[awm]

Where, if anywhere, is the flaw in this reasoning? It seemed fairly clearcut to me, but ends up contradicting the laws in effect for virtually every SO...

Players are allowed to use judgement in deciding what to bid. It's often the case that hands which seem on the surface to be appropriate for a particular call end up not being bid that way. Some systems allow more than one call for a hand based on location of values. Some people have implicit high card location or distributional requirements. Common examples of these judgement calls in a fairly natural system include:

(1) Choosing to open 1nt or 1M with 5332 shape.
(2) Choosing which suit to open with 4-4 in two suits (especially in a four-card major method).
(3) Deciding whether to preempt based on factors like suit quality, side suit honors, length in unbid major(s), side voids, etc.

Not only is judgement allowed, but it's perfectly accepted for this judgement to depend on issues like form of scoring, state of the match, perceived skill level of the opponents, etc.

Bridge is a game of full disclosure. Both explicit and implicit agreements should be disclosed when requested. Among other things, this implies that consistent "judgement calls" do need to be disclosed. If partner always opens 1nt with 5332 in range, and opponents ask about his 1M opener, I should disclose that "he won't be 5332 with a certain point range." If partner never opens a weak two with a side void I should disclose this. The rule seems to be that "anything I know about partner's judgement or tendencies, opponents should also be allowed to find out if they ask." Obviously the more partner and I have played together, the more disclosure responsibilities we have for implicit things like this.

A reasonable regulatory structure should encourage full disclosure. It should not be the case that disclosing patterns in partner's judgement somehow incurs a penalty against the disclosing side. It would therefore be dumb to have a rule that "you're allowed to use your judgement whether to open 1M or 1nt with 5332 hands in range, but an agreement to open 1nt with all 5332 hands in range is illegal." In this case it could be the case that partner's "judgement" is to always open 1nt, and you know this, and he knows this, but as soon as you disclose this to opponents you have an illegal agreement. In this case something very very bad has occurred. If it's legal to "use judgement" to bid a certain way, you should also be allowed to formally agree to bid that way, and to disclose it.

If it's legal to agree that a certain opening shows any one of a particular set of hands, it should also be legal to agree that the same opening shows any one of a strict subset of that particular set of hands. Otherwise, we are back to a situation where the "formal agreement" could be the legal one, but the players are allowed to use their judgement. Their judgement may be quite consistent in not opening hands outside the strict subset. They could even have another alternate opening bid with an overlapping definition that handles the other hands. Such a situation encourages non-disclosure rather than really changing the set of allowed methods. This is not an acceptable situation.

-------------------

Unfortunately, the last of these statements in bold contradicts the set of legal agreements most everywhere. The case in question is a convention called Wilcosz, by which an opening of 2♦ shows a weak hand with two five-card suits, at least one of which must be a major. One can see that this is a strict subset of the "weak hands with a five card major" -- a convention known as Multi which is frequently allowed. The last bold statement above would imply that if Multi is allowed on a five-card major, then Wilcosz should also be allowed which is not generally the case.

Of course, the ACBL regulations violate the statements above in any number of additional ways.

[jtfanclub]

awm, on Jun 29 2007, 07:03 PM, said:

If it's legal to agree that a certain opening shows any one of a particular set of hands, it should also be legal to agree that the same opening shows any one of a strict subset of that particular set of hands.

Nope, sorry.

As an obvious example, it's legal to use a nebulous club opening. That means the club opening can be as few as 1 card.

That does not mean you can define 1♣ as 1+ clubs and 5+ spades.

Disclosure is nice, and all, but you're missing that big old forest out there. Partnership agreements are defined in part by response structure. If my responses to 1 club assume a 5 card spade suit by partner, or his rebids assume such, then it's clear there's a partnership agreement. 2 diamonds can be weak and natural. It cannot be weak with an unknown second suit, even though that's a subset of weak and natural. How can we tell if it's really weak and second suit? Look at the response structure.

But, it was a nice argument up until that point.

[hrothgar]

Hi Adam

A few years back, I made a similar argument on the Bridge laws mailing lists. I argued that basic concepts from set theory should be applied to convention regulation. More specifically, I stated that if it is legal to make opening X with a given set of hands, then it should also be legal to make the same opening with any subset of that hand.

jtfanclub has raised an obvious criticism of this theory. However, I think that it is possible to reconcile the two approaches:

jtfanclub assumes that his (1♣ and 5+ Spades) example invalidates the "set theory" based approach. I would argue that the set theory approach is perfectly valid so long as you exercise sufficient care in defining the initial set of hands that can be shown by a given bid. If you combine the set theory with any opening that can show any hand type imaginable you're going to run into problems. More specifically, you probably don't want to license a completely nebulous 1♣ opening.

