[blackshoe]

MFA, on Aug 19 2009, 03:16 AM, said:

What irritates me here is that I know that many outcomes other than 5♣ would have been possible. But we can't really include them all in a big 7-8 way weighted score mix, can we?

Why not?

[MFA]

blackshoe, on Aug 19 2009, 05:59 PM, said:

MFA, on Aug 19 2009, 03:16 AM, said:

What irritates me here is that I know that many outcomes other than 5♣ would have been possible. But we can't really include them all in a big 7-8 way weighted score mix, can we?

Why not?

Because we would be implying a degree of precision in our ruling that is really not there.

[blackshoe]

MFA, on Aug 19 2009, 04:32 PM, said:

blackshoe, on Aug 19 2009, 05:59 PM, said:

MFA, on Aug 19 2009, 03:16 AM, said:

What irritates me here is that I know that many outcomes other than 5♣ would have been possible. But we can't really include them all in a big 7-8 way weighted score mix, can we?

Why not?

Because we would be implying a degree of precision in our ruling that is really not there.

I don't understand that. Can you elucidate?

[MFA]

blackshoe, on Aug 19 2009, 11:05 PM, said:

MFA, on Aug 19 2009, 04:32 PM, said:

blackshoe, on Aug 19 2009, 05:59 PM, said:

MFA, on Aug 19 2009, 03:16 AM, said:

What irritates me here is that I know that many outcomes other than 5♣ would have been possible. But we can't really include them all in a big 7-8 way weighted score mix, can we?

Why not?

Because we would be implying a degree of precision in our ruling that is really not there.

I don't understand that. Can you elucidate?

Ok.

If we make a 7-way weighted score, it will be along these lines:

A 40%
B 27%
C 14%
D 8%
E 5%
F 4%
G 2%

We can't really claim to have a good justification for numbers like that.
On the other hand, if we adjust to 1 (or 2 or perhaps 3 at most) scores, it seems clear that we are giving it our best shot at a reasonable ruling. Which is just what we are doing.

For the same reason, a 2-way weighted score shouldn't be 77-23 or something, but round numbers like 50-50, 75-25 or 60-40. It is not exact science and we shouldn't pretend it is.

[blackshoe]

Fair enough, I suppose, as far as it goes. But if the TD thinks that seven results have some potential, he should not ignore his responsibility on the basis of this argument, I think. Perhaps it depends on what one considers a reasonable probability of a particular contract occurring. For myself, I think 2% is too small, for one. I might, taking your example, do something like:

A 40%
B 30%
C 15%
D 5%
E 5%
F 5%

And toss G. There are other reasonable weightings, I think. Six results, rather than seven, but I think we're talking principles here.

[Cascade]

That is the fundamental problem with this approach.

Even something simple like 50/50 or 60/40 is very unlikely to be the correct weighting.

Is there any precedent for an appeal committee changing these weightings in a minor way - say from 50/50 to 55/45? I think resolving doubt in favour of non-offenders is much better than giving them some weighting of results that they might not ever attain at the table.

[blackshoe]

You seem to be suggesting that Law 12C1{c} is bad law. That may be the case, though I doubt it. It is, nonetheless, the law most of the world's TD's must use.

[Cascade]

blackshoe, on Aug 20 2009, 11:54 AM, said:

You seem to be suggesting that Law 12C1{c} is bad law.

In theory no. In practice yes.

This is what the law says.

"In order to do equity, and unless the Regulating Authority forbids
it, an assigned adjusted score may be weighted to reflect the probabilities
of a number of potential results."

The weightings according to the law must "reflect the probabilities". Basically this is an impossible task.

[helene_t]

What else do you suggest, Wayne? Just assuming that the most likely, or most beneficial for the non-offenders, result would always have happened absent the infraction?

But OK, for an appeal committee it will typically be easier to decide whether they agree with the TD's assessment as to the most beneficial result possible than as to a whole array of probabilities.

[blackshoe]

reflect, v.: embody or represent something in a faithful or appropriate way.

I do not see there any requirement for the degree of precision you seem to think necessary.

