[lalldonn]

bluejak, on 2012-August-02, 15:50, said:

The poll results are 25 for double and 15 for pass. I have no idea whether some of these are tainted as someone put it, but I shall assume not.

Probably an incorrect assumption, although necessary if you are to use the results.

How many passers out of the 40 respondents would you think is the minimum that makes pass a logical alternative?

[mjj29]

lalldonn, on 2012-August-02, 17:17, said:

Probably an incorrect assumption, although necessary if you are to use the results.

How many passers out of the 40 respondents would you think is the minimum that makes pass a logical alternative?

In the EBU I believe it's something like 8 would consider it of whom 2 would actually select it. So even if you discount a lot of the 'would pass', or change them to merely 'consider', it's still going to be an LA.

You could also argue about whether they are peers, of course (-:

[barmar]

mjj29, on 2012-August-02, 18:01, said:

You could also argue about whether they are peers, of course (-:

Which could result in circular reasoning: anyone who takes action different from what the player actually did isn't a peer, so there's only one LA and UI is irrelevant.

[campboy]

barmar, on 2012-August-02, 21:16, said:

Which could result in circular reasoning: anyone who takes action different from what the player actually did isn't a peer, so there's only one LA and UI is irrelevant.

Only by begging the question of whether the player's call was influenced by UI. If it might have been then there is no reason why peers without UI should be expected to produce the same call.

[iviehoff]

bluejak, on 2012-August-02, 15:50, said:

I was flabbergasted to be ruled against. ... I put this here to see if everyone agreed with the TD. It appears the general feeling is Yes.

It just goes to show how very difficult it is to comply with the law on UI, the best you can do is enter into the spirit of it, and be ruled against if your guess of the LA situation turns out to be wrong.

It also reminds us that LAs can be irrational, simply because there is an element of irrationality in a game that involves making decisions in conditions of risk and uncertainty. I chose to pass because I'm a wuss: because I have too many unpleasant memories of 5-level doubled contracts like this making, sometimes because I chose (guessed) the wrong lead. But of course that is irrational: in reality it is only a small percentage of them, which are prominent in memory only because of their unpleasantness. It is a human trait well known to psychologists to make decisions which avoid the risk of unpleasant outcomes.

[barmar]

campboy, on 2012-August-03, 02:25, said:

Only by begging the question of whether the player's call was influenced by UI. If it might have been then there is no reason why peers without UI should be expected to produce the same call.

But isn't that putting the cart before the horse? The law says you determine the LAs first, then decide if the choice among them was influenced by the UI.

You come to the table, and the player says "I'm an aggressive bidder, and would never even consider passing. Anyone who does is a total wuss, and should not be considered my peer." If you consider this to be true, can a poll be useful?

I think we've had variations of this discussion before: how similar do players have to be to be considered "peers"? For polls to be useful, they probably have to be very broad categories. The narrower they become, the more biased the results are likely to be. When we say that criminal cases should be decided by a jury of the one's peers, we obviously don't mean that it's full of accused criminals (and as this site points out, even broader categories like gender or race are too narrow -- the main point is to ensure that peers are not excluded).

[campboy]

barmar, on 2012-August-03, 08:22, said:

But isn't that putting the cart before the horse? The law says you determine the LAs first, then decide if the choice among them was influenced by the UI.

The hypothetical argument you mentioned is what puts the cart before the horse: it starts from the assumption that the player would always have made the same call and deduces (unsurprisingly) that there are no LAs.

[Jeremy69A]

I think it is clear to double. However the poll suggests pass is a LA so I would double(because I think it is right) and not sulk(much) if it were taken away.

[mamos]

Jeremy69A, on 2012-August-04, 03:22, said:

I think it is clear to double. However the poll suggests pass is a LA so I would double(because I think it is right) and not sulk(much) if it were taken away.

