Tuesday, 17 April 2007

Using 1S:2D as a Transfer

This follows on from the explanation of the artificial 2C response to major-suit openings. If playing natural 2/1 responses (apart from 2C), 1S : 2D and 1S : 2H would both promise five-card suits. This works fairly well, but there is a lot to gain from switching the 2D and 2H responses round, so that 1S : 2D shows hearts.

Clearly this will improve the bidding of hands with hearts at the expense of hands with diamonds. But because of the importance of major suits, this seems to be worthwhile.

The most important gain is on hands not worth forcing to game.

S K5
D 52
C 8742

Suppose that you respond to 1S with 1NT on this hand, and partner rebids 2D. This gives you a difficult rebid problem. You could try 2S or 2NT, but neither of these show your lovely heart suit. Alternatively you could rebid 2H, but this doesn't express the strength of the hand very well. This is the sort of hand that the 2D transfer was designed for. You respond 2D on this hand, and if opener completes the transfer by bidding 2H, you continue with 2S. The 2S rebid is non-forcing and mildly invitational. Not only does this sequence get the heart suit into the picture, it also perfectly describes the strength.

The 2D bid should promise a decent hand. The minimum strength is about the same as a traditional Acol 2H response:

S 8
H AKJ963
D 9863
C 84

This hand is just about worth a 2D response: this time we will pass opener's 2H rebid. However, when holding only five hearts and shortness in spades, it is better to respond 1NT unless the hand has genuinely invitational values (that is, it should be good enough for a 2NT rebid).

Of course, game-forcing hands with 5+ hearts must respond 2D as well. After opener's minimum rebids, they have to bid a minor suit in order to create a forcing auction, like in Acol.

Opener is not obliged to complete the transfer, but, apart from raising hearts, the only alternative with a minimum hand is to bid 2S. This shows 6+ spades, and will tend to be short in hearts (but if the spade suit is sufficiently good opener might have heart tolerance). Opener's minimum rebids are not forcing: responder would pass with the minimum 8-point hand above. With a stronger hand, opener bids as in Acol, except that I would play 2NT as showing a good hand with 6+ spades (forcing to game).

As was said above, the problem with playing 2D as a transfer is that you have to work out what to do when responder actually has diamonds. Balanced hands with diamonds are not a worry, since they can be put into the artificial 2C response. But unbalanced hands with diamonds have to respond 2H to 1S. The loss of a step can make life difficult - especially from opener's point of view, when he holds hearts. Some further artificiality is likely to be needed to try to deal with these problems.

An alternative approach is to put all the hands with diamonds into the 2C response. This frees up 1S : 2H to be used as a spade raise. (And similarly, 1H : 2D could be used as a heart raise.) I rather like this idea, but it means that the 2C response becomes incredibly complicated, though there does seem to be just enough room to make it work. I wrote out some notes for a complete system of artificial 2/1 responses here.

Saturday, 14 April 2007

2NT Showing a 6-card Suit

This is another way to try and improve bidding after a 5-card major opening, but unlike the artificial 2C response this convention is very simple. It involves opener's 2NT rebid after a 2-level response, for example 1H : 2D , 2NT. Particularly if you play a strong NT opening and often open 1NT on balanced hands with a 5-card major, the natural meaning for 2NT does not get used very much. Minimum balanced hands should be happy rebidding the major instead.

A more useful meaning for the 2NT rebid is showing a hand with 6+ cards in the suit opened and better-than-minimum values. These hands are extremely difficult to bid if you do not have such a convention available. In Acol, you would have to jump to the 3-level. This is problematic because it takes up so much space: there is a well-known problem with trying to decide which of responder's rebids should be cue-bids and which should be natural, and the reason this is so difficult is because there is not enough space to have both things. In 2/1, the jump to the 3-level promises an excellent suit, and hands without such a good suit have to rebid the major at the 2-level. It then becomes virtually impossible to show the extra strength of the hand.