[Cascade]

1. At least "the most favourable result that was likely" for the non-offenders is objective

2. At least "the most favourable result that was likely" for the non-offenders does not profit the offending side

3. Weightings if wrong could easily profit the offenders

4. I have never had anyone talk to me before weighting my score (I haven't actually had many) and ask about our methods etc that would have a material affect on any proper weightings.

5. How do you weight a contract that depends on a guess - I played hand recently where the relevant guess was in hearts with KJxx opposite xx. If you guess correctly you make your contract if not then you go down. 50-50 can't be right - well certainly not for everyone. There might be information that affects your decision but how in practice do you translate that into a percentage. I have no idea how the directors determine these percentages and I doubt that they could satisfactorily tell anyone. What say you settle on 60-40 for the above problem and then the non-offending declarer tells you "I would always get this particular play right". Do you change to 100-0? What if instead it was the offenders who were declaring and they explain why they would always get this right. Do you readjust? It is much simpler just to give the non-offenders 100-0 and the offenders 0-100. Maybe it is not equitable but there is at least no danger of benefiting the offenders. What if as in the actual hand around 3/4 of those who played the contract got this decision right - does that really tell you the appropriate percentage for the table under consideration. It is a mine field. I have seen adjustments that include a weighting for results that I am convinced could never actually happen. What about a weighting for the probability that one side or the other will stuff up - everyone does that some time. The stuff-up percentage could be as high as 10% but I have never seen that included in a calculation.

Basically the weightings are never done properly and there is too much chance that the non-offenders will benefit. As a non-offender even if the weightings by some miracle are correct it seems wrong to me that I have been deprived of getting the best out of the hand and a more favourable result to ensure that I have not been damaged by the infraction would seem more fair.

If weightings are to be used then I think there needs to be very clear guidelines as to how the percentages are calculated including a safe guard that will ensure that non-offenders are not damaged.

[blackshoe]

It is a fair distance from "it is possible that weightings are off" or "figuring out proper weightings is difficult" to "all weightings are always wrong". I don't believe the latter.

[helene_t]

All weightings are always wrong, but not doing any weighting would be more wrong.

[bluejak]

The advantage of Law 12C1C over 12C1E is that the former is equity based, the latter punishment based. However, all adjustments under either Law are based on the 'best guess' of the TD [after consultation] or the AC and total accuracy is impossible. It is just as obtainable as complete accuracy by officials in other sports and mindsports: a totally unreal objective that thoughtless people expect, and practical people know is not worth worrying about. Fortunately, errors/misjudgements [whichever you like to call them] matter little under Law 12C1C but can have disastrous results under Law 12C1E. A small change of opinion in a weighted score probably affects thngs by one imp or MP: in a Law 12C1E ruling a small change can make a 13 imp or 70% of a board difference. Accordingly it has been realised that the total accuracy people want just does not matter nearly as much with weighted scores so worrying about it is unecessary.

Michael Askgaard's views seem reasonable. While a TD who wishes to can give a 7-way weighting and/or use figures like 3% or 54.7% it seems unnecessary and we are not trying to get such accuracy. What for? An imp? Probably not: the difference between 54.7% and a rounded 55% will usually be zero imps.

EBU White book #78.1.3 said:

Recommended is a facility for up to 5 different results, and 2 decimal places of weighting (eg, 33% which is shown as 0.33).

While a TD may give a weighting that does not fall within this recommendation, it will have to be entered manually into the software, and the EBU view is that it is unnecessary except in exceptional cases.

Quote

s there any precedent for an appeal committee changing these weightings in a minor way - say from 50/50 to 55/45? I think resolving doubt in favour of non-offenders is much better than giving them some weighting of results that they might not ever attain at the table.

I do not understand this. If 55/45 is correct why should the TD not do it? Why the AC? Surely you are not reverting to the ideas of 40 year ago where the AC does the TD's job for him.

Naturally, the EBU recommends 'sympathetic weighting', where a small 'benefit of doubt' bias is included. If a TD gives a 50/50 weighting it probably means that he thinks the actual weighting should really be 55/45 or 60/40 to the offending side.

