[aguahombre]

Ed: as an ACBL director who seems to understand weighting of scores, even though it isn't done here, do you believe it should be?

My concerns (not trolling) are that:

--it would be used as a hedge to sort of adjust, rather than being a reflection of the TD's studied and genuine judgement.
--if we adapted weighting at this late date, there would be at least a year's worth of highly inconsistent rulings.
--perhaps weighting should only be allowed in specific categories of rulings

Your opinion? If you would rather move this question to "changing", or would rather not answer at all, o.k.

[blackshoe]

aguahombre, on 2012-March-01, 20:15, said:

Ed: as an ACBL director who seems to understand weighting of scores, even though it isn't done here, do you believe it should be?

My concerns (not trolling) are that:

--it would be used as a hedge to sort of adjust, rather than being a reflection of the TD's studied and genuine judgement.
--if we adapted weighting at this late date, there would be at least a year's worth of highly inconsistent rulings.
--perhaps weighting should only be allowed in specific categories of rulings

Your opinion? If you would rather move this question to "changing", or would rather not answer at all, o.k.

I think that 12C1c (weighted scores) is a better approach than 12C1e (most favorable result for NOS, least favorable for OS). I think the ACBL has stuck with 12C1e out of reluctance to change more than anything else (my guess, based on impressions only). I think that, given the ACBL's abysmal record wrt club level TD training, at least (I can't speak to tournament level TD training, because I have no idea what there is beyond word of mouth), your first two concerns are very important, particularly at club level, and probably at tournament level as well. I'm not sure how to address your third concern. In what categories should weighting be allowed, and in what not? Why?

[aguahombre]

blackshoe, on 2012-March-02, 06:03, said:

I'm not sure how to address your third concern. In what categories should weighting be allowed, and in what not? Why?

I was thinking that weighing the probabilities of different contracts might be a good thing, but that deciding the actual number of tricks to be taken in a particular contract, when the ruling already is weighted for multiple contracts would get to be too convoluted.

[blackshoe]

And yet you have to decide the number of tricks taken in order to score the board. What do you do, give the NOS the maximum number of possible tricks in each contract? The minimum? Some number in between? If the latter, why that particular number?

I don't think this idea is workable.

[aguahombre]

blackshoe, on 2012-March-02, 13:31, said:

And yet you have to decide the number of tricks taken in order to score the board. What do you do, give the NOS the maximum number of possible tricks in each contract? The minimum? Some number in between? If the latter, why that particular number?

I don't think this idea is workable.

Maybe not, but it is what is done, now. The number of tricks awarded in an adjusted contract is a number of tricks in ACBL, not a figure based on a percentage of the time one or the other number of tricks will be taken.

It just would seem a bit much to figure a percentage for arriving at two or three different contracts, and then another weighten calculation for the result of each contract.

[blackshoe]

aguahombre, on 2012-March-03, 09:29, said:

Maybe not, but it is what is done, now. The number of tricks awarded in an adjusted contract is a number of tricks in ACBL, not a figure based on a percentage of the time one or the other number of tricks will be taken.

It just would seem a bit much to figure a percentage for arriving at two or three different contracts, and then another weighten calculation for the result of each contract.

Come on, Agua. You think I'm not well aware that the ACBL doesn't use weighted scores? That has nothing to do with how weighted scores work. Sure weighting takes a little work. So what? It's not about making life easy for lazy directors.

[aguahombre]

blackshoe, on 2012-March-03, 10:53, said:

Come on, Agua. You think I'm not well aware that the ACBL doesn't use weighted scores? That has nothing to do with how weighted scores work. Sure weighting takes a little work. So what? It's not about making life easy for lazy directors.

I understand, and it will probably never happen anyway. Was just thinking it would be too much, too fast, for the vast majority of ACBL directors---not really a matter of laziness, more a matter of frying their brains .

[blackshoe]

Oh, I don't know. I think there may be a minority, perhaps even a large one, with that problem, especially at club level, but I don't think there's a "vast majority".