[jtfanclub]

hrothgar, on Jun 29 2007, 07:39 PM, said:

jtfanclub has raised an obvious criticism of this theory. However, I think that it is possible to reconcile the two approaches:

jtfanclub assumes that his (1♣ and 5+ Spades) example invalidates the "set theory" based approach. I would argue that the set theory approach is perfectly valid so long as you exercise sufficient care in defining the initial set of hands that can be shown by a given bid.

Gross.

Bad enough to try to define SAYC 1 club openers to opponents....you know, one club shows my longer minor except when it doesn't, denies a 5 card major except when I have one, shows 12-21 hcp except when I have 9 or 24....the list goes on and on.

And now I'm going to try to codify that in the rules? No thanks.

An opening bid or initial response cannot promise length in a suit other than the one bid, except under certain listed exceptions. An opening bid or initial response cannot promise length in an unknown suit unless it also promises a strong hand (15+), and even then only in certain circumstances.

It's simple, it's easy to understand and follow (even if some people try to circumvent it). I even agree with it for GCC. I wish all the rules in the GCC were so easily defined.

[awm]

It seems that the current approach on the ACBL general chart allows 1♣ showing 1+♣ and 5+♠, since any 1♣ opening which guarantees 10+ points is allowed. The set theoretic approach will never allow me to agree to open 1♣ with a hand that I couldn't agree to open 1♣ before. It simply says that "if I am allowed to open 1♣ on hand type X, by agreement, then what I choose to open on hand type Y has no impact on the legality of opening 1♣ on hand type X." In any case, which would you rather defend, a 1♣ opening that shows "1+♣, any shape that we don't feel like opening something else" or "1+♣ and 5+♠"? Which is better disclosed?

Let me give some examples of ridiculous ACBL rules based on not using a set theoretic approach:

(1) A response of 1NT to 1M showing "invitational or better values" is not allowed. However, a 1NT response showing "invitational or better values or an unspecified 8-card suit" does not promise invitational or better values and is thus okay. How does this fit the spirit of that rule?

(2) An opening 2M showing a weak hand with 5 cards exactly in the major is allowed. Further, I am allowed to "use judgement" and never preempt with 5332 patterns. Therefore it's legal to open 2M showing exactly 5 in the major and a 4+ card side suit, as long as I don't disclose this four-card side suit requirement. How is this fair?

(3) Suppose partner and I agree to play four-card majors and four-card 1♦ openings. In addition, say we always open 1M or 1♦ if we have four cards in one of those suits. It follows that a 1NT opening promises 4+♣ in addition to being balanced, and denies any other four-card suit. This makes our 1NT opening conventional (promises length in a particular suit) and quite possibly disallowed. But what if we don't make this a formal requirement, and just say "well, we open 1NT with balanced hands when we feel like it" and "when we feel like it" happens to only be when we can't open 1♦ or 1M. Again, non-disclosure helps us out. How is this fair?

You can try to make arguments about follow-up methods, but this is a very difficult situation to resolve. Say I play relays in response to 2M in (2), and there is a followup for opener that shows 5332, but this followup is ridiculously high and would probably get us a terrible result if it ever came up. I could argue that "well there is a bid that shows 5332, so it's just judgement that we always pass the 5332s" or even "well we'd like to bid it on 5332 but usually we pass because the relay structure will screw us." In any case, a detailed exploration of all possible followup sequences is usually impossible for an opposing player (or a director) to peruse, so in practice "it depends on follow-up methods" is going to equate to "it's allowed."

[hrothgar]

For what its worth, I agree with Adam's logic.

Designing a regulatory structure that encourages incomplete disclosure is a disaster waiting to happen.

[kenrexford]

I completely agree as well.

There are a number of occasions that this type of logic and problem comes up, such as where any conventional bid tailors the meaning of what might otherwise be a normal-sounding bid. And then, we are told not to alert any inferences from other calls. Makes no sense. I refuse to follow that logic.

An example is rather clear.

I only open 1♦ with:

(a.) 4441 with four diamonds, or
(b.) five diamonds plus a stiff or void somewhere, or
(c.) 6+ diamonds

So, I play a "short club."