Furthermore, to avoid unnecessry appeals, EBU ACs are asked not to fiddle with the weightings, not to make small adjustments, and to treat as frivolous any appeal based solely on an attempt to change the weightings in a non-major way.

So, if a TD rules:
.. 40% of +620
+ 30% of +170
+ 30% of -100

then an AC should not amend it to:
.. 30% of +620
+ 45% of +170
+ 25% of -100

but the followng are acceptable:
.. 50% of +620
+ 50% of +170
since they have over-ruled the TD on the possibility of a third score,

and also:
.. 10% of +620
+ 10% of +170
+ 80% of -100
since this is such a major change as to again be overruling the TD.

Quote

The weightings according to the law must "reflect the probabilities". Basically this is an impossible task.

No, it is not. It just means within reasonable limits. Remember this is not BLML: we approach this as a practical exercise, not a theoretical play on words.

Quote

Weightings if wrong could easily profit the offenders

As can any wrong ruling. Do not throw the baby out with the bathwater.

Quote

How do you weight a contract that depends on a guess - I played hand recently where the relevant guess was in hearts with KJxx opposite xx. If you guess correctly you make your contract if not then you go down. 50-50 can't be right - well certainly not for everyone.

So you come up with a guestimate, and do not worry out it. Simple!

Quote

If weightings are to be used then I think there needs to be very clear guidelines as to how the percentages are calculated including a safe guard that will ensure that non-offenders are not damaged.

Of course the EBU has guidelines, and I would recommend others to follow them. In general, they do not damage non-offenders. Mistakes are made, but it is rare that a weighting mistake will make a serious difference, unlike adjusting or not under Law 12C1E, or adjusting to the wrong score.

Quote

All weightings are always wrong ...

... but it does not matter.

[Cascade]

bluejak, on Aug 21 2009, 11:39 AM, said:

Furthermore, to avoid unnecessry appeals, EBU ACs are asked not to fiddle with the weightings, not to make small adjustments, and to treat as frivolous any appeal based solely on an attempt to change the weightings in a non-major way.

Another way for the system to bully the players.

If the weightings are wrong then the players have a right to get them put right.

But instead we have a regulation that says effectively that the TD is GOD.

This sort of attitude makes me feel sick.

[blackshoe]

Yeah, well, don't throw up in here. It'll be a pain to clean up.

[Cascade]

I'm sorry but this is getting utterly ridiculous.

A player turns up to play.

The opponents cause some problem.

The director adjusts the score incorrectly (poor director judgement)

The regulators say tough if you appeal you the non-offenders will we slapped with a penalty.

If this is really the process it is shameful.

[jeremy69]

Quote

Another way for the system to bully the players.

If the weightings are wrong then the players have a right to get them put right.

But instead we have a regulation that says effectively that the TD is GOD.

This sort of attitude makes me feel sick.

The system isn't bullying anyone. Weightings are an approximation hence the EBU guideline to appeal committees that minor tinkering is best not done. If the appeal view that the TD has got it substantially wrong then, of course, it can be changed. Note that it is a guideline so that an appeal committee in a picky mood could make a minor adjustment if it wished.

Quote

The regulators say tough if you appeal you the non-offenders will we slapped with a penalty.

A deposit would only be forfeit in England if the appeal was frivolous or so clearcut as to make everyone think it was waste of time. No other penalty would result. I can imagine an appeal committee viewing that an request to change a weighting from 50% to 51% to fall in that category but, in reality, nobody ever appeals on those grounds. I've done a lot of appeals but cannot remember any deposit being forfeit for an appeal involving a weighting adjustment.

[FrancesHinden]

There seems to be this assumption that TDs and ACs somehow want to make people's lives unpleasant and difficult.

Quote

4. I have never had anyone talk to me before weighting my score (I haven't actually had many) and ask about our methods etc that would have a material affect on any proper weightings.

Maybe that's because you haven't had many. I can't see how to make a weighting based on the play that doesn't involve talking to the players.

