[helene_t]

 aguahombre, on 2013-October-30, 09:33, said:

Perhaps, at this club, the correct thing to do would have been give the peers two polls:

1N-3N &

1N- (P) 3N (X).

This would be time-consuming, however. Sjoerd would have to wait about 15 minutes for them to forget the first poll before doing the second one.

Sounds like a reasonable approach but still not bullet-proof. The slow pass could be interpreted as a good hand not suitable for a double (whatever a double would mean for that partnership).

[ggwhiz]

It sure seems like if a heart was led (small or top) and beat the contract the police would be summoned anyway for the previously stated reasons that a spade lead is a legitimate choice and by the UI that partner is not close to broke indicating that a heart may well work. Especially the King if pard has length but realized they aren't good enough quality to double or we control the suit in time to switch to whatever the tank was about.

A lead of either minor suit would look like taking advantage of the UI imo but the auction, not the UI suggests a major and I see nothing outrageous or indicated with the choice that was made.

[StevenG]

 lamford, on 2013-October-30, 10:17, said:

I agree that the BIT suggests a club over a spade, but it also suggests a spade over a heart.

But it changes the probability of my leading a heart from 0% to 0%. If I use the UI, it changes the probability of my leading a spade from 100% (I think) to 0%. How does that constitute suggesting a spade over a heart?

[campboy]

 StevenG, on 2013-October-30, 10:52, said:

But it changes the probability of my leading a heart from 0% to 0%. If I use the UI, it changes the probability of my leading a spade from 100% (I think) to 0%. How does that constitute suggesting a spade over a heart?

If the chance of leading a heart without UI was really 0% then it wasn't an LA so there is no issue. And that may be so for you. But suppose for the sake of argument that I am on lead; I believe a heart is an LA for me.

With UI -- well, if allowed to use that UI everyone would lead a club, of course. Therefore a club lead will be disallowed (if successful). So if I were unethical I might reason "I have the UI that partner wants me to lead his suit. The most likely lead to work is a club, but I can't lead a club because the director will rule it back. So I'd better try a diamond." Thus -- even though no-one would lead a diamond if allowed to use the UI -- it is suggested over a heart or a spade, because once you rule out a club it is the obvious lead. You can presumably see where this argument is going.

(In case it isn't clear, I am not in any way suggesting that the player at the table did anything unethical.)

[StevenG]

If we take the poll as gospel ( ) then we start with the probability of a spade as 50% and a top heart as 25%. After the UI, the probability of a heart is, presumably, 0%. But what if the probability of a spade is, say, 20%? The probability of a spade has still dropped by more than that of a heart. What is our methodology on this?

The difficulty is that a heart, while not a LA for me, is a LA according to Law 16B. But the perception of its importance seems to be skewed by whether you consider it a serious LA, or something that rather scrapes into the list. With multiple LAs, if a minor LA (call it X) becomes less attractive after UI, do you automatically adjust for a non-X choice? Or is it only if it's a LA you perceive (rightly or wrongly) as one of the most likely?

[campboy]

 StevenG, on 2013-October-30, 12:17, said:

If we take the poll as gospel ( ) then we start with the probability of a spade as 50% and a top heart as 25%. After the UI, the probability of a heart is, presumably, 0%. But what if the probability of a spade is, say, 20%? The probability of a spade has still dropped by more than that of a heart. What is our methodology on this?

The probability of a spade has gone down less, in the sense that it is has been reduced by 60% whereas that of a heart has been reduced by 100%. But all that is irrelevant anyway, as I keep saying. What matters is not the actual probabilities of a heart or a spade lead, but what they would be if you were forced to lead one or the other. And I think the chance of choosing a spade over a heart would go up from maybe 75% to 100%.

[Trinidad]

 campboy, on 2013-October-30, 11:57, said:

If the chance of leading a heart without UI was really 0% then it wasn't an LA so there is no issue. And that may be so for you. But suppose for the sake of argument that I am on lead; I believe a heart is an LA for me.

With UI -- well, if allowed to use that UI everyone would lead a club, of course. Therefore a club lead will be disallowed (if successful). So if I were unethical I might reason "I have the UI that partner wants me to lead his suit. The most likely lead to work is a club, but I can't lead a club because the director will rule it back. So I'd better try a diamond." Thus -- even though no-one would lead a diamond if allowed to use the UI -- it is suggested over a heart or a spade, because once you rule out a club it is the obvious lead. You can presumably see where this argument is going.

(In case it isn't clear, I am not in any way suggesting that the player at the table did anything unethical.)

This argument is going nowhere.. The probability that anyone would like to lead a spade with the UI is equal to the probability that anyone would like to lead a heart with the UI: zero.

As a matter of fact, I asked my wife. Without the UI, she picked a spade lead. I asked whether there were any LA's and she said: "not really, but I guess that someone just out of the bridge course might lead 'fourth best from longest and strongest'". Whatever..