The conventional 2NT rebid solves all these problems. It describes the important features of the hand immediately, while leaving plenty of space for further exploration. The best thing is that the continuations are very simple: all of responder's rebids at the 3-level are natural. For example, after 1H : 2D , 2NT if responder wants to set hearts as trumps and start cue-bidding he simply bids 3H. Whereas, if he wants to show a two-suiter or rebid his diamonds, he can do those things too. A jump to 4C would be a splinter.

The main drawback to this convention, apart from having to find another bid with a natural no-trump hand, is the possibility of wrong-siding a no-trump contact. However, because the bid promises a 6-card major, there is a high probability that the hand will be played in the suit instead. I feel that the advantages of being able to show the hand more than make up for this occasional problem.

This really is one of my favourite conventions. It can be added to most 5-card major systems (though it is not so good with a weak NT) and requires hardly any partnership discussion apart from the basic definition.

Friday, 13 April 2007

1M:2C Artificial

If you play 2/1 Game Force (or "GF except rebid"), you need to decide what you will do on a hand worth forcing to game over a 1-of-a-major opening but without a good suit. For example, over a 1S opening:

S K5
H AQ52
D A84
C J862

The standard method is to bid 2C, establishing the game force. However, if you do this then it can become difficult for opener to know whether you have a "real" club suit or not, which makes slam bidding harder. In fact, it may sometimes be necessary to make a two-over-one response in a three-card suit (change one of the clubs in the example above to a spade).

One attempt at solving this problem is to play a natural game-forcing 2NT response. Removing balanced hands like the one above means that 2-level suit responses virtually guarantee a 5-card suit. However, there are three problems with this idea. First of all, the 2NT response consumes a lot of space, making further exploration on those hands difficult. Secondly, the range of game-forcing hands which do not have a 5-card suit is very wide: they can be any strength from 12 HCP up, and include 4-4-4-1 distributions as well as balanced hands. It is not really practical to put these all into 2NT, and so the 2NT response does not completely solve the problem it was designed for. And thirdly, it means that you can't use the 2NT response for other purposes (it is more commonly played as showing a good raise of opener's suit, of course).

A much better idea is to put all game-forcing balanced hands into the 2C response. So, 2C becomes an artificial two-way bid: it shows either real clubs or a game-forcing balanced hand. In some situations the hand might contain only two clubs.

Obviously this clears up the other two-over-one responses, since they now promise a 5-card suit. You might think that there is still a problem with the 2C bid, since we are putting even more hands into it than in standard 2/1, but in fact by making it explicitly a two-way bid, it becomes much easier to distinguish in the subsequent auction when responder actually has genuine clubs.

To see how this works, let's assume for the moment that we will play natural continuations over the 2C response. Now, assuming that opener rebids at the 2-level, we will require that responder always rebids 2NT with the balanced hand. Any other bid will show an unbalanced hand, and thereby implies a decent club suit. So for example, with the hand from earlier,

S K5
H AQ52
D A84
C J862

we will respond 2C to 1S, intending to bid 2NT at our next turn, even if opener's rebid is 2H. In standard methods it would be more normal to raise a 2H rebid to three, but playing a two-way 2C response the direct raise should show the "natural" hand type and promise good clubs. Of course, there is no reason why you couldn't bid this way even if your 2C response was defined as natural, but turning it into a two-way bid makes everything seem much clearer, as well as improving the definition of the 2D response.

Notice how by rebidding 2NT we put ourselves in the same position as those people who play a natural game-forcing 2NT response, except that opener has had one extra chance to describe his hand. We have gained an entire round of bidding, which is extremely useful.

So far this has all been very simple. However, it is possible to add a vitually unlimited amount of artificiality to the continuations after 2C. There are a number of reasons why it helps to move away from natural bidding.

