Join Bridge Winners
All comments by Wayne Burrows
1 2 3 4 ... 38 39 40 41 42 43 44 45 46 ... 62 63 64 65
You are ignoring the author of this comment. Click to temporarily show the comment.
This has been well known for a long time.

Some claim otherwise but the definition of HUM included the standard short 1 opening by any plain reading of the language.
Dec. 16, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Peter, depends who is on lead if you miss the A so 50-50 grand in that case.
Dec. 14, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Actually 40A1(b).

40A3 is also relevant. “A player may make any call or play without
prior announcement provided that such call or play is not based on an undisclosed partnership understanding.”

To me, this plainly says if you have not disclosed your partnership understanding then you are not allowed to use that method.
Dec. 11, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
I think it would be problematic if a logical alternative was defined based on the opinion of a person who did not ask what the auction meant. Therefore at least ordinarily an explanation should be provided.
Dec. 10, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Michael: Perhaps we wrote at cross purposes. I meant the a priori odds without the consideration of whether or not there as a tempo break.

The comment that there is a huge chance of a tempo break with the A is interesting because it is almost now as if an in tempo pass conveys information that the A is not held. Is the situation really that sensitive?

The situation where Grue might have the A is also interesting. If this is a real possibility then I suspect they will need a partnership discussion.

Suppose Grue has the A and Levin bids Blackwood with a void then when he hears two key cards and assumes they do not include the A then they could be in a no play grand whatever the strain.

Further if Grue might have the A then you and others have ignored that possibility. Say this is the hand where Grue does have the ace if such a hand exists. And they get to grand and south is looking at another ace. Now he might hesitate not sure what to do and especially if he thinks that partner might think a double asks for a diamond lead. To me the possibility of Grue having the A argues against the hesitation suggesting a diamond.
Dec. 9, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
“Which is the better way to show ♦ A over 4♦ x
- XX
- 4NT”

I think this depends on:

1. Whether you want more information. If you redouble partner can conveniently cue 4 and you can gather more information. Alternatively 4NT is taking control. But as we see from Grue's correction to 7NT then it is important if you bid 4NT to confirm all the key cards with 5NT before bidding the slam opposite an unlimited partner.

2. On general principles it is better to allow the unlimited partner to take control of the auction and given west was limited by 3 then with the A it will frequently be best to redouble rather than 4NT.

3. Against 2. since Levin has a source of tricks KQJxx then it can be better for him to take control. If Grue took control on some auctions it might be difficult for him to find the Q.
Dec. 9, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Are players required to fill in an appeal form in the ACBL?

Amazingly my two ACBL tournaments - one regional and one national - did not see me before an appeal committee.
Dec. 8, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
I have heard a national representative tell her partner that he should have known that she had a good hand because of her sides tempo issues.
Dec. 8, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
The laws do not mention polling peers. Polling is some procedure that directors have used in practice to answer the question. The methodology of the polls is very often flawed. It is always flawed if very small samples are used.

It is also flawed if players will say in response to the poll things like what is reported above “the third said that partner would have doubled clubs with the ♣A and doubled 7NT with the ♦A.”

The law requires the judgement to be made with reference to a player using the same methods as the pair in question. There is nothing in the write up that suggests that this pair would have doubled with the A or the A as suggested by the pollee. The fact that the player did not double with the A suggests that a double might have suggested something else.

The poll results were “In response to the first question, three of the pollees said they would lead a diamond and the same players said the hesitation suggested a diamond lead. The other three players said they would not lead a diamond.”

This does not establish a logical alternative by any analysis. What is the logical alternative? The law says “A logical alternative is an action …”

No action has been determined. “(N)ot lead a diamond” is not an action. It is a cumulation of actions.

The subsequent poll is reported as follows “When asked what they would lead, two chose a diamond – with one, the king was considered. One considered a spade or a diamond; one considered a spade or a club and one chose a spade.”

So one person “chose” a particular lead that was not a diamond. That is one out of five. This is not conclusive as to the question of whether

“a significant proportion of the class of players in question, using the methods of the partnership, would seriously consider, and some might select. ”

I do not know if the ACBL has guidance on this law as to what percentage would be considered a “significant proportion” and what would be considered “some”. In New Zealand these are defined as: “significant proportion” - more than one in four; and “some” - more than an isolated instance.