But then I gave her the UI and asked what lead it suggested and we had the following, very fast exchange:
"A club."
- "What suit next?"
"A diamond."
- "What suit next?"
"A heart."
- "A heart?!? Why a heart and not a spade?"
"Partner has a long, broken suit and one or two entries. It's obviously a minor, so it doesn't really matter, but you wanted to know the difference between the major suits. So suppose -for argument's sake- that it's a major: If it is hearts, you will get the contract down immediately. If it is spades, declarer will hold the stopper(s), since dummy will likely show up with a good minor to justify his raise to 3NT on a combined 23-24 points or so."
- "All nice, but how did you get the idea that partner has a long, broken suit? Shouldn't he have a solid suit?"
"Of course not. He can't have a solid suit. If he would have had a solid suit, he would not have thought about doubling... He would have doubled... "

Rik

[campboy]

Ok, we agree at least that the order starts club, diamond. Obviously you disallow a club. Do you disallow a diamond? After all, just like hearts and spades, the probability of leading a diamond if allowed to use the UI is zero.

[blackshoe]

Could a club demonstrably be suggested over some other LA? Yes. Can't lead a club. Could a diamond demonstrably be suggested over some other LA? Yes. Can't lead a diamond. Could a heart demonstrably be suggested over some other LA? Yes. Can't lead a heart. Could a spade demonstrably be suggested over some other LA? Apparently not. So if a non-spade is led, we adjust to whatever would have happened with a spade lead. The fact that's the lead that would have been made without the UI is irrelevant. The fact that a heart is "only a little bit suggested" is irrelevant. Now you should probably look at the spades and ask "which cards are LAs, and which is suggested over which other one(s)?" In this particular case, I don't think that will be fruitful, but in some others perhaps it might.

This reminds me of a more general (and hypothetical) thread we had some years ago: A is suggested over B, B is suggested over C, C is suggested over A. Now what? I don't remember that we came to a satisfactory conclusion. The point being that in our actual case here if a spade could demonstrably be suggested over something else, we're stuck. The law does not provide for a construction such that "A is more suggested over B than is B over C" and so on. So I suppose the choices are to throw the board out for this table (score it as "not played") or treat it as "there are no LAs". Neither is palatable, so I hope I'm right that a spade is not suggested over another suit.

Added: the probability of a particular LA is IMO a red herring. The only question is whether it demonstrably could be suggested over another LA.

This post has been edited by blackshoe: 2013-October-30, 19:47
Reason for edit: Added a line.

[Trinidad]

I think it would be good to use a concept of "noise".

At some point, actions are so improbable, given the UI, as in a spade or a heart lead here that the difference between the two is less than the "noise".

One can say that a spade lead has a 10^-7 probability for succes and a heart lead has a zero probability of succes, if one assumes that the BIT is based on a solid suit. Following this assumption, one could argue that a spade lead is suggested over a heart lead.

However, we cannot be sure that the assumption is true. There is always some degree of uncertainty. This uncertainty creates "noise". In this case the uncertainty that our assumption is true (reasonable as it is) is much larger than the difference between the effect of a heart or a spade lead. That means that we cannot discern between the spade and heart lead.

In this case, I would think that the probability that the BIT is based on a broken suit is much higher than the probability that a spade lead will be succesful if the BIT is based on a solid suit. P(BIT because of broken suit) >> P(Spade lead is succesful | Partner has a solid suit)

Therefore, the error introduced by our assumption is much larger than the difference between the heart and spade lead, making this difference not measurable (below the detection limit of our system).

Rik

[pran]

 blackshoe, on 2013-October-30, 19:40, said:

Could a club demonstrably be suggested over some other LA? Yes. Can't lead a club. Could a diamond demonstrably be suggested over some other LA? Yes. Can't lead a diamond. Could a heart demonstrably be suggested over some other LA? Yes. Can't lead a heart. Could a spade demonstrably be suggested over some other LA? Apparently not. So if a non-spade is led, we adjust to whatever would have happened with a spade lead. The fact that's the lead that would have been made without the UI is irrelevant. The fact that a heart is "only a little bit suggested" is irrelevant. Now you should probably look at the spades and ask "which cards are LAs, and which is suggested over which other one(s)?" In this particular case, I don't think that will be fruitful, but in some others perhaps it might.

This reminds me of a more general (and hypothetical) thread we had some years ago: A is suggested over B, B is suggested over C, C is suggested over A. Now what? I don't remember that we came to a satisfactory conclusion. The point being that in our actual case here if a spade could demonstrably be suggested over something else, we're stuck. The law does not provide for a construction such that "A is more suggested over B than is B over C" and so on. So I suppose the choices are to throw the board out for this table (score it as "not played") or treat it as "there are no LAs". Neither is palatable, so I hope I'm right that a spade is not suggested over another suit.

Added: the probability of a particular LA is IMO a red herring. The only question is whether it demonstrably could be suggested over another LA.

Seems to me that you apply false logic:

"A is (demonstrably) suggested over B, B is (demonstrably) suggested over C and C is (demonstrably) suggested over A" is equivalent to the Logic-Mathematic statement : (A > B) AND (B > C) AND (C > A) which is a self contradiction and therefore impossible.

Law 16 implies that when a player having received UI has more than one LA available then TD must be able to rank these according to how they may be suggested by the UI in order to consider one or more (but not all) illegal.