First of all, consider a natural auction such as 1S : 2C , 2H : 2NT. What do opener's bids mean now? If you play natural methods then 3H, 3S and 3NT have obvious meanings, but 3C and 3D are less clear - presumably these would show a fragment, but considering that these are such cheap bids the meaning is not terribly useful. Furthermore, neither partner's strength is well defined - it would be better to give some information about strength at this point. So it seems better to use artificial continuations, for example 3C showing a minimum hand with 3D as a further artificial asking bid after that. This sort of scheme can be used throughout the system. Once you start doing things like this, the system starts to look rather like a relay system. Indeed, the 2C response is sometimes called a "relay".

Next, you might decide that it is not always most efficient to use 2NT as the bid which shows a balanced hand. Particularly after a 2D response to 2C, it makes a lot of sense to bid 2H with most balanced hands and use 2NT for something else. Doing this makes it look even more like a relay system.

And finally, it is possible to do a lot better than natural responses to 2C. In particular, using opener's 2D rebid to show diamonds is not particularly efficient: you want to use 2D much more frequently than that, particularly when responder's 2H rebid shows the balanced hand. Also, opener's bids at the 3-level need to be better defined. A typical scheme might look something like this (played by Bocchi / Duboin):

2D = any minimum without 4 cards in the other major
2H = any hand with 4 cards in the other major
2S and above = better than minimum, without 4 cards in the other major

Again, this looks very much like a relay system, with the minimum and better-than-minimum hands being treated symmetrically (1M : 2C , 2D : 2H , 2S/2NT/3C show the same shapes as 1M : 2C , 2S/2NT/3C).

Putting all these things together, the resulting scheme can be extremely complicated. But this is an area of system which rewards a bit of hard work. Balanced game-forcing hands are quite frequent, and are very difficult to bid in standard systems. A relay-like scheme is ideal for them: relays work best when the asking hand is balanced and fairly strong.

Tuesday, 3 April 2007

Announcements: More! More! More!

If my blog posts so far have been pretty esoteric, I suppose this is the one where I completely lose touch with reality. Yep, this is about how we should have more announcements. Never mind the fact that the ones we have already are so controversial that it would be suicidal for the L&E to try and impose any more - this one is straight out of the "If I Ruled the World" folder.

But I don't want you to think that I believe every situation can be dealt with using announcements. Far from it. The consequences of having too many announcements are much worse than those from having too few. We got along well enough without any announcements before last August. Announcements are only of use in a very limited set of circumstances. Let's start by listing three conditions that are absolutely necessary for announcements to work:

  • The rule must be easy to understand and remember.
  • Announcements must only apply in situations where players can be expected to know exactly what their agreements are.
  • The announcements must actually have a purpose - they must be justified by one or both of the reasons I gave in the previous post. (That is, avoiding alerts for common artificial bids, or avoiding unauthorised information from asking questions.)

These are absolutely non-negotiable. To give an example, it was once suggested that all take-out doubles should be announced as "take-out" and all penalty doubles should be announced as "penalties". This scores well in terms of being easy to understand, and does solve disclosure problems, but it still would be an awful regulation because of the second condition: in complicated auctions, no-one has firm agreements about the meanings of doubles (though they might have general agreements which are relevant), so they can't be expected to make an announcement.

In addition to the above points, there are two other things that would be helpful:

  • Ideally, new announcements should only affect pairs playing unusual methods.
  • Announcements work best when the agreement can be described easily and effectively in a short phrase rather than with a long explanation.

However, these are not as essential as the three main conditions above, and indeed these two things tend to work in opposite directions: unusual methods will generally be more difficult to describe.

Given these conditions, there are only a very restricted number of situations where announcements might usefully apply. Two in particular stand out, and I would like to talk about those.

1. "Negative" Doubles

This forms part of my wider ideas on the alerting of doubles. In contrast to the proposal mentioned above where all penalty and take-out doubles would be announced, which I think is a terrible idea, there is one particular situation where announcing doubles might well be useful. This is when there has been an opening bid, followed by a natural suit overcall from opener's LHO (below 3NT) and a double from opener's partner. Ignore for the moment the importance of making the rules in this situation consistent with the rules for doubles in general. What would be the best rule in an auction such as 1S : (3D) : Dbl ? My suggestion is this:

  • If the double is for penalties, then do not alert or announce;
  • If the double is for take-out, then announce as "negative";
  • If the double has some other meaning, then alert.

