[BudH]

Regarding the wording below of Law 16B1(b) regarding the definition of logical alternative:

A logical alternative action is one that, among the class of players in question and using the methods of the partnership, would be given serious consideration by a "significant proportion" of such players, of whom it is judged "some" might select it.

1. What percentage of players is needed to be a "significant proportion"? (let us call this percentage X)

2. Of that "significant proportion", what percentage of those players is needed to satisfy the phrase "some might select it"? (let us call this percentage Y)

In the last few years, for the typical case where there are very limited choices, I have lowered my "X" from near 50% to about 33% and Y to be about 15 to 20%, which would result out of 100 similar players about 5 to 7 players minimum to be a logical alternative.

I have seen some EBU literature lately which has X at about 20% and Y a very small number which would result in only 2 or 3 out of 100 being required, which seems far too low, in my opinion.

How do other major bridge organizations view the terms "significant proportion" and "some [of the significant proportion sub-group] might select it"?

[blackshoe]

I'm not sure that the drive to assign specific numbers to these concepts is worthwhile or desirable. That said, let's see:

From my Oxford American Dictionary:

[

Quote

b]Proportion:[/b] a part, share, or number considered in comparative relation to a whole.
Significant: 1. sufficiently great or important to be worthy of attention. 2. Statistics of, relating to, or having significance.
Significance: 1. the quality of being worthy of attention; importance. 2. (also statistical significance)the extent to which a result deviates from that expected to arise simply from random variation or errors in sampling.

So "significant proportion" means "a part of the whole sufficiently important to be worthy of attention". It seems pretty clear that 5% is too small a part. 10%? Maybe that's enough. 20%? Now we're getting into it, I think. At some point (33%? 40%? 50%?) you're too high, you miss out on some opinions that would be "worthy of attention". I think 50% is clearly too high in that respect.

Of "some" my dictionary says "an unspecified number or amount of people or things; at least a small amount or number of people or things". It seems clear that 2 or 3% is not too small.

Considering the statistical definitions of "significant" and "significance", keep in mind that our usual sample size is, statistically speaking, way too small to have much, er, significance. :-)

I don't think I've seen any guidance from ACBL on specific numbers for these concepts.

[BudH]

If you found 100 similar players and at least 1/4 to 1/3 of them SERIOUSLY considered a choice, then, in my opinion, at least "some" of them will always choose that particular choice, especially if "some" is defined to be a very low percentage.

Because it will be very rare for the "some" criteria to not be met if the "seriously considered" is met, I think it would be easier to define a LA as simply one that "some" would actually choose, where "some" is something between 10% and 20% and leave the "seriously considered out of the definition.

Or, alternately define an LA as one SERIOUSLY considered by approximately 1/4 to 1/3 of similar players and leave out the "some" part. Emphasis on "seriously".

[mycroft]

We have actively moved away from fixed percentages, because they lead to players not "carefully avoiding using the UI", but "carefully working out whether they're going to pass the percentages if they use the UI".

The ACBL has case law that "zero is not 'some number'", if that helps...

[Vampyr]

mycroft, on 2015-March-12, 11:34, said:

We have actively moved away from fixed percentages, because they lead to players not "carefully avoiding using the UI", but "carefully working out whether they're going to pass the percentages if they use the UI".

This is poor, because this is exactly what players should do -- they cannot choose among logical alternatives if they do not know how a LA is defined.

[campboy]

"Some" is there solely to rule out the rare situation where virtually everyone will realise that a particular action is wrong, but some people might require serious thought to do so. It should be very unusual to rule that an action meets the first criterion to be an LA but not the second.

I agree that "significant proportion" needs a proper definition. However, failing to define it in the laws effectively just delegates the responsibility to RAs. The EBU, for example, interpretes "a significant proportion" as "at least 20%" (White Book 8.16.6.1).

[mycroft]

They need not choose from among logical alternatives. They should carefully avoid using the unauthorized information. If they think "this is clearly suggested, but I think I can get away with it because most won't take the other choice", they're not carefully avoiding anything.

If they say "this is suggested, but seriously, that call?" then they're probably good.

The EBU used to have a 70% rule. I have quotes of people who tried to work out if 75% or 65% of people would take the action they clearly wanted to take, but was also clearly suggested by the UI. I don't want to go back there.

I agree that we also don't want to go back to "If it hesitates, shoot it" - which is why "zero is not 'some number'". And certainly there will be guidelines.