The one in five that considered a club would not be a significant proportion even if the sample was considered definitive and the two in five that considered a spade would not, in my opinion, convincingly show that there were a significant proportion. There needs to be some understanding of the amount of error in such a poll. If five people are polled and less than 25% would consider such an action then we cannot expect that every poll will give zero or one respondents saying that they would consider it. That is not how statistics work.

These are the rough numbers that I calculated. For a poll of six people from a population of 100 (my ball park figure) where 25% would consider an action then you will get two or more saying they would consider it around 47% of the time.

If you conclude that such a bid is a logical alternative because 2/6 is bigger than 1/4 then you will be wrong nearly half of the time when in fact it was not a logical alternative by your definition. That is how flawed this method is.
Dec. 8, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Jim: I do not accept that this is the end of the case under the current rules. As I have said elsewhere in this thread proving a threshold of say 10% from a poll of fewer than 10 people is extremely problematic. A much larger sample would be needed to determine that more than 10% would choose an action with anything like certainty.

In this case we have additional problems that have manifest. For example, did the pollees consider the methods of the partnership?

Logical alternative was in the law before the definition which now is included in the law. I do not think the definition is close to optimal. In particular I have often thought about the situation with weaker players. For weak players in many situations almost anything is by definition a logical alternative because there are some other weak players out there that would consider it. This is still the case if the particular person would never take the action claimed to be logical alternative by this definition. The problem is also caused by their being no definition of “class of player”.

The problem here as I see it is an extension of that. Obviously these are not weak players and in particular Glubok whose lead is under scrutiny is a not a weak player. Nevertheless when the logical alternative is defined as what someone else might do and then that is determined by a proxy from a loosely defined poll are we really determining the logical alternatives for that player with any certainty.

There will always be questions about the poll when it is so unscientific.
Dec. 8, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Michael: Thanks for your numbers.

I am a bit surprised that you think that the A being held is so low at 30%. That means that 70% of the time Levin has the A and did not either redouble 4 to show it or bid 5NT to investigate a grand including the possibility of 7NT as Grue can easily have significant extra values over what was shown by 4.

I am not sure what hands you are imagining that Levin might have to bid this way with the A. The hand I suggested earlier based on the actual hand but changing the QJ for the A in high cards - Kxx KQJxx A Axxx - counts for 12 top tricks in no trumps opposite most bare minimum hands for Grue's bidding. The K, in a suit that has been cue-bid, could easily be the thirteenth as would a sixth spade.

Grue arguably made a poor bid with 7NT, although if Levin's bidding could hide the A then the gamble on the A could be the percentage action with extra tricks. However, to me it is almost incomprehensible that Levin would not investigate or at least allow the unlimited Grue to investigate. Therefore I am inclined to think from a class player like Levin that there is a much higher chance that he does not have the A and therefore a priori that partner does have that card.

“In general, I don't want to send the message that you can turn an ‘almost sure thing’ into a ‘lock’ via UI.”

I understand what you are saying but I am not convinced that this is what the laws say.

If I am defending 7NT and even with your numbers if there is a 30% chance that the opponents are off a specific ace and I know that their long suits are breaking well and will almost certainly produce 10 tricks and then counting two aces and the extras that Grue has announced with 7NT then I am not convinced that we should be forcing a person to give up the 30% chance of a set outright.

“Especially in a case such as this - where the ‘winning’ lead might not even occur to the player without ‘help’.”

The problem is that this is speculative. The law does not and cannot deal with what is in a player's head at the time he made a lead. The law has an objective test.

If the impropriety you suggest exists then it needs to be dealt with by other mechanisms and needs to be dealt with based on evidence of that specific impropriety and not by taking away a close to 100% action speculatively.

Based on your numbers, if a diamond lead is going to lead to a set at least 30% and another lead is going to lead to a set virtually 0% how can we really say that there is a logical alternative to a diamond.

The approach you are advocating is at least in this particular case “if it hesitates shoot it.” I do not believe that is what the law requires.

You seem to be arguing for that approach in what you describe as particularly tempo sensitive situations. You are entitled to that view but we cannot force that on a situation when it is plainly illogical to lead a non-diamond absent any UI.
Dec. 8, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Ray: Without the break in tempo the odds are already heavily stacked in favour of needing partner to have the A to beat this contract.
Dec. 8, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
I think that is the most likely scenario but even before the poll it is unclear to me that the directors or panel determined what the NS methods or even lack of methods were.
Dec. 8, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
First answer: it is not possible.