[Vampyr]

blackshoe, on 2012-March-03, 10:53, said:

Come on, Agua. You think I'm not well aware that the ACBL doesn't use weighted scores? That has nothing to do with how weighted scores work. Sure weighting takes a little work. So what? It's not about making life easy for lazy directors.

It's not very difficult if you have a scoring program that allows you to just enter each score and the percentage assigned to it.

[blackshoe]

I wasn't referring to the math, but to the decision how much weight to give the possibilities. Not to mention figuring out what the possibilities are.

[Vampyr]

blackshoe, on 2012-March-12, 14:21, said:

I wasn't referring to the math, but to the decision how much weight to give the possibilities. Not to mention figuring out what the possibilities are.

Right. Yes that's harder! Best to get a consultant (or several) to perform this task. Which is usually what happens.

[bluejak]

aguahombre, on 2012-March-02, 08:34, said:

I was thinking that weighing the probabilities of different contracts might be a good thing, but that deciding the actual number of tricks to be taken in a particular contract, when the ruling already is weighted for multiple contracts would get to be too convoluted.

I don't see why. My guess is that you have failed to realise how much easier adjustments are with weighted scores. Why? Because the absolute accuracy which is unobtainable is not needed.

You get a situation where you have MI, and a pair defend 4♥ doubled as a result. They claim they would "obviously" bid and make 6♠, which is on a guess. The values are pretty meager for getting to slam.

You decide they would certainly play the hand, making 11 or 12 tricks, and might bid the slam but you are doubtful. You get a few views on the bidding and the play, and come to the conclusion that the players would have bid the slam about one time in three - say 30% - and made twelve tricks about 60% of the time. So the results seem to be:

6♠ = 18%
4♠+2 42%
4♠+1 28%
6♠-1 12%

Using sympathetic weighting, assuming the score in the other room was 4♠ +2, NS +480, you give the players

. 25% of 6♠ =, NS +980, .25 x 11 imps = 2.75 imps
+ 45% of 4♠+2, NS +480, .45 x 0 imps = 0 imps
+ 20% of 4♠+1, NS +450, .20 x -1 imps = -.2 imps
+ 10% of 6♠-1, NS -50, .1 x -11 imps = -1.1 imps

Total 1.45 imps, rounded to 1 imp.

Now you find people do not agree with your weighting, and you start to worry. People think it would be bid 40% of the time, and made 65%. So it could be more like

6♠ = 26%
4♠+2 39%
4♠+1 21%
6♠-1 14%

Using sympathetic weighting, assuming the score in the other room was 4♠ +2, NS +480, you give the players

. 30% of 6♠ =, NS +980, .30 x 11 imps = 3.3 imps
+ 40% of 4♠+2, NS +480, .40 x 0 imps = 0 imps
+ 20% of 4♠+1, NS +450, .20 x -1 imps = -.2 imps
+ 10% of 6♠-1, NS -50, .1 x -11 imps = -1.1 imps

Total 2.00 imps, rounded to 2 imp.

That made a huge difference, didn't it? The point is that over-worrying about the absolute accuray of the weightings is pointless - you are talking very little differences.

Now we cross the Atlantic.

6♠ = 18%
4♠+2 42%
4♠+1 28%
6♠-1 12%

At all likely? 18%? No, not enough. So the non-offenders get 0 imps, since they get 480.

How about this?

6♠ = 26%
4♠+2 39%
4♠+1 21%
6♠-1 14%

26%? that's ok, so now the non-offenders get 11 imps.

The point is that ACBL TDs need to be super-accurate in borderline cases when adjusting, because of the enormous swing a small likelihood causes: other TDs have it much easier because weighting does not need to be so accurate.

The sad thing is there are a lot of TDs this side of the Atlantic who do not use weighting not realising it makes life simpler and easier for them!

[aguahombre]

Yep, sounds easy to me.

[bluejak]

Of course, there is practice. But when someone shows me a UI problem, I think "Well, disallow 3♥, what's left? Play 2♠, how many tricks? No idea? 7,8,9? Ok, 30% of 2♠ -1, 30% of 2♠ =, 40% of 2♠ +1. They are non-offenders so 40% of the highest score because of sympathetic weighting." The point is that is easy, and takes no time at all.