Wednesday, July 28, 2010

When attitude is known

It's customary when signaling to show attitude when we lead a suit.  If they lead a suit, our attitude is assumed to be bad and we skip to showing count.  If both attitude and count are known, or if bridge logic says that neither of these is important (e.g. a singleton in dummy), we skip to suit preference.

So much is reasonably standard.  Some people like to show count even on the opening lead, others on the opening lead if dummy presents a certain number of cards in the suit, etc.

In this blog entry, I would like to propose a variation: when attitude is already known from the auction, we give count on the first trick.  It's a method I've been playing for a while with one of my partners.

How does it operate?  When is attitude already known?  The premise is that when we get to show a good suit during the auction, our attitude is considered known (and good).  What constitutes showing a good suit?
  • an overcall;
  • a rebid of a suit;
  • a "free" bid in a suit (when pass would be a valid alternative) [according to my principle of "stuff"];
  • a lead-directing double of an opponent's artificial bid;
Thus, if we are not obliged to try to win the trick, for instance when partner leads a high card, or we cannot beat the dummy, our carding shows count, not attitude (when attitude is known from the auction).

Here's why it works: in all of the cases given, we have suggested length and strength in the suit.  Because of the length, the suit will not be standing up for very many tricks.  But how many tricks?  That's why count is so important.  Fie, I hear you say, sometimes I make bad overcalls.  Well, that may be true, but assuming partner is going to lead your suit anyway, the damage, if any, will already be done.  Much of the time the play to the first trick will clarify the position.  Yes, it's possible that a tempo or even a trick may be lost when partner gets in and, assuming good attitude (because you weren't able to discourage at trick one because of the obligation to show count), leads the suit a second time.  But for that to cost, four conditions must be met:
  1. you have to have made a questionable bid during the auction;
  2. you must have been in a position to signal at trick one (i.e. you were not trying to win the trick);
  3. declarer/dummy must have sufficient cards in the suit for it to make a difference;
  4. it must not be obvious from dummy's holding what's going on in the suit.
The chance of all these happening at the same time is actually quite small.

The same idea applies when the bidder is the one leading the suit.  On opening lead, it is normal to show count in any case so there's really no difference there.  But during the hand, it's common to make attitude leads of new suits.  Again, it's better to show count when we are leading our own "good" suit.

Note that this scheme may also apply (according to partnership agreement) when partner is leading his own known-to-be-good suit: our carding should show count if we can't win the trick.

As always, comments welcome.

Thursday, July 15, 2010

Light third-hand openers

I've never found a really good written formula for when and how to open a sub-standard hand in third seat.  I've presented some ideas myself earlier in this blog: Third and Fourth Seat Openers.

A hand came up yesterday evening at the club which suggests a new rule for light third-hand openers: once partner has raised your "suit", never bid a new suit (unless partner forces you to).  This should never be necessary if you start with your best suit.  My partner picked up the following hand: ♠– AJT8 KQ432 ♣T843.  Only the opponents were vulnerable.  After two passes, partner bid 1.  My hand was ♠A976 Q53 J6 ♣K975.  My RHO passed and I contributed 2.  This went around to my RHO who backed in with 2♠.  Considering that I had a flat maximum, and somewhat forgetting that I had passed originally, I doubled.  Partner now felt that 2♠ was a likely make (it was) and decided to retreat to 3, giving me a choice of red suits.  I took this as a good distributional but solid opener, and jumped to the heart game.  My LHO was happy to double this and we went four off for a round zero (even two down would have given us the same matchpoint result).  Double-dummy, we can make 1, 2 or 4♣ (the par result).

So, what lessons if any are to be learned?  I've always believed that a passed hand should not take any questionable actions.  Doubling 2♠ was not automatic so I should not have done it.  But I think partner should have opened the hand, if at all, with 1 (his best suit).  If I bid 1♠ (as I surely would, assuming no intervention), he can then retreat to 2♣.  I would assume a "full" opener of course but the hand really is a full opener (27 Zar points which is an automatic opener in any seat) and, as noted, we can actually make 4♣.

So let me restate my guidelines for light third seat openers (with the new addition):
  • Open only in a good suit of four or more cards, one which you'd like led [no "prepared" bids];
  • Be ready to pass partner's one-level new suit bid [you can't rebid 1NT because you have too few points, by definition];
  • Unless there's a suit you really want led, tend not to open with a very balanced hand, especially vulnerable;
  • Don't bid a new suit of your own unless partner makes a forcing bid (he shouldn't, as a passed hand, unless he has a very good hand and fit for you).
Here are a few hands that I think should open 1 in third seat:
  • ♠A86 KQ93 J6 ♣T975 (probably should pass this if vulnerable);
  • ♠K862 AQT5 6 ♣T975, assuming you can't open this with an artificial 2 (if partner bids a natural 2, you'll just have to suck it up);
  • ♠A86 KQ963 86 ♣T95.
 Here are a few hands that I think should probably be passed in third seat, especially if vulnerable:
  • ♠A86 KQ93 J65 ♣T97;
  • ♠A86 KJ93 Q5 ♣T975;
  • ♠A6 K83 Q65 ♣QT975.
Comments welcome.

Tuesday, July 6, 2010

Protecting Equity

The concept of protecting your equity applies to any form of bridge scoring and at any bidding level but it is especially pertinent for part-scores at matchpoints.  The idea is that if you think you were going to make your contract and the opponents bid on, you have to double them in order to protect your equity.