If I were to simply note "short club," I'd be failing to mention that I'd open 1♣ with more than 4432. I'd also open 1♣ with:

(a.) 4342/3442
(b.) 3352/3253/2353
(c.) 3343/4243/2443
(d.) 2245/2254

That's much different. Now, tweak 1♦ further when I play that 2♦ shows 13-16 and minors. Now, 1♦ is further modified, such that any hand in that range will feature either six diamonds or 4M/5♦. 1♦...2♣ is now very weak.

Tweak further, adding a weak 2NT for the minors. Now, 1♦...2♣ shows 17+.

These are subsets of hands, meanings that may or may not be "regulated" in some sense, but meanings that are forced by allowing other bids to be made. That nuance, or at least some nuances, are not even supposed to be alerted or explained, which makes no sense.

So, I just alert and explain them anyway. If some bone-head wants to complain because I gave him more help than I was supposed to give him, I'll take my punishment as handed out.

[jtfanclub]

hrothgar, on Jun 29 2007, 09:23 PM, said:

For what its worth, I agree with Adam's logic.

Shocking. Don't you believe no systems should be banned, or am I thinking of somebody else?

Look, if the ACBL makes a rule, and somebody does something cute in order to obfuscate that they're violating the rule, then they're cheating, pure and simple. You treat them like you treat other cheaters. The fact that people can, and will, cheat doesn't mean that we shouldn't have any rules at all.

And if you don't understand why I say no rules at all, by all means publish your set theory rules. It's a lot longer than you'd think. Or maybe shorter, if your real intent is to allow everything.

The rule is very simple. The fact that you can think of ways to try to get around the rule shows how clear it is. Leaving out the listed exceptions, an opening or initial response may only show length in the suit bid. It may not show, or imply, length in a known suit, or length in an unknown suit. If you make an opening a two suiter, whether by pretending to include a few other hands you never actually bid, or through negative inferences, or through other methods, you're cheating. I think it's wonderful that you found lots of ways to cheat where you think people won't get caught. It shows, well, you. Whether that's good or bad, I don't know.

Look, bridge will always have this problem. Those meanies won't allow tranfer openers. So you just psyche a heart bid when you mean spades. And now your opponents don't even get to know that you were making a transfer opener! Wow, that's horrible. I guess we'll just have to allow transfer openers, then.

Do I think the rules are too restrictive? Probably. Are they far too vague and capricious? Absolutely. Is it a real problem that different directors interpret in different ways? Definitely. Is it a crime that directors will allow you to wink-wink nudge-nudge play things that aren't permitted as long as you don't draw attention to it? Oh yeah. And do I have a serious problem with certain experts getting their pets installed while things that meet the same criteria are excluded? Don't get me started.

But is the problem that the rules restrict definitions of hands without necessarily restricting the hand? Um, no. Because, as you're quickly figure out if you try to put SAYC to set theory, it's easy to make a nightmare system simply by using the subsets of SAYC. You have to exclude subsets, if you're going to make a workable set of rules.

And I am out of here. See you in a week.

[awm]

My intent is not to "allow everything." In fact I don't particularly care what is and isn't allowed! What I would like to see is:

1. Clear rules, where it's easy to tell what's allowed and what isn't.
2. Fair rules, where what's allowed doesn't depend on who the players are.
3. Unbendable rules, where the agreement itself, and not the manner in which it is explained, determines whether it's a legal agreement.
4. Enforceable rules, where the legality of a bid doesn't depend examining and making a complex judgement on the sequence of followups.

The problem is that the current set of rules doesn't seem to conform to these standards. There are many gray areas where no one really knows what the rule is, and in practice it often depends on the players. There are many ways to bend the rules, to explain your agreement via negative inferences instead of positive rules and thereby make it legal. And I see numerous suggestions that "well we should look at the followups" which will never happen in practice. I'm just asking for a level playing field. I don't appreciate being accused of cheating.

[rbforster]

hrothgar, on Jun 29 2007, 09:23 PM, said:

Designing a regulatory structure that encourages incomplete disclosure is a disaster waiting to happen.

Depending on your definition of a disaster, I'd say that kind of flawed regulatory structure already exists here in the US. Whether or not it's commonly abused is, for the reasons given above, hard to tell. In any case, it would be a welcome area for improvements.

