Wednesday, November 28, 2012

You just can't make this stuff up!

Some players don't like to preempt in one suit when they have a good holding in a major suit on the side. Others feel that preempting with a void on the side is somehow evil, especially when opening with a weak two. Is it OK to preempt with a side void? How about two voids? Two voids and a five-card major? Just ask rising star Zach Grossack. Here he is playing with Kim against the robots on BBO a couple of months ago. Did someone say "unfavorable vulnerability?"

Even Victor Mollo and his Hideous Hog never quite pulled off anything like this. Note that, despite the defense holding all four aces, all four kings and a jack, they can do no better than three tricks! And, no, this was not part of a "Goulash" tournament.  Kim and Zach won a whopping 15.5 imps on this hand.

If you follow the play, you will see that the West robot makes an error when he doesn't capture the diamond queen with his king.  This leads to an ignominious -990 and, adding insult to injury, the robots helplessly (or haplessly?) each contribute an ace to the last trick, won with Zach's last trump. You just can't make this stuff up!

When I first saw this hand, I thought that the hand actually belonged to North/South even though they held only 11 hcp! But then I realized that E/W can legitimately make four spades (with careful play). They can also make 3NT (or four clubs) but not if North has already bid four hearts!

Zach was the only human North to play this board.  Fifteen robot North's failed to appreciate the offensive potential of a three diamond opening.

Sunday, November 18, 2012

Suit quality in slam bidding (part 2)

This is the second part of the topic: Suit quality in slam bidding (see part 1). We will go into a little more detail about the issues raised there. Our first example illustrates the danger of taking the strong route to a contract versus the weaker route. As always we must recognize that the more (forcing) bids you take to get to a bid that you could have reached via a shorter route, the stronger your hand. When two sequences are equal in their numbers of steps, the one which more precisely describes your hand is the weaker sequence because it limits the possible hands the bidder can hold.

You pick up this hand and hear partner open 1: ♠A75 KJ5 T9852 ♣A5. You already know what contract you are going to play in, unless partner has significant extras (and they will have to be quite significant). That pre-ordained contract is 4. What's the quickest (and most limiting) way to get there in 2/1? It's to start with a forcing 1NT (essentially just a relay bid to keep the auction alive) and then jump to 4 over partner's rebid. He will now know you have game-forcing values with three hearts and no slam aspirations of your own. Isn't that exactly what you have?

What will happen if you bid 2? Nothing bad will happen if partner rebids 3♣, 2, 2♠ or 2NT. Over all of those continuations, you will now jump to 4 (fast arrival). No damage done!

But what if partner actually likes diamonds. There are several ways the auction might continue and most of them are awkward. The simplest thing is partner bids 3 and you bid 4. You will probably get away with this because it's generally understood that a raise of your minor is overruled by a delayed major suit raise. It's a jump so it can't really be a control-showing bid. Additionally, we don't bid splinters in partner's suit so it can't be a splinter for diamonds.

What if partner gets excited and bids 4♣ or 3♠ (splinters in support of diamonds)? How will you recover from that? A bid of 4 now will be ambiguous at best. Is it a control-showing-bid in search of a diamond slam? Is it an attempt to sign off in 4 (in which case how is your hand different from the one above where you went through 1NT?).

Even worse is when 4 happens to be your key-card asking bid for a diamond slam. Now, it can get really ugly. Maybe the best thing here is to simply bid 5 (at teams) in the hopes of arriving quickly in the diamond game.

This sequence actually occurred at a regional earlier this year. Kim and I were playing a very pleasant match against the eventual winners of the KOs. At our table were Pamela and Matthew Granovetter. We knew we were somewhat behind (actually we were a lot behind but we didn't realize that at the time). I opened 1 with ♠QT8 AT972 AJ76 ♣8. After Kim bid 2 with the hand shown above, I took a very optimistic view of my hand. In retrospect, I was clearly trying for a "swing." So much so that I bid 4♣. Now, Kim continued with 4. Hmm! Was that "keycard" for diamonds (we play "kickback")? Or was it a sign-off in 4? Well, if it was the latter, at least we had a double fit. And, although we had never specifically discussed the principle described in this blog, Kim's hand should be better than a minimum game force and have a good suit. Maybe slam wouldn't be such a bad prospect. In these ambiguous circumstances, our rule is generally to keep bidding if we think the bid might be forcing. It's usually more fun to play a tough slam in a good fit rather than play game in a suit where there's no fit at all. The predictable result was down 2 in 6.