This would solve two problems that we currently have in this situation: firstly that some players who play double as penalties don't realise they have to alert, and secondly that when a double is not alerted, some opponents feel the need to check that it really is a take-out double as it is supposed to be.

But perhaps the best thing about this rule is that it also works when the opening bid was at the 2-level, or was in NT. In the auction 1NT : (2H) : Dbl , for example, the traditional and most popular meaning is penalties: so again the current rule means we have problems with people not alerting penalty doubles, but here it's worse because opponents might get the wrong idea even if it is disclosed correctly. And in the auction 2S : (3C) : Dbl , I've seen people very angry to find out that they were supposed to have alerted a penalty double. The EBU's current rule (an unalerted double is take-out) does well in minimizing alerts when the opening bid was a natural 1-of-a-suit, but people find it terribly confusing in the other situations.

Of course, my suggested rule wouldn't fit in with the rules we have at the moment about alerting for doubles in other situations. It is designed to be part of a system where, with the one exception for negative doubles, only very unusual doubles are alertable. In fact my specific suggestion was that apart from the negative double situation, no doubles would be alertable except:

  • Doubles of natural opening bids. (Below 3NT, a double of a natural suit opening is expected to be for take-out; a double of a natural no-trump opening is expected to be for penalties);
  • Certain very specific unusual doubles. (Perhaps just anti-lead-directing doubles.)

So really while the rules for the specific negative double auction may appear slightly more complicated than the current ones, the scheme as a whole is much simpler.

[Incidentally, I could do with some help in lobbying about this, so if you think this is a good idea, or even if you just agree with the basic principle that neither take-out nor penalty doubles should be alertable in auctions more complicated than the "negative double" auction, please grab your nearest L&E member or other influential person and tell them.]

2. Opening bids which are currently alertable

This is an obvious candidate for announcements because opponents will very often want to know the meaning of such a bid. As with a natural 1NT opening, the problem is opponents asking questions. For example, after a 1C opening it is easy for an opponent to ask a question in such a way that he reveals that he has clubs himself. While we could say that opponents should be able to find out the meaning by looking at the convention card, experience suggests that even when a convention card is available, opponents still like to ask directly.

In theory, all opening bids which are currently alerted could be announced instead. That would certainly be a simple enough rule to understand. But it might be wise to look at narrower categories of opening bids.

  • Two-level suit opening bids. Currently all two-level opening bids are either alerted or announced. If the alertable bids were announced instead, then that would make the rule for such bids very simple: they would all be annouced. This makes a lot of sense, particularly when the opening bid is 2 of a major, since the introduction of announcements means that an alerted 2H or 2S bid is now quite surprising, so opponents will very often ask anyway.
  • "Short" 1C and 1D opening bids. An announcement for these would be useful in order to distinguish them from the "genuinely" artificial minor-suit openings which are currently alerted. The same goes for a natural opening bid which is alertable because of some unexpected agreement (e.g. 1D "natural but could have longer clubs"). This would also have useful consequences for the alerting of doubles under the current rules - a double of a natural or short minor is expected to be take-out whereas a double of an artificial bid is expected to show the suit bid.

So there are particularly good arguments for those opening bids to be announced. Note that this wouldn't make the rules much more complicated, since "announce natural 2-of-a-suit bids" becomes "announce all 2-of-a-suit bids" and the "short" 1C/1D is an unusual method which most players would not need to worry about. Having said that, I think there is still a good case for all alertable opening bids to become announcements instead.