[Free]

See WBF color defined systems: it's clear, you can make all kinds of categories, and if you want you can easily ADD some conventions if it's not clear enough. Problems there are, for example, Polish club, which is defined RED, but in Poland it's a standard system (which should make it GREEN). This means we need better definitions of "standard" (more like "natural"), but that's about it...

In Belgium we use these color codes with success imo. Yellow systems and BSC are usually banned except at high level tournaments and competition. Red systems are allowed most of the time, which makes sure most useful artificial methods and conventions are allowed. However, at low level competition only green systems, the basic natural systems, are allowed. Maybe blue as well (strong ♣ or ♦), which isn't that different from pure natural anyway. There's also a list with common conventions that are allowed (like puppet stayman, muiderberg, stuff like that).

Nobody complains, except poor bridge players playing at a level which is too high for them.

[han]

I agree that it would be good if the laws could be implemented that follow the guidelines Adam suggests. I don't expect that this will happen anytime soon.

Quote

(2) An opening 2M showing a weak hand with 5 cards exactly in the major is allowed. Further, I am allowed to "use judgement" and never preempt with 5332 patterns. Therefore it's legal to open 2M showing exactly 5 in the major and a 4+ card side suit, as long as I don't disclose this four-card side suit requirement. How is this fair?

This is a great case for your viewpoint. If you also agree not to open 2M with 4 cards in the other major (as many LOL do) then you'd be playing Muiderberg.

By the way, I think this thread should be in a different subforum, I think it is interesting.

[han]

jtfanclub, on Jun 29 2007, 10:56 PM, said:

It's easy to make a nightmare system simply by using the subsets of SAYC.

Perhaps you can show us?

For example, try making a nightmare 1NT opening. Of course you are only allowed to use hands that have 15-17 HCP and are balanced. For example, you could play that 1NT shows some 4432 pattern, with a 4333 you'd have to open 1m and with a 5332 you'd have to open the 5-card suit. But would that really be a problem to play against?

A nightmare 1H opening that shows 5+ hearts, an opening hand and at most 1 card in the other major? That's a wet dream to play against, not a nightmare. Besides, you can't open anything else with the other hands so it is not even possible.

How about a pass that shows less than an opening hand but denies a 6+ suit and 5-10 points? You know what, I'd treat it as a normal pass.

A 2C opening showing a gameforcing unbalanced hand?

I don't think it is possible to come up with a subsystem that would scare anybody.

[hrothgar]

jtfanclub, on Jun 30 2007, 06:56 AM, said:

hrothgar, on Jun 29 2007, 09:23 PM, said:

For what its worth, I agree with Adam's logic.

Shocking. Don't you believe no systems should be banned, or am I thinking of somebody else?

I believe LOTS of things...

For the purpose of this discussion, I think that its useful to distinguish between two different issues:

1. The rules set that governs how convention regulations should be written
2. What conventions should be legal

I have some strong opinions about both.

In an ideal world, there won't an any conflict between the two. On a more practical basis, I think that that I am more concerned with issue 1 than and am about issue 2. Simply put, a poorly written regulatory structure has a much more significant impact on the game than any particular set of convention regulations.

[helene_t]

Hannie, on Jun 30 2007, 01:39 PM, said:

I agree that it would be good if the laws could be implemented that follow the guidelines Adam suggests. I don't expect that this will happen anytime soon.

Quote

(2) An opening 2M showing a weak hand with 5 cards exactly in the major is allowed. Further, I am allowed to "use judgement" and never preempt with 5332 patterns. Therefore it's legal to open 2M showing exactly 5 in the major and a 4+ card side suit, as long as I don't disclose this four-card side suit requirement. How is this fair?

This is a great case for your viewpoint. If you also agree not to open 2M with 4 cards in the other major (as many LOL do) then you'd be playing Muiderberg.

By the way, I think this thread should be in a different subforum, I think it is interesting.

Yes, or agree to open such hands 2♦ and you're playing the Scollaard/Jacobs 2♦ opening which may or may not be a BSC depending on the mood of the TD (so far all events they have played in together, AFAIK, allowed BSCs but I wonder what happens the first time they open 2♦ in a non-BSC event).

