UI and probabilities

[Vampyr]

Yes, but only if is a consequence of the UI that you believe a borderline good try is more likely; not if that is AI anyway.

So what you are suggesting here, Frances, and what has been suggested elsewhere, is that absent the UI partner is still likely to have a weaker or less shape-suitable hand than average for an invitation, so this information is AI?

 

I don't buy it, but I am not sure why.

[bluejak]

I see the logic, but somehow it still does not feel right.

 

So with no UI, partner's raise is

a 16 count 23% of the time

a 17 count 68% of the time

an 18 count 9% of the time (Bayes' theorem)

 

His expected HCP given a raise is 16.9

 

Now we are suggesting that the UI tell you that partner doesn't have a 17-count - he is either thinking of passing with a 16-count, or bidding game with an 18-count. Now he has a 16-count 5/7 of the time, and an 18-count 2/7 of the time, making his expected HCP 16.6

Yes, but the problem is that is not the way people think. Suppose he has 8 HCP. Now he might think "This is teams, I dare not miss a routine game, and over half the time we have 25 HCP. So I shall bid game."

 

Now, assuming a less than ethical player, who has UI, how does he think now? He might think "Partner does not have 17, and 16 is over twice as likely as 18, so I shall pass."

 

So he is staying out of game when here is UI because of the UI, and his action is correct because of the UI. Surely illegal.

[gnasher]

Suppose you play the 'normal' forcing pass approach where bidding at once is weaker than passing then bidding if partner doubles. If you think for a long time before making a forcing pass, the UI doesn't tell partner anything useful.

That depends. What if hands where it's close between passing and doubling are more common than hands where it's close between passing and bidding?

[Vampyr]

I see the logic, but somehow it still does not feel right.

 

So with no UI, partner's raise is

a 16 count 23% of the time

a 17 count 68% of the time

an 18 count 9% of the time (Bayes' theorem)

 

His expected HCP given a raise is 16.9

 

Now we are suggesting that the UI tell you that partner doesn't have a 17-count - he is either thinking of passing with a 16-count, or bidding game with an 18-count. Now he has a 16-count 5/7 of the time, and an 18-count 2/7 of the time, making his expected HCP 16.6

Yes, but the problem is that is not the way people think. Suppose he has 8 HCP. Now he might think "This is teams, I dare not miss a routine game, and over half the time we have 25 HCP. So I shall bid game."

 

Now, assuming a less than ethical player, who has UI, how does he think now? He might think "Partner does not have 17, and 16 is over twice as likely as 18, so I shall pass."

 

So he is staying out of game when here is UI because of the UI, and his action is correct because of the UI. Surely illegal.

You are restating the original premise here. Since partner is likely to be weaker, is the ethical player forced to bid game?

[bluejak]

Yes, it seems so. In this particular case, anyway.

[Vampyr]

Yes, it seems so.  In this particular case, anyway.

That's what I think. Except that there is now the suggestion that partner's invite is statistically more likely to be under-strength than over-strength anyway, without UI.

[bluejak]

If you have a hand that refuses a try anyway, it matters not. But we are talking of accepting a try without the UI and refusing it with: that's illegal.

[campboy]

With Bluejak's 8-count the prior chance that we have 25+ points between the two hands is 77%; with UI this drops to 28%, which seems like a pretty good argument for pass being suggested.

 

On the other hand, if our hand was a 7-count the chance of 25+ points is 9% without the UI but 28% with the UI, so now perhaps 3NT is suggested.

 

This makes sense: if we were told that partner had exactly 17 points we would want to bid 3NT with the first hand and pass with the second; it is reasonable that having the opposite information would suggest the opposite action in each case.

[Vampyr]

If you have a hand that refuses a try anyway, it matters not.  But we are talking of accepting a try without the UI and refusing it with: that's illegal.

Of course that's illegal. I think that the discussion here of what the UI does or does not suggest is geared toward rejecting the suggested call, if anything.

[shyams]

My (admittedly quick) read of this post tells me that the consensus is heading towards bidding 3NT rather than passing; given the UI.

 

The next question arises: How do you think most TDs would rule if the declaring side makes 9 tricks (on some lucky distribution) with a 16 HCP hand for opener?

 

My guess is that despite logical reasoning leads to a 3NT bid, when cards turn out to be favourable and declaring side makes 9 tricks, TDs or ACs would tend to disregard the rationale and adjust back (e.g. 2NT +1?) for UI.

[Cascade]

I had a situation like this recently.

 

I argued that there were two possibilities for the slow bid - weak or strong.

 

It turns out from simulations that I have done since that one possibility was approximately three times more likely than the other. Nevertheless the committee ruled unanimously that a break in tempo suggested the less 3:1 underdog hand which happened to correspond to the actual hand.

 

While they made the above determination they happened to rule for other reasons in my side's favour.

 

It would seem to me that at least on the surface this is wrong in that the UI information is more likely to be based on the more frequent hand type - possibly discounted as for example with one hand type you might always have a problem whereas with another hand type you might have a clear action some of the time and a problem only some percentage of the time (or something similar).

 

Having said that this is very much a theoretical position as in practice I suspect that many partners particularly in established partnerships can read their partner's mannerisms sufficiently well to know when there is an underbid and when there is an overbid.

 

I am not sure when the threshold is crossed and the probabilities are sufficiently high that a particular action is or is not demonstably suggested.

 

Personally subject to the practical problem that partner might well be an expert in reading this particular partner's mannerism I am increasing more inclined to not accept in these invitational auctions for one action to be suggested over another by the break in tempo.

[greenender]

I'm afraid I'm in the camp which discounts this as a practical issue.

 

Whilst I have to accept as a matter of mathematics that the removal of the middle band of the inviter's possible hand types exaggerates the difference between the relative probabilities of him being at the lower or higher ends of the range, when the problem is expressed in point count terms (as to which see below), it seems to me that there are too many other imponderables.

 

I kinow that Andy wanted a theoretical answer to a theoretical question, without side issues, but I guess he has got that from bluejak.

 

In practice the theoretical problem will be sidetracked by one or more of the following:

(1) alternative possible invitations which may account for the BIT

(2) the need for inviter to give relative weight to different features of his hand, which may suggest different things, in deciding whether to invite

(3) partnership and individual player style

(4) other UI such as body language which gives a clue to what the inviter has.

 

Besides, I don't think the analysis that (say) inviter will have 16 or 18 for the slow invite, from the original 16-18 range for any old invite, really stands up. If I invite, I tell my partner that I am too good to pass/sign-off and not good enough to blast game. In the absence of special agreements I assume that partner is invited to accept on the top 50% of hands for his bidding so far. Both inviter and his partner will take into account other things than just points in arriving at their decision.

 

I certainly wouldn't be thinking in rigid points terms, and I don't think most players experienced enough to appreciate what Andy is asking would either. Indeed, there is a good argument IMO that a slow invite (when extraneous factors are stripped out) tends to show the top or the bottom 20-30% by probability of the possible hands for any old invite, and to eliminate the middle 40-60%. After all, when I invite, I am trying to assess the probabilities of being able to make game. My view is that that is perhaps a better approximation of the way bridge players think than the point count anaylsis which leads to the conclusion that a slow invite tends to be at the weak end. It also gets rid of the problem, does it not?