stayman).  It's a lot less clear what we do over interference over a 2NT opener.  In my view, 2NT is "to play" by our definition (two-level or higher) and therefore partner's double would be penalty.  But partner sees a double here (other than over 3C which would be Stayman) as negative.  Who's right?  It all depends on which rule ("1" or "A") takes precedence.

Life becomes further complicated when we overcall notrump.  Again it's "to play" (bid in the face of competition) but what about "systems on"?

So here's the hand where this came up.  I held ♠KT953 AQ5K8 ♣AQT in third seat (no-one was vulnerable).  Kind of an awkward hand, right?  I suppose I have to open 1♠ and then rebid 2NT over 1NT but then partner will be declarer and my hand which is chock full of tenaces will be exposed as dummy.  No matter, my RHO came to the rescue with 2.  Now, this was easy: 2NT showing a balanced 15-18 possibly with a five-card major (since we play Romex including opposite an overcall).  LHO now raised to 4 and partner doubled with ♠AQ72 K732T2 ♣753.

Here's how I interpreted that double:  I bid two NT in competition so it's 100% penalty.  Here's how partner interpreted it: the opponents have strongly bid and raised a suit and our doubles are always for takeout (or two-way) in such cases.  We play systems on, and therefore double should be for takeout [we had never actually discussed a four-level overcall of a 1NT or 2NT opening].

I was of course overjoyed to pass this double, whatever its meaning and we duly collected 1100.  We could have made slam for 980/990 but in any case nobody else bid it (although all made 12 tricks in spades).  6NT (on 27hcp!) is the best contract but from my side of course in case the A is in the wrong place.  It was only afterwards we discovered that this hand could so easily have been a disaster.

The last word?  I think that there are always three major considerations when discussing the meaning of calls: efficacy, frequency and simplicity.  To a point, I'm a big fan of simplicity when it comes to doubles, which is why I have developed my set of rules.  What about frequency and efficacy?  LHO has bid something, partner has overcalled in notrump, and RHO has raised.  Are we more likely to want to double for takeout (i.e. negative) or to penalize?  To be honest, I think we're much more likely to want to double for takeout so that "systems on" should take precedence, thus supporting Kim's use of the double card.  But, suppose that RHO didn't raise but bid a new suit.  It's almost impossible for us to have much in the way of values now.  Do we really want to make a takeout double?  I don't think so.  The only time I think we'd want to double would be when we happen to have some values in RHO's suit.

Really, I think the last word can be summed up in the typical experts' response when asked about this kind of double: "cards".  He who made the notrump overcall has to decide what to do based on bridge logic.  At IMPs and sometimes at matchpoints, the expert will generally "take the money".  The rest of the time?  Your guess is as good as mine.

Honey, let's agree that 2NT (3) X would also be negative (by extension of the 1NT sequence), thus overriding the other rule that a bid of 2NT is "to play".  We'll agree therefore that no opening notrump bid is ever "to play".  The next time it comes up, in about 2015, we should be well prepared :)


  1. I would have overcalled 2S on your hand. Seems a bad idea to hide 5-card majors when overcalling.

    Also, 2NT(3D)X being negative does not seem very good. Brian once told me that if double is negative, then he will interfere over 2NT with any hand, since he can't be penalised.

  2. I agree with double at high levels as showing cards when neither pure penalty nor pure takeout seem clearcut.

    As for 2NT(3D)X being negative, I'm more optimistic than Vincent. I do know that I've had good results with 1NT(3Y)X being negative, because without a fit, it's easy for opener to pass. I can see that working quite well after a 2NT opener.

  3. I agree with Vincent that an overcall of 2S is the preferred call. My second choice would be dbl. I don't care for 2NT because I can anticipate my diamond stopper being knocked out early. If partner does not deliver a second diamond stopper, that means I will need to run seven more tricks right away. That would require too many specific cards from partner for my taste. Not to suggest that 2S is bullet-proof ...

    Never thought much about the meaning of 2NT-(3X)-dbl. But I believe I prefer the pure penalty treatment. Takes very little (maybe a three count and three trumps) to expect a nice result from a penalty double, while it could be tough to unravel our best fit if starting at the three level over a negative double.