Second answer: you can do this if you make assumptions. The quality of the results will depend on your assumptions and those assumptions would be subject to subjective analysis.

Third answer: Yes I have done some simulation analysis on this problem. I used dealer and leadsolver. Lead solver is a double dummy analysis which is of limited use since play in the real world is not double dummy.

I am working from memory here because I am on a different machine than I did the simulations and I probably did not record every single run that I did.

Firstly, I had to decide what the auction meant. This is where assumptions come in.

1. Levin has an opening bid (1) and some extra values (3). These are relatively easy to constrain but not perfectly. What Levin jumps to 3 on might not be the same as what I jump to 3.

2. Grue has five spades assuming that is what 1 showed.

3. Grue has enough to make a slam try. I have no idea whether they play a serious/frivolous mechanism so do not know how serious the slam try is with 4 - his hand makes it look serious but we do not know how much ligher (or stronger) he would make the same bid.

4. After this we do not know what hands Levin would bid 4 as he did compared with some other action like 4NT or 5.

5. After 4 X came back to Levin we do not know what hands he would have taken some other action - XX to show his A etc.

6. After 5 we do not know when Levin would have continued with 5NT or some other bid.

Many of these things that we do not know there is no way of knowing for certain. Levin might not know precisely what hands he would bid one way or another until he holds the hands at the boundaries.

One inferential issues that I think is very relevant is Levin's heart holding to bid a grand given that Grue has not denied three hearts and has not shown any strength there other than possibly the ace in response to Blackwood - this is not the actual Grue hand but all hands consistent with the bidding. I doubt that Levin would gamble on the K but I do not know how much he would gamble on the Q - is AKJxx enough what about AKTxxx etc?

I started with some ballpark assumptions that coded the auction:

a/ Generic descriptions of Levin's 1H … 3S sequence and Grue's 1S response.

b/ Some extra values for Grue for 4C.

c/ A club cue for Grue - Ace, king, or shortage.

d/ No diamond ace for Grue.

e/ Either the diamond ace or void for Levin!!!

f/ Some conditions that meant that Levin had a source of tricks which was solidish hearts, or extra heart length with at least the king, or solid clubs on the side.

g/ I also gave north his actually hand obviously.

When I first ran these leadsolver told me the best double dummy lead was a diamond. Obviously catering to the diamond void. Interestingly it was not the diamond king which caters to the negative problem of stiff ace opposite Qx(+). A diamond lead was best by a big margin. Sorry I do not remember the numbers. I ran a sample of 10000 hands.

I then looked at how often Levin would have a void and I got numbers about 15% for this first iteration.

Subsequently, I have looked at scenarios where a diamond void is more likely. Others have pointed out reasons for this as Levin would want Grue to correct to 7NT with extra values at MPs if he held the A. However again it is impossible to know how often if ever Levin would bid the way he did with the A.

All I can conclude is that it is likely somewhere between 15% and 100% that Levin has a diamond void for this auction depending on inferences about his style (or possibly propensity to error in leaping to 7).

And it is very likely with no other information than the explicit information in the auction and no additional inferential information that a diamond lead is far and away the best double dummy.

Also I eyeballed smaller samples of hands consistent with my naive assumptions that allowed for diamond ace or void and about 2-3 in a hundred a diamond lead gave up the contract compared with 10-20 hands where a diamond lead beat 7NT.
Dec. 8, 2018
Wayne Burrows edited this comment Dec. 8, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Michael what are your probabilities without the hesitation (and with pass being consistent with the A being possibly held)?

Unless Levin has bid incredibly badly, I think that the probability of the A being missing is incredibly high - some say almost certain.
Dec. 8, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Another thought.

If we accept that a hesitation from south suggests the diamond ace then we have the implication.

Hesitation implies diamond ace.

This is logically equivalent to

Not the diamond ace implies no hesitation.

Now consider the case where the bidding has gone off the rails in some other way and south holds the A.

Unless we are convinced that if south had the A then he would have an auto double or auto pass then

club ace and not diamond ace would also be consistent with a hesitation.

And therefore

Hesitation implies (club ace or diamond ace) - and or possibly some other holdings.
Dec. 7, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Kit, if south was thinking and not distracted then your first assertion is true.

