[pran]

dburn, on May 16 2010, 07:00 AM, said:

blackshoe, on May 15 2010, 09:46 PM, said:

awm, on May 15 2010, 10:25 PM, said:

It's also interesting that the laws say that if a player "chooses, from among logical alternatives..." and don't really say anything about what happens if a player chooses a bid that isn't a logical alternative at all.

This is something that's been bothering me for a long time. I've been informed by several pundits, including David Stevenson and Grattan Endicott, that this law does not mean what it literally says, and that if a player who has UI chooses an action that is not an LA, and that action is successful, he has violated this law, even though a literal reading would say otherwise.

The ACBL now has a regulation to the effect that any action chosen at the table is a logical alternative, because it seemed logical to the player who chose it, and that player is a representative of his class of player.

There can be no doubt that any chosen action shall be counted among the logical alternatives available, but TD and AC must still judge if this action could demonstrably have been suggested by the UI. If they rule negative to this question the result must stand however successful it is for the alleged offending side because Law 16 has not been violated.

[Cascade]

Cascade, on May 15 2010, 11:05 PM, said:

dburn, on May 15 2010, 11:44 AM, said:

he should bid a slam, because it will make considerably more than half the time.

I remain unconvinced that this is demonstrable.

With one heart (and not six spades and a weak suit)

hcp P[>12 tricks]
8 16.6%
9 21.2%
10 31.7%
11 50.1%
12 64.1%

With two hearts

hcp P[>12 tricks]
6 12.5%
7 18.7%
8 27.3%
9 39.5%
10 54.8%
11 65.4%
12 82.9%
13 88.0%

With three hearts

hcp P[>12 tricks]
10 66.5%
11 77.9%
12 87.3%

8-10 with 2 hearts P(>12) = 34.6%

7 with 2 hearts or 11 with 2 hearts or 10 with 3 hearts or 8-11 with 1 heart (and not six spades and some weak suit) P(>12) = 34.3%

The odds of 12 tricks with a slow raise based on a hand at one of the boundaries is relatively close to a normal raise with an 8-10 hand.

The relative frequencies of the various hand types are approximately

        Low      1      2      3   High    Sum
Low       0      0      0      0      0      0
   6      0      0    106      0      0    106
   7      0      0    111      0      0    111
   8      0     66    112      0      0    178
   9      0     66     91      0      0    157
  10      0     38     79     55      0    172
  11      0     33     68     44      0    145
  12      0     18     53     31      0    102
  13      0      0     29      0      0     29
High      0      0      0      0      0      0
Sum       0    221    649    130      0   1000

I still remain unconvinced that a slow raise can demonstrably suggest 6H over PASS.

[gwnn]

I don't like that simulation. It is impossible to effectively delimit 'problem hands' from non-problem hands with numbers alone (I think most problem hands are slam tries but many people thought this was false so I am not going to sustain this claim as a part of my argument). However how about this: I bid my hand exactly and if partner said 4H wtp then slam should be 0%. If partner said 4H but maybe this is not the best contract then maybe 6H is the best contract.

Meanwhile I'm still not convinced that writing pass with all caps is correct English.

[cherdanno]

pran, on May 16 2010, 04:04 AM, said:

cherdanno, on May 16 2010, 12:26 AM, said:

I don't even understand what you are disagreeing with.

Maybe I misunderstood your post?
I am not aware of any reference to Law 16 from Law 12, neither express nor implied.

I quoted a reference from Law 16 to Law 12. In any case, I really don't think we disagree about anything.

[blackshoe]

dburn, on May 16 2010, 02:00 AM, said:

The ACBL now has a regulation to the effect that any action chosen at the table is a logical alternative, because it seemed logical to the player who chose it, and that player is a representative of his class of player.

Yes, I'd forgotten that. However, "seems logical" does not mean it is logical.

[blackshoe]

Vampyr, on May 16 2010, 03:16 AM, said:

Haven't we covered this ground more than once? The call a person actually chose is deemed to count among the logical alternatives. This makes sense, because after all, this is not just a bid that some small number of the player's peers might hypothetically choose; it is the bid that the player in question (his own #1 peer?) actually made.

Does anyone want for there to be a loophole in the Laws such that, if you are in receipt of unauthorised information you are allowed to take an "illogical" action? This would become a standard tactic for those who feel they would not be permitted to "get away with" the suggested action, when they are sure that it is best and that they will get a poor score if they choose a "logical" alternative.