But there is one difficulty which I've avoided talking about so far. That is, there is such a wide variety of possible meanings for these bids, it would be impossible to define "official" wordings for announcements for them all. I used to think that this made the idea of announcements for these bids a non-starter, but now I don't believe it would be a serious problem at all. The solution is to allow players to word their annoucements in whatever way they think best, but to have a list of recommended announcements for the more common methods. (These recommended announcements would include the short ones we already have for natural two-level suit bids.) Add to this a little bit of common sense - for example if a relatively long explanation is given the first time a bid comes up, a shorter form would be acceptable if it comes up again against the same opponents - and I think it would be fine.

The 3NT Opening

It's difficult to find a good meaning for the 3NT opening. The "natural" strong balanced meaning doesn't work very well because there's too little space to bid constructively over it. More often 3NT gets used as showing a pre-empt in one of the minors, either "Gambling" (promising a solid suit) or showing a normal 4-minor pre-empt so that the 4C and 4D opening bids can be used for other purposes. However, this wrong-sides the possible 3NT contract, and, particularly when 3NT does not promise a solid suit, it can be difficult for partner to know whether to play in 3NT anyway. Also, using 3NT for any sort of pre-empt gives the opponents more options than they would have if you had made a natural opening bid.

I prefer to use 3NT to show a hand which is not interested in playing in no-trumps at all. One possibility which a few pairs have adopted is to reverse the usual "Namyats" opening bids so that instead of 3NT showing a minor-suit pre-empt and 4C/4D showing good major-suit pre-empts, 4C and 4D become natural and 3NT shows a good pre-empt in either major suit. This seems like a reasonable thing to do. But what I would like to recommend is the simpler approach where 3NT shows specifically a good pre-empt in hearts.

You make life a lot easier for yourself by showing a specific suit. As I said here, I'm generally wary of agreeing to play multi-meaning pre-empts because there is so much which needs to be discussed. Also, slam tries are more effective when the suit has already been shown because you don't need to waste time confirming which suit opener has.

Of course, this gives up the chance to show a good 4S pre-empt. But there are good reasons why being able to show hearts is more important than being able to show spades. When you have hearts, you are worried about the opponents competing in spades. You want to make a pre-emptive bid to keep them out of their possible good spade fit - or to put them into a bad spade fit. And also if you open at a lower level you might get pre-empted yourself. For example, if you open 1H, the opponents may be able to bid up to 4S and leave you unable to show your one-suiter without going to the 5-level. When you have spades you are less worried about keeping the opponents out since you can outbid them on any level. Note also that a 3NT opening prevents the opponents from bidding spades below the 4-level, the same as a 4H opening would. If you have spades, then an opening bid which shows a good 4S pre-empt will allow the opponents to get in a heart bid at the 4-level, so this is significantly less pre-emptive than a 4S opening would be.

Monday, 2 April 2007

Announcements in the EBU

Eight months on from the big changes to our alerting regulations, announcements are still a very controversial topic. It's difficult to tell exactly what the level of opposition is, because inevitably it's the people with the strongest opinions whose voices are heard, but it's certainly clear that a significant number of people don't like announcements. One interesting feature of the debate is that there is an apparent distinction between "club players" and "tournament players", with a common perception amongst the club players being that the EBU's rules are there to cater for the tournament players, at the expense of the "ordinary" club players. While the EBU might disagree, it is difficult to shake off this criticim, since it is genuinely difficult to cater for everyone at once, and there is no doubt that announcements work best in more serious events. But in spite of all the criticism, I still believe that announcements are a good thing. I'd like to explain why I think it is that despite announcements being good for the game, there are so many people who are unconvinced.