So, here is the relevant principle: avoid bidding a suit in a 2/1 auction that you would be unhappy to see as the trump suit, unless you have no choice. You have some safety if the suit you really like out-ranks the one you are bidding. But, as we've seen, even that has its possible snags. Another example situation might be when you hold the following hand: ♠AKQ92 – AQJ6 ♣8642. You open 1♠ and partner responds 2, game-forcing. You have sufficient strength to bid a minor at the three-level, but do you really want to suggest clubs as trumps? Here, you do have an alternative rebid: 3. You might miss a good 6♣ slam. But you might also stay out of a bad 6♣ slam when partner's support is only ♣KJ92 or something similar – a holding that partner will consider to be magnificent support.

I have also observed that many players will bid a bad four-card minor suit rather than a forcing 1NT. At least Kim's suit had five cards in it! While 1NT typically denies the ability to bid game in a different strain, it certainly does not deny a hand which can raise to game in opener's suit. But, since it is forcing, it could also be used for those rare hands that already know where they plan to play, five-of-a-minor for example.

I'm happy to be the "goat" in the story because I obviously overbid. But it got me thinking about this topic and I think I now understand the concept behind "Picture Bids". A picture bid is the 2/1 version of the principle of fast arrival. See for example a good discussion of this topic by Eric Rodwell. Although Eric doesn't put it this way, it seems logical to me – based on everything we've discussed here – that since responder could, with a minimum balanced hand, bid 1NT followed by a jump to game in partner's suit then a 2/1 followed by a game bid shows a different type of minimum. What type? Well, since responder doesn't have much enthusiasm for game, it suggests that he doesn't have much in the way of values, certainly not controls, in the other suits. He's basically saying that I don't think we have slam unless a) you have significant extras of your own or b) we have a good double fit and you, opener, have control of the other suits.

Here's an amusing – for my partner and me – result from a recent club appreciation game. My hand was  ♠AKT8762 A3 84 ♣KT and I opened 1♠ in second position at unfavorable vulnerability. Partner bid 2♦ with ♠QJ K75 Q9763 ♣AQ3 and I jump-rebid my spades. I don't recall for sure what happened next but I think partner raised to 4♠. We ended up in 6♠. I was not concerned about needing a control in diamonds because my partner had bid this suit as a 2/1 bid and then supported my spades. This time, however, we got lucky. The DA was singleton and the slam is cold.

Admittedly, this is a tricky hand. With 14 hcp, the hand does seem a bit too good to make a forcing 1NT followed by 3NT. But if I had to come down on one side or the other, I would take the low road, starting with 1NT, simply because of the quality of the diamond suit.

Tuesday, November 13, 2012

Suit quality in slam bidding (part 1)

This is the fourth contribution of what I now realize has become a series on slam bidding. I wrote the first one three years ago: Slam Tries. Since then, there have been two more: When to bid a slam and When to bid an NT slam. And while I note that Marty Bergen, Mike Lawrence, Eddie Kantar et al have written volumes on the subject, they have not covered the stuff that I am presenting in these short articles. This topic, however, has become so long that I am going to offer it in two parts.

One of the great things about playing "Two-over-one" is that responder to an opening bid in a major is able to distinguish between those hands he expects to play in game only and other hands where he would like to suggest a slam. Of course, opener might have slam ambitions even when responder wants to stop in game. And, before I get started, let me say up front that my ideas here are controversial but, I hope, compelling.

So, let's start by looking at the types of hand an un-passed responder might hold opposite an opening bid of 1♠ where he has fewer than four cards in spades:
  1. a weak hand that is not worth any response;
  2. an intermediate-strength hand that cannot force to game:
    1. a minimum responding hand:
      1. with three spades (this hand will immediately raise to 2♠);
      2. with two spades (or perhaps a really bad flat hand with three spades);
      3. with fewer spades (this hand will have to exercise some judgment at rebid time);
    2. an invitational hand;
  3. a hand which expects to play in game but has no slam ambitions opposite a non-jump rebid:
    1. responder knows where the hand will likely play (4♠ or 3NT typically);
    2. responder does not yet know where the hand will play;
  4. a hand that will at least make game:
    1. with support for partner – and may be worth a slam;
    2. without support for partner.
Clearly, hand type 1 will pass and all hands of type 2 (except for 2.1.1) will bid 1NT forcing. All of the other hands will start with a 2/1 bid, right? Wrong. I will allow that the standard 2/1 texts do not address this issue at all – all of them, as far as I can tell, limit the 1NT forcing response to 11 or at most 12 hcp – however, there would be little point in me regurgitating the standard dogma here.

Types 4 and type 3.2 will start with a 2/1 bid. But type 3.1 should start with 1NT forcing! That's why it's forcing! We won't miss game. It doesn't deny game-forcing values. It only denies a desire to play in a slam or to find the best game. Responder will clarify his hand at his next turn. I first wrote about this principle almost two years ago: The Forcing 1NT.