[Vampyr]

mycroft, on 2015-March-12, 17:16, said:

They need not choose from among logical alternatives. They should carefully avoid using the unauthorized information. If they think "this is clearly suggested, but I think I can get away with it because most won't take the other choice", they're not carefully avoiding anything.

If they say "this is suggested, but seriously, that call?" then they're probably good.

Buzzt. You have obviously tried, but you have got it wrong. There cannot be one criterion that players must use, and another by which they will be judged by the qdirector/appeals committee.

[blackshoe]

Vampyr, on 2015-March-12, 21:57, said:

Buzzt. You have obviously tried, but you have got it wrong. There cannot be one criterion that players must use, and another by which they will be judged by the qdirector/appeals committee.

Cannot? But there is - Mycroft is talking about Law 73C vs. Law 16B. I don't remember where I heard it first, possibly from David Stevenson, but the former has been described as a law for players, the latter for directors.

[mycroft]

There have been several attempts to determine if it is possible to violate L73C without violating L16B (I think yes, obviously; one is a matter of intent, the other is a matter of action); whether it is possible to violate L16B without violating L73C (I think yes, again; this is the source of all the "I realize you tried your best to ... but here's what the poll showed" comments - the ones that aren't either white lies or "listen, you know what I'm telling you is BS, and you know I know that, but let's pretend you tried to do right, and you are going to do right in future, correct?"); and whether we can and should penalize people for either infraction separately.

But I know for a fact that if you are trying to see if you can get away with doing the suggested action (as opposed to trying to argue that the non-suggested actions just don't make sense), you are being imProper; and being deliberately imProper is an infraction. And the 70% rule caused people to do just that - and they didn't think it was a problem admitting it publicly in front of TDs.

Something's wrong with that - and the thing that's wrong with that is the thing that meant that in 197mumble the Proprieties moved from being an Appendix to the Laws to Laws themselves.

[StevenG]

mycroft, on 2015-March-13, 11:17, said:

Something's wrong with that

So, you try to work out which bids are legal under both L16B and L73C and then pick the best option. And then "something's wrong with that"? I don't see why - it's not as if the laws say that you have to throw the board when in receipt of UI.

[mycroft]

I think if you work out what calls are legal under L73C, you'll be good - apart from a few times when I had my blind spot pointed out to me, I have never had an issue with any call that I truly thought was L73C valid (those I did have a blind spot to *I* called the TD once my eyes were opened and accepted the ruling with grace). I know for a fact that there are those who deliberately subvert L73C, hoping that L16 will let them get away with it; I know for a fact that there are those who only take the call not suggested by the UI after determining that they probably wouldn't survive a L16 challenge. That is *not* carefully avoiding using the UI.

You don't have to throw the board. You have to carefully avoid using the UI. If the thought of making any other call is either ludicrous or makes you go green, then fine. If it's "I know this is wrong this time, but if I wasn't told it was wrong, I might have done it", then off you go to a poor score. It's still a better score than you'll get after the TD both assigns a contract and "the best defence at all possible" against it, most likely.

[campboy]

mycroft, on 2015-March-16, 15:36, said:

It's still a better score than you'll get after the TD both assigns a contract and "the best defence at all possible" against it, most likely.

But this is exactly why a player might want to go further than 73C requires, to avoid an adjusted score. If I know I would never have taken the non-suggested action (typically because I had already decided what I was going to do before receiving the UI), but expect that the TD will think it is an LA, then I am better off changing my plan and taking it. This is not uncommon in my experience.

[Vampyr]

mycroft, on 2015-March-16, 15:36, said:

I know for a fact that there are those who only take the call not suggested by the UI after determining that they probably wouldn't survive a L16 challenge. That is *not* carefully avoiding using the UI.

I don't agree. And I think that the phrase "carefully avoid using the UI" is pretty vague.

[BudH]

From the Houston NABC 2009 ACBL Law Commission Minutes:

"There was a consensus that in Law 16B1 where the word 'some' is used it should suggest to the tournament director and a committee more than one, and the word 'significant' should suggest to the tournament director and a committee more than a minor proportion (e.g. 2/100) but less than a major proportion (e.g. 40/100)."


So it seems between 2% and 40% is roughly what the ACBL Laws Commission is deeming a "significant proportion" and anything more than one is enough to satisfy "some" of that significant proportion sub-group.

That seems in line with the 20% to 30% (at least that many) that I typically consider and in line with the 20% figure that the EBU has for a guideline for "significant proportion".