This has been around a long time. I'm sure I'm not the first to have pointed it out.

I agree that we don't want such a loophole, if only because I don't want, as a TD, to have to deal with players who make silly calls just because they have UI. My problem, however, is that the phrase "logical alternative" has a specific meaning, and the lawmakers have had plenty of opportunity to change the laws if what they want is a phrase that means something quite different. Instead, they have elected to redefine the words so they do not mean what they say in English. To me, that's just plain stupid.

[blackshoe]

pran, on May 16 2010, 05:13 AM, said:

There can be no doubt that any chosen action shall be counted among the logical alternatives available,

Sure, so long as you accept the premise that the lawmakers, like Humpty Dumpty, can make words mean whatever they want them to mean.

[Cascade]

gwnn, on May 16 2010, 11:07 PM, said:

I don't like that simulation. It is impossible to effectively delimit 'problem hands' from non-problem hands with numbers alone (I think most problem hands are slam tries but many people thought this was false so I am not going to sustain this claim as a part of my argument). However how about this: I bid my hand exactly and if partner said 4H wtp then slam should be 0%. If partner said 4H but maybe this is not the best contract then maybe 6H is the best contract.

Meanwhile I'm still not convinced that writing pass with all caps is correct English.

1. I agree that numbers are not an effective criteria in delimiting the problem and non problem hands. Therefore in each simulation there will be included problem hands and non problem hands. The simulations are not intended to be definitive but rather to give rough guide as to how many hands there are of different types that could potentially give a problem decision and the trick taking expectation of similar hands.

2. With hands with one heart I made some attempt to qualify the problem hands by forcing the responder to have one suit with a poor holding. This will be relatively common with weak hands.

3. It is unrealistic to expect slam (12 tricks) to be 0% when a simulation shows that over 30% of the time when partner has a routine 4H raise - 8-10 hcp or so with two hearts that 12 or more tricks are available. Perhaps if we assume partner's judgment is good then the numbers for 12 tricks when partner has a good but marginal problem hand will be lower than simulated and those when partner is bad but marginal will be slightly better. As stated above i did make some attempt to qualify the problem hands when partner had only one heart so maybe the numbers are more reasonable there.

4. If partner said 4H but maybe this is not the best contract then maybe 6H is not the best contract or maybe 6H is the best contract. This is born out by the simulation where the numbers for hands that are on the three borders that I simulated for i/ weaker than a normal raise, ii/ stronger than a normal raise (two cases with two or three trumps) and iii/ Normal raise strength but few hearts than normal are similar to the numbers for hands within the normal range. Granted (as above) not all hands so quantified will be problem hands but I doubt that that is sufficient reason to 6H is more likely.

5. While the simulations do not show the expectation for "problem hands" they do show the expectation for boundary hands on three (or four boundaries) -
i/ too strong with two trumps
ii/ too strong with three trumps
iii/ too weak with two trumps
iv/ one trump
Nevertheless I would suggest that the numbers indicate that it would be non-trivial to "demonstrate" that knowing that partner had a problem hand from one of the boundaries above that 6H was suggested over PASS or indeed that in fact it is suggested.

An argument that the problem can not be quantified is only one step (or maybe a few steps) away from saying that it 'can not be demonstrated that the UI suggests one action over another'.

The law requires an acknowledgment that a suggestion can be demonstrated. Such an acknowledgment will necessarily be more difficult when the potential problem hands are diverse and suggestive of different and even conflicting actions. This will often be the case when it is a slow action as compared with a slow PASS (non-forcing - a slow forcing PASS is likely to be similar to a slow action in that at least in a vacuum partner will not be able to tell whether the uncertainty was due to being close to double or to some bid).

[cherdanno]

We are talking about a pair that bids 1H 1S 3H 4H 6H. A hand that is too strong for 4H may well be not a borderline slam try, but a clearcut slam try where responder doesn't know how to try for slam (except bidding 4NT).

[pran]

blackshoe, on May 16 2010, 03:30 PM, said:

pran, on May 16 2010, 05:13 AM, said:

There can be no doubt that any chosen action shall be counted among the logical alternatives available,

Sure, so long as you accept the premise that the lawmakers, like Humpty Dumpty, can make words mean whatever they want them to mean.

This doesn't matter because what you apparently objects to be considered a logical alternative can then definitely not be suggested by the UI (whatever we call this alternative).

Or are you asserting that UI can demonstrably suggest an action that you consider not to be a logical alternative?