Yet I'm more on line with jtfanclub. Law is far from an exact science. A friend of mine wrote his M.Sc. thesis about an effort to reconstruct implicit laws by applying machine learning to court verdict data. To make the problem as easy as possible he focused on corporate tax law, which one might think is one of the more "exact" branches of law (and Denmark follows the French philosophy of law so the problem should be easier than it would be in the U.S. or Britain, one might expect). Yet he found (or maybe it was his assumption) that the implicit laws must necessarily be based on fuzzy logic or some such. And the predictions were far from perfect. Maybe this could be explained by judges' brain farts or by incomplete or inaccurate data, but it could also be that the problem is too complex for the ML tool.

Maybe it's reasonable to expect bridge law to be more exact than real law, but in principle the problem is the same.

It is debatable how much freedom TDs should have to interpret the spirit of the law instead of applying bureaucratic criteria, and it is debatable to what extent one can rely on players' sportsmanship rather than making the laws easy to enforce. When different parties have different answers to these two questions, tensions will arise.

I think this is a useful thread since the system is clearly prone to improvement. But we should not expect an easy solution. Making the laws a little bit clearer step by step, or at least making the different parties more conscious about the muddiness of the laws, is what we can aim at.

[DrTodd13]

The system is broken. That has been known for as long as I've played bridge. Do you think they care that their rules are inconsistent? So long as the rules appeal to the average bridge player that is all they care about.

[awm]

DrTodd13, on Jun 30 2007, 11:40 AM, said:

The system is broken. That has been known for as long as I've played bridge. Do you think they care that their rules are inconsistent? So long as the rules appeal to the average bridge player that is all they care about.

In ACBL-land, the vast majority of players do not know or understand these rules. Neither do directors, and I've had some long and frustrating encounters with directors over whether certain methods are legal.

I don't think the rules "appeal" to the average bridge player, so much as that the average bridge player doesn't understand them and doesn't feel like it matters.

Then again, there are sometimes complaints at the club when people are playing a strange (but legal) system. And there are definitely complaints by people who adopt some new convention they read about somewhere or made up on their own, only to find out later that it is illegal. I think there are indications that a set of rules which were clear and consistent would make almost everyone happier (although to varying degrees, it probably isn't a big deal to the average player one way or another).

The only people who really seem to benefit from the current situation is the people on the rules committee, because they can interpret the charts to mean whatever they want and the directors basically have to just go along with what they say (since no one else really understands what the rules are).

[awm]

There is one caveat that probably should be added: methods need to be explicable. While I don't think the following convention is hard to defend, per se, I can see being very annoyed to have to play against it:

"1NT shows 15-17 high balanced. In addition, it guarantees that the number of aces is at least the number of jacks, that the total number of red cards is either 5, 7, or 9, that the square of the number of jacks minus the number of queens is not divisible by three, that the number of ZAR points is even, and that opener would prefer that partner not lead a spade."

I'm sure I'd just treat it as a natural strong notrump, but having to listen to all this nonsense could be bothersome.

[TimG]

awm, on Jun 29 2007, 07:03 PM, said:

It would therefore be dumb to have a rule that "you're allowed to use your judgement whether to open 1M or 1nt with 5332 hands in range, but an agreement to open 1nt with all 5332 hands in range is illegal." In this case it could be the case that partner's "judgement" is to always open 1nt, and you know this, and he knows this, but as soon as you disclose this to opponents you have an illegal agreement.

I think you have the agreement before you disclose. It's the disclosure that reveals the agreement, but that doesn't mean the agreement doesn't exist without the disclosure.

I think there are two separate types of judgment. One is the judgment you use in creating your agreements, such as deciding that 5332 hands should be opened 1NT. The other judgment is deciding what to do with the bidding structure which you have chosen, such as deciding which 5332 hands to open 1NT once you agree that it is right to open some such hands 1NT.

I understand that there can be a fine line (which I may not be able to define) between the two. The players in an established partnership will probably know each other's tendencies (use of judgment) better than the opponents, the extent to which this understanding must be disclosed is unclear to me. That is, how much of the understanding is based upon partnership experience and how much is based upon bridge knowledge is often a fuzzy line.

As an example, I seem to recall an article in which Michael Rosenberg outlined something like 11 factors in his decision whether to open 1D or 1C when 44 in the minors. If given an example hand that is 44 in the minors, Zia could probably guess better than most which Rosenberg would open, but how much of that is agreement and how much of it is understanding Rosenberg's bridge judgment is unclear. To me, agreement should always be disclosed, while understanding of bridge judgment might not be disclosed.

[blackshoe]

Zia's understanding of Rosenberg's bridge judgement comes from partnership experience - and that is disclosable.