I'd sulk

[Fluffy]

The winning action can be double or bid on, pass can never be the winner, this doesn't mean that pass cannot be best on average in some circumstances, but this really doesn't look like one of them, opponents are never making overtricks to 5♠ this is enough for me to double.

IMO UI suggests biddig 5NT, double is the only LA not suggested by UI.

[mamos]

Law 16 B 1 (b)

(b) 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.

In the past few years the "poll" has become the normal way that TDs and ACs have decided on Logical Alternatives.

I'm not sure that these polls are always as trustworthy as it seems to be assumed. Anyone, with any vague understanding of social science, knows that manner of asking questions, the wording of the questions, even who it is that asks the questions, can all affect the answers that are given.

Two polls have been mentioned in this thread. The first poll was conducted within this forum by Bluejak himself. The thread title in itself (in my no so humble opinion) completely invalidates the whole process. Those answering it are immediately clued in to the purpose of asking the question. To establish LAs we need to discover what players would do at the table. Such a poll is invalid - it does not measure what it purports to measure.

Moreover Law 16 requires that we consider LAs in relation to "the class of players in question." Bluejak has told us that he was the player in question. High level decisions are ones that expert players are much better at making than average players. Of the 40 or so who responded I doubt if half maybe even a quarter are players at Bluejak's level. My own experience is that expert players are quite good at predicting the actions of weaker players but the reverse is not true.

As someone who trains TDs, often works as TDIC and who reads and reviews Appeal Forms, TD Polls concern me greatly. Certainly in the EBU, TD training has not really yet concerned itself with this a great deal but here are a whole host of issues. Proper consideration of player peers is one concern - I don't want to criticise TD colleagues but I've seen little (no?)documentation that reassures me. Players at an event like the one played at Scarborough have all played the same boards. I doubt if players with prior knowledge of the hands and their personal outcomes can ever be used in polls like this. At the very least polls should be written - perhaps with a pro forma and presented to the AC. It's just too easy to (unintentionally) bias the answers you get by the way you ask questions. Quite worryingly TD and ACs seem prepared to take these straw polls as total evidence of what counts as LAs

Bluejak wrote
The TD told me he had polled some good players [he did not say how many] and three considered pass, of whom two passed. So I could not fault his methodology.

It seems to me that such a poll is unlikely to be worth the back of the fag packet it was probably written on.

[c_corgi]

mamos, on 2012-August-07, 15:04, said:

Proper consideration of player peers is one concern...

Indeed. It seems clear now that for Bluejack pass was not an LA. For me it marginally is. I wasn't in Scarborough, but if I had been (and been in Bluejack's situation) I would imagine that the TD would have polled much the same group of people as for Bluejack. Worse, he might have polled me for Bluejack or vice-versa, which would mean that pass would be more likely to be deemed an LA for Bluejack than for me, when it should be the other way around.

Incidentally, is there an onus on a player to self-regulate his actions in these UI situations? I would have thought that if I think I have an LA, then I have one, but instead it appears that it is only an LA if the poll determines that it is.

[gordontd]

mamos, on 2012-August-05, 08:15, said:

I'd sulk

I can visualise it.

[mamos]

c_corgi, on 2012-August-07, 15:43, said:

Indeed. It seems clear now that for Bluejack pass was not an LA. For me it marginally is. I wasn't in Scarborough, but if I had been (and been in Bluejack's situation) I would imagine that the TD would have polled much the same group of people as for Bluejack. Worse, he might have polled me for Bluejack or vice-versa, which would mean that pass would be more likely to be deemed an LA for Bluejack than for me, when it should be the other way around.

Incidentally, is there an onus on a player to self-regulate his actions in these UI situations? I would have thought that if I think I have an LA, then I have one, but instead it appears that it is only an LA if the poll determines that it is.

But you see the questions you should be answering (and indeed be asked) is "What calls do you consider?" - "What call do you make?" -

As I understand it your original choice would be "Double".

You are not being asked "Is Double a LA?"