I prefer including the selected action among the logical alternatives available and in case rule that this particular alternative was not demonstrably suggested by the UI. (That was exactly what we did in the case I quoted in my OP.)

[blackshoe]

I am saying that it is illogical to, for example, blast to 6♥ when the previous bidding indicates that 6♥ is unlikely to have any reasonable play. So 6♥ is not a logical alternative. Law 16 says that a player in receipt of UI

Quote

may not choose from among logical alternatives one that could demonstrably have been suggested over another by the extraneous information.

(Emphasis mine). The choice of 6♥ here then does not violate this law. Yet we are instructed and taught to rule as if it did. That is not, IMO, a good way to administer the rules of the game.

Stephanie made a good point when she suggested that if this law were interpreted literally, some players, faced with a situation where any legal option looks to lead to a bad result, might do something completely illogical which, if successful, would be allowed to stand. I agree that we don't want that, but I'm not happy with a precedent that "the law means what it says, except when it doesn't".

I suppose we could use 73C here, but I'm not sure that will always work. In particular, is doing something completely illogical "taking advantage" of UI? Is it even not carefully avoiding taking advantage?

[nige1]

blackshoe, on May 16 2010, 03:58 PM, said:

Law 16 says that a player in receipt of UI

Quote

may not choose from among logical alternatives one that could demonstrably have been suggested over another by the extraneous information.

(Emphasis mine).

A perennial discussion topic on BLML. It was regularly suggested that the WBFLC remove the superfluous word "logical" But that would have been a "logical" simplification, probably anathema to some law-makers

[dburn]

nige1, on May 16 2010, 04:36 PM, said:

blackshoe, on May 16 2010, 03:58 PM, said:

Law 16 says that a player in receipt of UI

Quote

may not choose from among logical alternatives one that could demonstrably have been suggested over another by the extraneous information.

(Emphasis mine).

A perennial discussion topic on BLML. It was regularly suggested that the WBFLC remove the superfluous word "logical" But that would have been a "logical" simplification, probably anathema to some law-makers

It's not so much a question of removing the word "logical". What the Law probably ought to say is that the player "may not choose any action that could demonstrably have been suggested by the extraneous information, unless there is no logical alternative to the chosen action."

[nige1]

dburn, on May 16 2010, 06:11 PM, said:

It's not so much a question of removing the word "logical". What the Law probably ought to say is that the player "may not choose any action that could demonstrably have been suggested by the extraneous information, unless there is no logical alternative to the chosen action."

Yes. That is much better.

[awm]

There's something problematic about Cascade's simulation.

My intuition is that making slam on this hand will very often depend on negotiating the heart suit for no losers. Examining a set of randomly generated hands confirms that this is often the case.

Looking at one small heart opposite AKJTxx, the normal play in the heart suit for no losers is to finesse the queen. This means hoping for one of Qx, or Qxx onside. The odds of one of these holdings are about 15/64, or roughly 23.4%. There are also a fairly large number of hands where you can't make six even if the hearts behave, or where you need another finesse or something. Yet Cascade's numbers have 6♥ making substantially more often than this opposite a singleton even when responder is fairly light in high cards (yes, sometimes partner has the singleton QUEEN but this is not nearly enough to compensate).

One possibility is that double dummy, you can also pick up Qx offside which improves the odds to around 31.3%, a fairly substantial improvement.

The same situation comes up when partner has a small doubleton. The normal play is to finesse, picking up the onside queen (except possibly queen-fifth onside). This gives you a bit less than 50% odds of success. However, double-dummy you can also pick up Qx and stiff queen offside, which improves your odds to around 65.6%. Again, Cascade's numbers seem high and probably reflect this.

Perhaps the point is that the chances of making 6♥ in practice are often quite a bit worse than the double-dummy odds, in particular because you often have to negotiate the heart suit for no losers. I understand that in general double dummy and table results tend to be similar in the long run, but here it seems like there are a lot more opportunities on this particular hand for declarer to go wrong.

[peachy]

dburn, on May 16 2010, 06:11 PM, said:

nige1, on May 16 2010, 04:36 PM, said:

blackshoe, on May 16 2010, 03:58 PM, said:

Law 16 says that a player in receipt of UI

Quote

may not choose from among logical alternatives one that could demonstrably have been suggested over another by the extraneous information.

(Emphasis mine).