The last time I asked for a MI ruling was in a national KO event. The TD went about it as follows:
- asked what the MI was (there was some dispute about exactly what had happened, but for the sake of argument assume that we all agreed there had been MI and what it was)
- asked what the result was
- asked how the play at the table had gone
- asked how the play at the table would have gone differently without the MI. All I could say at the time (without having had a great deal of time to consider it) was that I had taken a line in the trump suit that was obviously wrong with correct information and reasonable with the MI, and this had clearly cost me a trump trick, so I assumed I might well finish up a trick better.

At this point I was expecting a weighted ruling.

- the TD consulted a colleague, and thought about it.
- Eventually he came back and said that he and his colleague had spent a long time going through the play, and assuming I then got trumps right and the defence misdefended in the way they had at the table, they still couldn't see the source of another trick, because the endplay that had then developed now wouldn't happen as I now had one fewer trump loser to endplay anyone with. However, if I could suggest an alternative line they woud consider how likely it was to be taken at the table.

- When I actually had a chance to think about the alternative line in detail I couldn't see a sensible route to another trick.

OK, in the end this wasn't a weighted ruling, but it's the way they are usually approached.

And just to make the point to some of the paranoid people around:
- The TD who ruled (with agreement from the opponents) was the TD of the club which my team were representing in this event. He also happens to be a very well qualified EBU national TD.
- My team includes the chairman and one other member of the EBU laws & ethics committee. The other TD consulted is another member of this committee. None of the opposing team has anything to do with ebu committees etc.

[Cascade]

jeremy69, on Aug 21 2009, 08:57 PM, said:

Quote

Another way for the system to bully the players.

If the weightings are wrong then the players have a right to get them put right.

But instead we have a regulation that says effectively that the TD is GOD.

This sort of attitude makes me feel sick.

The system isn't bullying anyone. Weightings are an approximation hence the EBU guideline to appeal committees that minor tinkering is best not done. If the appeal view that the TD has got it substantially wrong then, of course, it can be changed. Note that it is a guideline so that an appeal committee in a picky mood could make a minor adjustment if it wished.

Who said they were an approximation?

The law says "weighted to reflect the probabilities of a number of potential results".

A reflection is a likeness of the original.

Ok so it is impractical to know the exact probabilities. At least in part that is my point.

Who knows the probability distributions of potential bridge results for a given hand?

I doubt that there is a person on the planet who can realistically calculate these for a random hand. I mean you might find some hands where at least superficially it is easy to calculate the probabilities.

The potential results depend on: the technical skill of the players; their optimism or pessimism; their likelihood of making a mistake; the consistency of their style etc. The final result depends upon multiple decisions in the bidding and play the precise or even approximate result of which is far from clear.

I have done many many bridge simulations. Usually of situations that I have encountered at the table or read about. Many times I have formulated an opinion about a hand only to find after single dummy or double dummy simulations that my estimation is out by 10s of percentage points. This isn't exactly the same problem but it illustrates how easy it is to be not even in the right ball park with these complex estimations.

Some years ago I was involved on an appeal committee that had to determine the likely result after some misinformation. We decided to adjust the board to 3NT. There there was a problem of determining the result. There was a choice of opening leads which led to different likely results. As far as I know there is no recognized way to determine the probabilities that a particular player will lead one suit or another. I asked for the traveller to look at the various leads that had been recorded. There was a significant minority who had made the less successful lead. After seeing that I had no trouble arguing for us to assume that lead when the criteria were the most favourable result that was likely. How do you start when you must determine some probabilities - because for example 7/17 other players made a particular lead does not necessarily mean that this particular player is going to make that lead with that frequency.

Sure no one appeals these things. That doesn't mean that players are happy with the ruling. An appeal is necessarily less likely when the issue is only a small number of IMPs than when it is several IMPs.

Yes I think it is bullying when the process denies non-offenders justice and then threatens them with penalties if they try to dispute the ruling.

Matches are won or lost by 1 IMP or tournaments won or lost by 1 MP. If the dispute is enough to win or lose a tournament it is not frivolous even if it involves adjusting from 50% to 51% or less.