[c_corgi]

mamos, on 2012-August-07, 17:06, said:

But you see the questions you should be answering (and indeed be asked) is "What calls do you consider?" - "What call do you make?" -

Whatever the form of wording, I would have said something to the effect of "I double but consider pass to be an LA". I would not feel comfortable answering the question without mentioning pass as an LA. It seems unlikely that if peers of Bluejack are being polled they will fail to realise the underlying issue no matter how carefully the question is phrased.

[mjj29]

mamos, on 2012-August-07, 15:04, said:

As someone who trains TDs, often works as TDIC and who reads and reviews Appeal Forms, TD Polls concern me greatly. Certainly in the EBU, TD training has not really yet concerned itself with this a great deal but here are a whole host of issues. Proper consideration of player peers is one concern - I don't want to criticise TD colleagues but I've seen little (no?)documentation that reassures me. Players at an event like the one played at Scarborough have all played the same boards. I doubt if players with prior knowledge of the hands and their personal outcomes can ever be used in polls like this. At the very least polls should be written - perhaps with a pro forma and presented to the AC. It's just too easy to (unintentionally) bias the answers you get by the way you ask questions. Quite worryingly TD and ACs seem prepared to take these straw polls as total evidence of what counts as LAs

Indeed, there are many problems with polls as used in practice. This is why bluejak often writes that they are aids to a TD's judgement, not the be-all and end-all. However, when at an event and all of the players have played all of the boards, save for the TDs, who will only coincidentally be 'peers' of the players in question, and often not at all, what do you suggest is done instead?

Certainly there are bad polls and one should be asking 'what actions do you consider and what would you take', without disclosing the UI, but even the good polls suffer from those problems. In that case, however, there is noone to poll but the player in question and while just asking them would be simpler it may suffer from other problems instead... Should TDs and ACs not poll at all and rely on their own judgement when they themselves are not peers of the players in question?

[bluejak]

c_corgi, on 2012-August-07, 15:43, said:

Incidentally, is there an onus on a player to self-regulate his actions in these UI situations? I would have thought that if I think I have an LA, then I have one, but instead it appears that it is only an LA if the poll determines that it is.

The problem with self-regulation is shown by this example. I considered my double so completely automatic that it never occurred to me that anyone would pass. I was surprised by the opponent even asking for a ruling - I would not have in the reverse situation - but confidently expected a ruling in my favour.

I think I can sulk better than mamos, can't I, Gordon?

[billw55]

mjj29, on 2012-August-07, 17:29, said:

Indeed, there are many problems with polls as used in practice. This is why bluejak often writes that they are aids to a TD's judgement, not the be-all and end-all. However, when at an event and all of the players have played all of the boards, save for the TDs, who will only coincidentally be 'peers' of the players in question, and often not at all, what do you suggest is done instead?

Certainly there are bad polls and one should be asking 'what actions do you consider and what would you take', without disclosing the UI, but even the good polls suffer from those problems. In that case, however, there is noone to poll but the player in question and while just asking them would be simpler it may suffer from other problems instead... Should TDs and ACs not poll at all and rely on their own judgement when they themselves are not peers of the players in question?

Agree. Polls aren't perfect and a good TD should consider the possible flaws. But would it be better to have all TDs rule on LAs unilaterally? How can it be bad to gather more data to consider when making a ruling?

[barmar]

billw55, on 2012-August-15, 06:40, said:

How can it be bad to gather more data to consider when making a ruling?

More data is not necessarily more good data. The utility of poll data is very dependent on selecting the right pollees.

[lamford]

Conducting the poll without the UI is useful, but it is just as important to conduct the poll with the UI. In this example, I regard double as automatic, but would pass because I think that Pass would be chosen by enough of my peers to make it an LA. I am surprised by the size of the Pass vote, but not that there are some votes for Pass, so the fact that "some of them would select it" is clear.

And it is a pity that it is not a public poll, so that I could see which loony voted for 6♦