In the double-or-pass decision, only their vulnerability is relevant.  In most situations, where our part score would score more than 100, doubling has much more to gain if they are vulnerable. Otherwise, doubling only makes good sense if we expect to set them two or if we were planning to get 90 (in 1NT or 2-of-a-minor).  Doubling to protect our equity in the latter case would be extremely rare.

Let's give an example.  Only they are vulnerable and both sides have found a fit.  We have bid 3H, expecting to make (+140) and the opponents counter with 3S.  We basically have two choices: bid on, hoping that we can actually make 4H (or be down only 1 while they can make 3S) or double them, hoping that they can't make 3S.  There is of course the option of passing, which is appropriate if we are vulnerable and/or we think that they have the balance of power.  If we pass and they are down 1 only, our +100 will not score well against those making +140, which is why we would want to double to protect our equity.

Let's say we encounter this board in the last round of a 13 round duplicate.  So far, the board has been passed out once and 3S their way has made twice and gone down one undoubled twice.  3H our way has made four times and gone down once.  One pair our way went down in 3NT doubled and another pair actually made 4H.  Assuming that our defense is going to be accurate, we expect to score +100 as is, earning a 6 (on a 12 top) for an average score.  In other words, if we do nothing, we will get an average board.

We might make 4H (for 11.5) but at the risk of going down (only 2.5 if they double) [gain:5.5 or lose 3.5], but let's say our hand is fairly balanced in light of the auction so we are not seriously considering 4H.  If partner still has a bid, he might bid 4H regardless of what we do, but let's suppose that whatever we do will be final.  Let's say that we decide to double to protect our equity.  If we're right, we will earn an 11 (gain 5).  If we're wrong, we will get an absolute bottom (lose 6).  On the other hand, if we had allowed them to make 3S undoubled, that would have scored only 1 matchpoint so, in fact, our double actually has a rather good reward/risk ratio: 5 to 1. 

Note how important it is that they are vulnerable.  I was reminded of this by my partner Steve yesterday in the morning STAC game.  On board 16, I elected to double 3S.  On paper, this wasn't a terrible decision, because I rated to gain 4.5 mps on a 15 top (there were 9 pairs in 3S the other way) and, assuming that we weren't setting it, we probably were only getting something like 6 mps anyway.  But now the reward/risk ratio is actually only 3/4 (nothing like the 5/1 we had above).  This is primarily because they weren't vulnerable!  So, our hypothetical gain was only from beating the 9 defenders of 3S (who we would otherwise have tied).  Because of the opponents' lack of vulnerability, the hoped for +100 wouldn't have restored our equity which in this case was +110 for 3C.  Assuming that there were a few pairs our way making 110 or 130, we'd have gained a whole point against each of these if the opponents were vulnerable, meaning that our R/R ratio would at least be greater than 1.

As it happens, of course, we couldn't defeat 3S (actually we let them score an overtrick) and 4C would have been down 2 vulnerable, possibly doubled.  Despite our 21 hcp, they can take 9 tricks in spade while we can take 8 tricks in clubs.  The commonest number of total tricks (17) but on this occasion biased in favor of the other direction.  Allowing them to make 10 tricks would only have earned us 3 matchpoints so in a way we only lost those 3.  But maybe our defense would have been more passive if we weren't trying to set the contract, in which case we'd have scored 7.

Saturday, July 3, 2010

Jacoby Two-Notrump Rebids

Not much to write about these days from the bridge trail.  I haven't seen any really interesting hands.

However, one topic has been on my mind for a while and an example came up just the other day in the STAC tournament (Friday morning, board 4).  As dealer, with both sides vulnerable, you pick up the following hand: ♠QJ953 3 KQ ♣AK542.  You have a decent 15 point hand although you couldn't say that it was a great 15 points.  Anyway, with silent opponents throughout, you open 1♠ and partner responds 2NT (game-forcing with four or more spades and either a good 13 or more hcp or a balanced hand, or both).  What should your rebid be?  What is the most useful thing for partner to know?  The singleton 3 or the good club suit?  Of course, that depends on partner's hand.  But one thing is for sure, partner will know you have a singleton (or void) regardless of which bid you choose.  A computer simulation is probably the best way to determine the best call. With ♠QJ953 3 K5 ♣AKQ42, I think that 4♣ stands out a mile.  Even with the actual hand, I think that 4♣ is best, even though it uses up more bidding space.  On this particular hand, responder held ♠AT72 AK865 92 ♣Q6.  Given that we're off an ace and the trump K, no good pair is going to bid the slam (which makes by the way).  Nevertheless, the responding hand will have cause to get very excited and bid 4 over 4♣ because whether opener's shortness is in hearts or diamonds, it really doesn't matter much as there's no wastage either way.  Over the actual 3 rebid, responder bid 3♠ leaving the door open but we ended up subsiding in 4♠.  However, if opener's hand had been ever so slightly different, ♠KJ953 3 KJ ♣AK542, we probably would have missed an easy and biddable slam.

So, what sort of suit quality does a 4-level rebid promise.  I'd say the usual "two of the top three" honors is about the minimum (similar to a positive response to a 2♣ bid).  Missing the ace or king, we'd probably want the J too, or at least the ten.  So AKxxx, AQJxx would all qualify but nothing less.  KQTxx would be OK when every other aspect of the hand is perfect.