[axman]

 pran, on 2013-October-31, 01:52, said:

Seems to me that you apply false logic:

"A is (demonstrably) suggested over B, B is (demonstrably) suggested over C and C is (demonstrably) suggested over A" is equivalent to the Logic-Mathematic statement : (A > B) AND (B > C) AND (C > A) which is a self contradiction and therefore impossible.

Law 16 implies that when a player having received UI has more than one LA available then TD must be able to rank these according to how they may be suggested by the UI in order to consider one or more (but not all) illegal.

It is the law that is false logic. It requires the player to do the impossible: Prove a Negative [and to boot- without supplying the standard against which he wlll be judged.]

[Trinidad]

One thing that is demonstrated very nicely by this thread is how TDs can get biased by seeing the whole deal.

Suppose that South had running clubs and North, with the same hand and same auction, would have led a club. We would have all condemned North for leading a club (and rightfully so). But no one would have said that North would be forced to lead a heart. Everybody would condemn the club lead, perhaps somebody would remark that diamonds would not be allowed either, but no one would have thought anything about the majors.

But now that we all see that South has a running spade suit and that we know how devastating the spade lead turned out to be, it is easy to condemn the spade lead. Just as easy as it is to forget that South didn't need to have that hand, or -in fact- was extremely likely not going to have that kind of hand and that North may have had UI, but that he didn't see the South hand either.

I strongly suspect that there would have been people crying foul if North would have led a heart and would have found South with:
♠xxx
♥AJTxxxx
♦x
♣Ax

They would have found a way to rationalize why the UI suggested a heart lead over the spade lead (e.g. by reasoning like my wife).

(In fact, I secretly hope that Sjoerd will post: "Fooled you all. In reality North did lead the ♥K and found South with 7 to the AJ. Everybody said that spades was an LA and that the UI suggested hearts over spades." Come on, Sjoerd, I am waiting. )

Rik

[gnasher]

It's almost impossible that partner was considering doubling for a heart lead when we have KQ9x. With 10xxx of spades, it's certainly possible that partner was considering a double, but rejected it because he didn't have ♠10.

[blackshoe]

 pran, on 2013-October-31, 01:52, said:

Seems to me that you apply false logic:

"A is (demonstrably) suggested over B, B is (demonstrably) suggested over C and C is (demonstrably) suggested over A" is equivalent to the Logic-Mathematic statement : (A > B) AND (B > C) AND (C > A) which is a self contradiction and therefore impossible.

Law 16 implies that when a player having received UI has more than one LA available then TD must be able to rank these according to how they may be suggested by the UI in order to consider one or more (but not all) illegal.

Hm. I don't recall anyone bringing up that argument in the previous discussion, but perhaps I've simply forgotten the details. IAC, what we're talking about is not "A is greater than B" or even "if A then B", so I'm not sure your statement is a valid description of the situation. It's more (UI > A) AND (UI > B) AND (UI > C).

The law doesn't imply that. We can infer only that the law overlooked a possibility, or that the hope was that it would never come up.

[campboy]

 blackshoe, on 2013-October-31, 08:33, said:

Hm. I don't recall anyone bringing up that argument in the previous discussion, but perhaps I've simply forgotten the details. IAC, what we're talking about is not "A is greater than B" or even "if A then B", so I'm not sure your statement is a valid description of the situation. It's more (UI > A) AND (UI > B) AND (UI > C).

It's not that, because each alternative has to be suggested over the next. So UI has to increase the amount you prefer A to B, and the amount you prefer B to C, and the amount you prefer C to A. Assuming your preference for A over B can be described as (amount you like A) - (amount you like B), this situation is impossible.

[lamford]

The poll (and common sense) tells us that a heart is an LA. We do not need to decide if a spade is demonstrably suggested over a heart. All 16B1a requires is that it could demonstrably have been suggested over a heart. If South was thinking of doubling to say "find my suit", then a non-heart lead could demonstrably have been suggested.

[pran]

 campboy, on 2013-October-31, 08:54, said:

It's not that, because each alternative has to be suggested over the next. So UI has to increase the amount you prefer A to B, and the amount you prefer B to C, and the amount you prefer C to A. Assuming your preference for A over B can be described as (amount you like A) - (amount you like B), this situation is impossible.

At least one who understood logic.

The statement "alternative A could be suggested over alternative B" establishes a ranking between these two alternatives, this ranking can logically be expressed as (A > B).

A statement like "A could be suggested over B" and (at the same time) "B could be suggested over A" is self contradicting and at the best illustrates a TD failure to properly analyze the situation. This does not illustrate any flaw in Law 16, in fact there is IMHO no such flaw.

[StevenG]

I thought Campboy dismissed this approach (or, at least, this metric) in post 46.

[campboy]

 StevenG, on 2013-October-31, 11:33, said:

I thought Campboy dismissed this approach (or, at least, this metric) in post 46.

No, I didn't. Here I am talking about how much you like the various alternatives, not the probability of actually selecting them. And when there are two alternatives you would never select you can still like one much more than the other.