First of all, we need to understand what the point of announcements is. I feel that the EBU did not do a good job of explaining this, which was a big mistake. The most common short explanation was, "It gives the opponents information immediately, without them having to ask for it." But that's nothing more than a restatement of what an announcement is, it doesn't explain why we should have them. In fact there are two distinct reasons why we have announcements:

1. Avoiding alerts for very common artificial bids.

This is the reason why we announce Stayman and transfers. The "basic" rules for alerting in the EBU say that artificial bids should be alerted. If this rule applied here (as it did before 1st August 2006) then Stayman and transfers would be alertable. However, this would mean that when the auction went 1NT : 2C / 2D / 2H, this would nearly always be alerted. This in itself would not be a problem: the problem comes when a pair turns up who don't play normal Stayman or transfers, but play some more unusual convention instead. The danger is that their opponents will see the alert and assume it was Stayman or a transfer since that is nearly always the reason for the alert. I used to play "Keri" over 1NT where the 2C response forced partner to bid 2D: many opponents went wrong because they had assumed it was Stayman, and occasionally some of them got very upset about it, even though we had done all we were required to do. The announcement of Stayman and transfers solves this problem, because now if such a bid is alerted opponents will know that it is something unusual. (Or of course it could be a Stayman bidder who has simply forgotten the regulations, but this will be clarified easily enough.)

2. Avoiding unauthorised information from asking questions.

This is the reason why we announce the range of 1NT openings. The point is that opponents will very often want to know what the range is. But asking about the range can lead to problems - particularly if opener's left-hand opponent asks and then passes. Asking tends to suggest that the player was interested in bidding, and now the Laws say that the player's partner must "carefully avoid taking advantage" of that information. Sometimes this is not a problem, but on other occasions a director may have to be called to sort things out. Or perhaps even a director should be called, but the players don't like calling the director. Either way, it would be so much better if the unauthorised information had not been passed in the first place: then this problem would not arise. This is why we announce 1NT openings: it gives the opponents the information they need, without them having to ask a question which might give their partner ethical problems.

For the other type of announcement that we have - natural two-level suit openings - both of the above reasons apply. (Before the announcement was brought in, an alerted 2H or 2S opening was nearly always a natural "weak two". And, like with a 1NT opening, opponents will very often need to know the meaning of a 2-level suit opening.)

As I said, I do not think that the EBU did a good job of explaining these things. So it is no wonder that so many people thought the announcements were pointless. The way the announcements were presented made them seem very arbitrary - just rules for rules' sake. Even if you ignore the attempts at explaining why we should have announcements and just look at the explanations of the rules themselves, there was a lot that could have been improved. The completely unreadable "easy guide" that the EBU produced is one of the most hopeless attempts at an explaining something that I have ever seen.

But of course it can't all be blamed on the publicity. Even if the L&E had handled that as well as possible, announcements were always going to be controversial. The fundamental problem is that the two reasons for having announcements aren't really things you'd expect a club player to be interested in. Let's look at those two reasons again:

1. Avoiding alerts for very common artificial bids. The announcement is designed to make opponents aware when a pair is playing an unusal method (ie. one which is alertable but is not Stayman or a transfer). However, these unusual methods are so rare at club level that club players will hardly ever come across them. So there is little need for them to be protected by annoucements.

2. Avoiding unauthorised information from asking questions. Most club players aren't interested in subtle Laws issues like this. Some people won't understand what the problems are in the first place, while others that do have some idea of the Laws are happy to ignore any problems that come up, since they don't like calling the director anyway. In a "friendly" game where people don't mind little bits of unauthorised information occurring, there seems little need to use announcements to prevent it.

Notice that the underlying problems - unsual systems and ethical issues - are not just specific to announcements, but cause a variety of difficulties in trying to apply the Laws of bridge to non-serious events. People who enjoy themselves playing in a club where everyone plays the same system and the director is hardly ever called about unauthorised information are not going to appreciate rules which are designed for unusual systems and scrupulous adherence to the Laws.

But to a "serious" bridge player, this looks so wrong. Announcements do genuinely solve the problems with unusual systems and unauthorised information that I have mentioned. And they come at little cost since once you've got used to them an announcement is no more difficult than an alert. But how do we sell this to our club players - do we tell them they're wrong not to care about these things? I think that is pretty much what we're saying. So while we may believe that announements improve the game for every type of player, it's no wonder that so many club players think that the rules are not designed with them in mind.