A perennial discussion topic on BLML. It was regularly suggested that the WBFLC remove the superfluous word "logical" But that would have been a "logical" simplification, probably anathema to some law-makers

It's not so much a question of removing the word "logical". What the Law probably ought to say is that the player "may not choose any action that could demonstrably have been suggested by the extraneous information, unless there is no logical alternative to the chosen action."

I like this. Plugs the hole without any ripple effects.

[Cascade]

awm, on May 17 2010, 12:18 PM, said:

There's something problematic about Cascade's simulation.

My intuition is that making slam on this hand will very often depend on negotiating the heart suit for no losers. Examining a set of randomly generated hands confirms that this is often the case.

Looking at one small heart opposite AKJTxx, the normal play in the heart suit for no losers is to finesse the queen. This means hoping for one of Qx, or Qxx onside. The odds of one of these holdings are about 15/64, or roughly 23.4%. There are also a fairly large number of hands where you can't make six even if the hearts behave, or where you need another finesse or something. Yet Cascade's numbers have 6♥ making substantially more often than this opposite a singleton even when responder is fairly light in high cards (yes, sometimes partner has the singleton QUEEN but this is not nearly enough to compensate).

One possibility is that double dummy, you can also pick up Qx offside which improves the odds to around 31.3%, a fairly substantial improvement.

The same situation comes up when partner has a small doubleton. The normal play is to finesse, picking up the onside queen (except possibly queen-fifth onside). This gives you a bit less than 50% odds of success. However, double-dummy you can also pick up Qx and stiff queen offside, which improves your odds to around 65.6%. Again, Cascade's numbers seem high and probably reflect this.

Perhaps the point is that the chances of making 6♥ in practice are often quite a bit worse than the double-dummy odds, in particular because you often have to negotiate the heart suit for no losers. I understand that in general double dummy and table results tend to be similar in the long run, but here it seems like there are a lot more opportunities on this particular hand for declarer to go wrong.

That doesn't surprise me. Certainly Qx offside is a big gain for double dummy play over single dummy play.

The chance of a stiff queen is interesting.

By simulation over the range 6-12 hcp the chance of the stiff being the queen varies considerably:

6 9.1%
7 11.7%
8 13.6%
9 16.3%
10 20.1%
11 24.7%
12 27.7%

Given we have the ♥8 there is one other case where we can play the suit for no losers. That is partner has the stiff 9 and either opponent has a singleton queen. Admittedly this is very unlikely at about 0.3%.

[dburn]

nige1, on May 16 2010, 07:16 PM, said:

dburn, on May 16 2010, 06:11 PM, said:

It's not so much a question of removing the word "logical". What the Law probably ought to say is that the player "may not choose any action that could demonstrably have been suggested by the extraneous information, unless there is no logical alternative to the chosen action."

Yes. That is much better.

Well, maybe. It certainly does what the ACBL wants, which is to prevent people from fluking good scores by choosing an illogical alternative that happens to work. It also makes clear that you are allowed to do what the AI suggests even when it is also what the UI suggests (this is obvious anyway, but the wording of the current Law renders it less obvious than it ought to be). Perhaps it also resolves the conflict between Law 73 and Law 16, and it may even cure cancer and neutralize greenhouse gases.

As an afterthought, for next to nothing you could remove the words "that could demonstrably have been" without detracting from the sense.

But there is probably something wrong with it that I haven't spotted. Well it was said by the bard:

A Romanian rhymer I met
Used a system he based on roulette.
His reliance on chance
Was a def'nite advance
And yet... and yet... and yet...

[pran]

dburn, on May 17 2010, 03:01 AM, said:

As an afterthought, for next to nothing you could remove the words "that could demonstrably have been" without detracting from the sense.

The word "demonstrably" implies that there must be some substance in the alleged suggestion of the chosen action.

The word "could" reduces the requirement for a proof to showing a likely possibility.

These two words are both essential.

[dburn]

pran, on May 17 2010, 02:45 AM, said:

dburn, on May 17 2010, 03:01 AM, said:

As an afterthought, for next to nothing you could remove the words "that could demonstrably have been" without detracting from the sense.

The word "demonstrably" implies that there must be some substance in the alleged suggestion of the chosen action.

The word "could" reduces the requirement for a proof to showing a likely possibility.

These two words are both essential.

Neither of the words "could" and "demonstrably" is even useful, let alone essential. We begin with the proposition "X is suggested". We strengthen this to "X is demonstrably suggested", and then weaken it again to "X could demonstrably be suggested". In so doing we accomplish precisely nothing, except to ensure that we follow standard legal practice in not using one word when three will do.