Let's say that opener bids 1♠ and responder bids 1NT. So far, we've ruled out hands of types 1, 4 and 3.2. Now, let's suppose opener rebids two of a new suit. Responder's rebids will now make it clear what type of hand he has:
  • 2♠ (type 2.1.2)
  • something else (type 2.1.3)
  • 3♠ or 2NT (type 2.2)
  • 4♠ or 3NT (type 3.1)
Now, let's look at the sequences that start with, let's say, 1♠ – 2♣. Suppose opener rebids 2, 2 or 2♠, 2NT or 3♣ (the most common rebids). Here are the possible rebids by responder:
  • 2♠ (if available), 3♠ or 4♠ (various types of 4.1 – different partnerships might place different interpretations on these three possible ways of supporting opener's suit);
  • anything else (types 3.2 or 4.2).
In the first of these, responder is making a slam try! Why make a 2/1 bid and then support partner if you didn't want to at least mildly suggest slam? You could have responded 1NT and then jumped to game (or even jumped straight to game as Precision players sometimes do) with no slam interest at all.

In the latter case (anything else), we cannot be sure where responder is heading. If responder rebids his own suit, he's interested in game or slam in his suit. We don't know which yet, but if we don't support his suit next, he's likely to give up on slam unless he has a very good hand. Incidentally, this raises the issue of whether the partnership plays strong jump shifts because if they do, then certain hand types are eliminated from a normal 2/1. However, since many 2/1 players don't see the need for SJSs, I am assuming for now that we are not playing them.

But for now, we should simply assume that the first order of business is to find a playable game, in keeping with the well-accepted principle "game before slam." Thus, we will group 3.2 and 4.2 type hands as one for now.

So, what's the point of all this? The difference between hand types 3.1 and 4.1 is ambition. Game versus slam. Now, would you seriously be suggesting slam in partner's suit if you couldn't control your own suit? Ay, there's the rub. By the process of deduction then, if responder makes a 2/1 bid and next supports partner, his suit will be a good one. Opener has a right to expect at least the A or K of the suit. This follows from the obvious inference that responder can hardly expect opener to control this suit.

And further, suppose you have a 4.2 type hand (no support for opener). You will be most pleased when opener supports your suit at his next turn suggesting honor-third or perhaps four small. When this happens you're likely to find yourself as declarer at the six level. How happy would you be if your suit was Qxxxx opposite Kxx? The answer is: not ecstatic – you will have to find one player with exactly Ax.

Let's look at an example of how not to bid – courtesy of my Robot partner in a recent BBO tournament:

Don't concern yourself too much with the play. I misclicked at trick 2 (I have a new trigger-happy mouse) but the damage was done during the auction – 6♠ is unmake-able!

The robot North has a minimum game-going hand with no apparent fit. I would argue that starting with 2 is overly optimistic with such a moderate diamond suit. After my "high reverse" of 3♣ made, admittedly on moderate "extras," my robot partner chose what is, I believe, an absolute no-no in 2/1 bidding: he showed three-card support for spades with his 3♠ call. Although I knew we were missing the ♠Q, I believed us to be in a 9 card fit with all five "key" cards, so I bid the slam.

It's better in this case to start with 1NT forcing and to see what happens next. If opener rebids clubs, then responder can confidently jump to 3NT. If opener rebids spades (promising six), then 4♠ looks right. Only red suit rebids would cause problems because we will not then be in a game-forcing situation. Do we jump to 4 (likely a Moyesian fit) or 5 (perhaps with only an eight-card fit) as appropriate? I think responder should simply bid game only unless opener makes a jump rebid, in which case responder will certainly be thinking about a slam. Poorly fitting hands are never easy to bid, regardless of system. But rest assured that the partnership will not be missing a game contract if there is one.

Not every hand is suitable for a 1NT pre-game force. You must be able to handle all of opener's rebids. Otherwise, you will have to start with a 2/1.

In the next part of this topic (coming soon), I will discuss "picture bids" and fast arrival with more example hands. As always comments are welcome.

Friday, November 2, 2012

A well played match?

Bridge is never dull, at least for a student of the game like myself.  I love the curious little things that happen at the table.  That’s what this blog is really supposed to be about.

In a recent regional, for example, we had the following auction:

1 1 1 1
2 2 3 3

I don’t remember what happened after that but I think it went 4, spoiling the pattern a bit.  But it was an elegant display at that moment.

But there was one eight-board match in last weekend’s sectional Swiss that I thought so strange that I’m going to share the story.  With two rounds to go of the seven-round event, we had 52 victory points (a pleasing number for card-players).  So, with a strong finish we might end up in the overalls.  At this point, one team (our first opponents) had 94, an (almost) unassailable haul.

