[gnasher]

From another thread:

blackshoe, on 2012-May-21, 21:46, said:

Grattan Endicott once suggested to me that the intended meaning of the phrase "logical alternative" is something more like "plausible alternative for the class of player concerned". That was before the 2007 laws were drafted, so for the life of me I don't know why they didn't change the wording in the law.

Vampyr, on 2012-May-21, 22:32, said:

"Plausible" is a good word, because it seems to leave no doubt that the call actually chosen is included.

Changing "logical to "plausible" doesn't help - that still allows the possibility of choosing an action that is implausible in the authorised auction but a good idea in the unauthorised auction.

As with many of the laws, they should just have used a few more words. "May not choose an action that could demonstrably have been suggested over a logical alternative" would do.

[gwnn]

Out of pure curiosity, what happens when there are three logical alternatives? Suppose we think there are three calls after some slow action by partner: pass, x and 4S. If partner had not blessed us with the benefits of UI, we judge the three actions had an IMP expectancy of +3 -1, -2 respectively. With the UI it would be +6, -2.6, -3.4 (these numbers are absurd but at least sum up to 0). Must I bid 4S even though double lost the most EV?

Another situation is if instead of -2.6, double is only a -1.2 (and 4S became -4.8), in this case double is still not suggested by the UI, although it improved relative to 4S.

[gnasher]

gnasher, on 2012-May-22, 01:11, said:

they should just have used a few more words

Actually. that's fewer words, isn't it?
"may not choose from among logical alternatives one that could demonstrably have been suggested over another" = 16
"may not choose an action that could demonstrably have been suggested over a logical alternative" = 15

[campboy]

gwnn, on 2012-May-22, 01:31, said:

Out of pure curiosity, what happens when there are three logical alternatives? Suppose we think there are three calls after some slow action by partner: pass, x and 4S. If partner had not blessed us with the benefits of UI, we judge the three actions had an IMP expectancy of +3 -1, -2 respectively. With the UI it would be +6, -2.6, -3.4 (these numbers are absurd but at least sum up to 0). Must I bid 4S even though double lost the most EV?

Another situation is if instead of -2.6, double is only a -1.2 (and 4S became -4.8), in this case double is still not suggested by the UI, although it improved relative to 4S.

I think in the first case that neither double nor 4♠ is demonstrably suggested over the other, since without the UI double is slightly better than 4♠ and with the UI double is slightly better than 4♠. Even if those are the exact numbers, you're not going to be able to estimate them accurately enough to tell whether the gap between dbl and 4♠ has increased or decreased.

In your second case double is slightly better than 4♠ without the UI, but significantly better with the UI, so probably is demonstrably suggested over 4♠.

[bluejak]

gwnn, on 2012-May-22, 01:31, said:

Out of pure curiosity, what happens when there are three logical alternatives? Suppose we think there are three calls after some slow action by partner: pass, x and 4S. If partner had not blessed us with the benefits of UI, we judge the three actions had an IMP expectancy of +3 -1, -2 respectively. With the UI it would be +6, -2.6, -3.4 (these numbers are absurd but at least sum up to 0). Must I bid 4S even though double lost the most EV?

You may not choose an action suggested over an LA by the UI. So, if partner's UI suggested pass over 4S then pass is illegal. If partner's UI suggested 4S over double then 4S is illegal. So you must double. Just apply the Law separately to each pairing.

I do not understand your figures but fancy it is the wrong approach: if you know the effect of various calls then you may easily consider there are no LAs to the action seen as most successful.

No doubt someone will ask [they always do] what happens if the UI suggest A over B, B over C, and C over A. That is the trouble with theoretical examples: they do not always make sense. My simple answer is that there is no such case. When I have said this before on various forums people have told me I am wrong but in fact have never come up with a convincing case. I shall worry about it only if they do, and not worry very much. There are always cases that are difficult, as we have seen many times on this forum.

[gwnn]

bluejak, on 2012-May-23, 05:00, said:

I do not understand your figures but fancy it is the wrong approach: if you know the effect of various calls then you may easily consider there are no LAs to the action seen as most successful.

Sometimes I know that one particular action is clearly best but I also know that many people of my peers would disagree with this (a simple example for this is that I gave a lot of thought to particular classes of bidding sequences and hence I'm better than my peers in those situations whilst being worse in many other situations). In this case there are LAs, as per the laws. Am I wrong about this?

PS: sorry gnasher for hijacking your thread.

[mjj29]

gwnn, on 2012-May-23, 06:11, said:

Sometimes I know that one particular action is clearly best but I also know that many people of my peers would disagree with this (a simple example for this is that I gave a lot of thought to particular classes of bidding sequences and hence I'm better than my peers in those situations whilst being worse in many other situations). In this case there are LAs, as per the laws. Am I wrong about this?

I'm of the opinion (although it's not shared by everyone) that as a player L73 is the important thing - are you carefully avoiding taking any advantage of the UI? If for you there is only one LA, then I believe you have done that.

However, the TD (since he is not telepathic) can't use that basis and instead uses L16 for guidance. This may mean that you don't violate L73, but still get a ruling under L16 that changes your result on the board. This is fine and you should take the ruling with good grace (or appeal).

If you want to try and get a better result by playing well in the contract you think the TD would likely adjust the score to then that's also fine, but I don't think you need to anything other than follow L73 at the table.

Matt

[mjj29]

bluejak, on 2012-May-23, 05:00, said:

No doubt someone will ask [they always do] what happens if the UI suggest A over B, B over C, and C over A. That is the trouble with theoretical examples: they do not always make sense. My simple answer is that there is no such case. When I have said this before on various forums people have told me I am wrong but in fact have never come up with a convincing case. I shall worry about it only if they do, and not worry very much. There are always cases that are difficult, as we have seen many times on this forum.

If you ever think you have this situation I think that actually none of them are suggested so you're free to take any of them. Like you say though, I don't think it's really possible (you're wrong about one of them).