Your second assertion is not true as I see it. That might depend a little on what you mean by a bad split. But south could hold a solid holding in hearts so that he knows that suit won't run, say JT9x. Or better still and perhaps more needed for a contemplation of double, a holding in both hearts and clubs and knowing that partner has diamonds held he thinks there is unlikely to be a route to 13 tricks. Levin was expecting to be able set up hearts in 7 but Grue bid on just based on extra spade length which he translated into extra tricks.

Your third assertion is based on believing Levin's auction to 7. However if you believe that auction then you get into an immediate problem that not only has Levin not shown the A he has probably denied it. Is it really conceivable that a world class player with 16 hcp opposite an unlimited partner holding all of the key cards does not investigate a grand slam in no trumps. Take away the QJ and give Levin the A then he would have Kxx KQJxx A Axxx opposite a known AQxxx A and a club cue. That gives something close to a favourite for five spades, five hearts, a diamond and a club and there will be a thirteenth top trick if Grue has the K. Yes we have to deal with the spade break but we need that in 7 and the heart break but partner might have two or three hearts and that improves immediately. Grue also could easily have a sixth spade for a thirteenth top trick. This leads us to conclude that at MPs Levin would have done something different with the A.

What this means is that even if the UI suggested a diamond then the reasoning to demonstrate that suggestion requires that there is also AI that overwhelmingly suggests diamonds too.

Further I am unconvinced that the information in south's hesitation and what it suggests should be properly demonstrated by inference from the opponent's bidding. It is the opponent's bidding that suggests that if south has an ace it is most likely the A. South's hesitation does not in itself suggest one lead over another.

And finally, I think a poll is a very poor way to demonstrate something. Five opinions do not a demonstration make. To demonstrate is to clearly show with proof or evidence, not to ask some people what they think is suggested. Asking what someone thinks is suggested just introduces another layer of biases.
Dec. 7, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
Ray, of course it is not inconceivable. He did have the A. I have no idea of what point you are trying to make.

Nothing he did announced the fact that he held the diamond ace.
Dec. 7, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
“You continue to ignore my main point; the total change of mindset that is brought about by the BIT. All the highfalutin analysis and arguments for the ♦ lead are basically irrelevant. Partner should not make a BIT which basically announces ‘I have an ace’.”

Previously, you said “without UI (or ‘help’), the mindset of a player is totally different.”

I do not accept that the break in tempo announces (basically or otherwise) “I have an ace.” That is only one of many possible reasons for the break in tempo.

If a break in tempo is being used to announce that a player has the ace then that is a whole different matter than UI. It is illegal communication between partners.

The notion that the mindset is changed is speculative and not established.

Adjustments in UI situations are based on bids or plays that are “demonstrably suggested” and “logical alternatives.”

We cannot determine the mindset of a player. There is no way of knowing for certain whether Glubok had thought of the diamond lead before the hesitation or not. If not then there is no way of knowing whether he was triggered into his thinking by the hesitation.

Therefore we need an objective measure. And the objective measure is based on actions being “demonstrably suggested” and that there is at least one “logical alternative”.

My contention is that a slow pass does not demonstrably suggest a diamond. And further that a zero play should not be considered by anyone to be a logical alternative whatever the result of a poll.
Dec. 7, 2018
You are ignoring the author of this comment. Click to temporarily show the comment.
'“ What is the hand you are hoping to defeat 7NT with a non-diamond lead that is consistent with this auction?” And similarly, elsethread Michael H:

That might SEEM like the logical question but, when it comes to UI, it is not the main question. It's barely even a relevant question.'

The law is defined in terms of logical alternatives which are defined in terms of one's peers, ‘class of player’, using the same methods. Unfortunately, ‘class of player’ is not defined.

I seriously doubt whether a player that makes a safe lead to allow the opponents to take their thirteen tricks is a peer or in the same class as a player who makes the best effort to defeat the contract when the opponents are in 7NT.

A safe lead when the opponents almost certainly have thirteen tricks is absurd when there is an appreciable chance the opponents have screwed up. And that information is available from their auction.

Even allowing for a poor bid by Levin, to not investigate 7NT opposite the unlimited Grue, a priori there is a significant chance that the A is not held and that makes other leads extremely poor on this auction - when a hand not in control corrects to 7NT not knowing that a particular ace is not held.
Dec. 7, 2018
1 2 3 4 ... 38 39 40 41 42 43 44 45 46 ... 62 63 64 65
.

Bottom Home Top