We were playing a less-experienced team and I thought we had a good chance of a solid win.  On the first board, our opponents bid to an unremarkable game and made it exactly.  It did seem that they might have made one or two overtricks but I dismissed it as a flat board.  Well, we won 13 imps because our teammates bid and made a slam.  Who knew?

The second board looked like we might get a 6 imp gain as our opponents stretched a little to a non-vulnerable game that went down 2.  I was pretty sure declarer could do better but I wasn’t expecting the 11 imp gain that we got when our teammates also stretched but played it a little more carefully.

The next board was a very boring little vulnerable 2+1 for us that looked like at most a couple of imps might change hands.  But no, we increased our lead by six when our teammates somehow came back with 100.  

At this point, it seemed that we might be up a bit, but not nearly enough for a significant victory point score.  I was very happy when the next board came up.  Our opponents were vulnerable while we were not.  My RHO dealt and opened 1.  My hand was  A8652  AKT3  void  QT74.  I decided to make a takeout double. This was followed by two passes.  Oh dear, partner’s going to be a bit disappointed to find that I have no trumps at all.  But never mind.  RHO opponent then decided to make an S.O.S. redouble (his shape being 4432).  This was passed out.  Interesting!  Partner had six good diamonds and our defense was accurate with the result that we took 8 tricks for the unusual score of 1000.  I knew this was a big gain, but it was hard to predict what would happen at the other table.  Apparently, they played in a more normal contract and our teammates were -110 for a gain of 13 imps.

So, half-way through the match we were up by 43 imps!  Things started to unravel a bit then, however.  On the next board, I picked up (not-vulnerable versus vulnerable) the following unremarkable hand: ♠ AQT953 ♥ QJ7 K5 ♣ 84.  I opened 1 heard partner bid 2NT (Jacoby) and I of course bid 4S.  Partner bid 5, I responded 5 and partner then bid 5.  RHO chimed in then with a double.  Hmm, obviously, RHO has the K and he’s now getting a heart lead.  If there’s another loser somewhere we’re going down in 6.  Maybe notrump would play better from partner’s side (RHO won’t be able to lead a heart then, will he?).  OTOH, maybe we’re already at our limit.  Well, surely if partner thinks we might have slam, we can make 11 tricks in notrump! [says I hopefully].

This hand was a true comedy of errors.  My partner didn’t really have his 5 bid (KJxxx Kx xxx AKQ) and we were about to be down 1 in 5.  My RHO didn’t have the K.  He had the A!  Did he really need a heart lead?  And, what was partner’s 5 bid supposed to accomplish anyway, particularly as the lead would be coming through his K?  And, perhaps silliest of all, we couldn’t take more than six tricks in notrump, whoever was playing it!  Partner passed (thankfully, nobody doubled) and then they took six diamonds and the A for -250.  Playing in the wrong strain cost us two imps, but no victory points.  Our lead was down to 31.  

We then lost an imp on a 1NT part-score.  On the penultimate board I invoked a somewhat unusual tactic: the preemptive game try.  My hand:  T986  AKQ93 6  Q97 .  With nobody vulnerable, partner dealt and passed, as did RHO.  I opened 1 followed by a pass.  Partner raised me to 2 and RHO passed again.  With a maximum of 20 (or possibly 21) high-card points between us and a singleton diamond, there seemed no likelihood of winning this declaration at 2. Indeed, the opponents might even be able to make 4 if I give them the chance. I decided to bid 2, a help suit game try.  That will deter the opponents, I thought, and partner will probably revert to 3 - the same contract that we would probably play in if they balanced.  Well, partner actually bid 4 given his excellent support for spades (Jxx).  No, seriously, he bid 4 because he had a 10-count.  All I needed was for the K to be onside to gain 8 imps.  Unfortunately, it was offside and I lost 3 imps (our side was -50 at both tables -- I don’t know what their actual contract was).  But my expectation on the hand of 2.5 imps was the kind of odds I can live with.  Lead down to 27.

Then came one of those unfortunate occurrences that you just have to take philosophically.  We bid to a normal 4 contract after a strong notrump and Stayman auction.  I don’t remember the dummy precisely but I recall that the shape was 4423.  Our trumps were pretty good ones, we were only missing the K, T and three small.  Unfortunately, my LHO held all five trumps and, try as I might, I couldn’t engineer an endplay on her because most of the adverse high cards were held by RHO.  I ended up down one while my counterpart at the other table played for some unknown reason in 3NT making easily.

We ended up with 14 imps which was good enough for 16 VPs.  Unfortunately, we lost the final match by six and finished out of the money.  As we were waiting for the last round, I was showing all of the crazy swings to my wife Kim (whose team ended up second in the event).  The director David Metcalf overheard this exchange and quipped “a well-played match.”