Misinformation and damage - but adjusted score?

[blackshoe]

I think those directors are using the EBU White Book about artificial adjusted scores, where it states:

Such a score is average plus (AVE+) if the side is not at fault, average (AVE) if the side is partly at fault and average minus (AVE-) if the side is fully at fault. This usually translates into 60% or +3 IMPs for AVE+, 50 % or 0 IMPs for AVE, 40% or -3 IMPs for AVE- (see §4.1.2 for other forms of scoring).

 

So, directors are indeed unaware that they can award an artificial adjusted score of 90-10. I think we need someone such as RMB1 to advise us whether you are right and the White Book is wrong.

Perhaps the White Book should also reflect that the director might award more than 60% for average plus, or less than 40% for average minus. It might even provide some guidance as to under what circumstances that might be done.

[Trinidad]

I think those directors are using the EBU White Book about artificial adjusted scores, where it states:

Such a score is average plus (AVE+) if the side is not at fault, average (AVE) if the side is partly at fault and average minus (AVE-) if the side is fully at fault. This usually translates into 60% or +3 IMPs for AVE+, 50 % or 0 IMPs for AVE, 40% or -3 IMPs for AVE- (see §4.1.2 for other forms of scoring).

 

So, directors are indeed unaware that they can award an artificial adjusted score of 90-10. I think we need someone such as RMB1 to advise us whether you are right and the White Book is wrong.

The White book is correct... and so am I.

 

You, and many other TDs, are confusing Law 12C1d and Law 12C2.

 

Law 12C1d deals with what to do when it is too complicated to determine an adjusted score (as a bridge result). There is a result at the table (e.g. 2♦-4), but there was an infraction and without it the result would have been different. Unfortunately, it is hard to say precisely what the result would be. So, we can immediately give the score in MPs / IMPs that seems right.

 

Law 12C2 tells what to do when no result could be obtained at the table. This could, e.g., be because one pair has played the board before when they shouldn't have. Then that pair could be at fault, and they will get Ave-, whereas their opponents are not at fault and will get Ave+.

 

Rik

[gnasher]

The White book is correct... and so am I.

 

You, and many other TDs, are confusing Law 12C1d and Law 12C2.

 

Law 12C1d deals with what to do when it is too complicated to determine an adjusted score (as a bridge result). There is a result at the table (e.g. 2♦-4), but there was an infraction and without it the result would have been different. Unfortunately, it is hard to say precisely what the result would be. So, we can immediately give the score in MPs / IMPs that seems right.

 

Law 12C2 tells what to do when no result could be obtained at the table. This could, e.g., be because one pair has played the board before when they shouldn't have. Then that pair could be at fault, and they will get Ave-, whereas their opponents are not at fault and will get Ave+.

12C2 begins "When owing to an irregularity no result can be obtained [and see C1(d)]". Doesn't that mean that this applies both when the board is unplayable and when an adjustment is made under 12C1d?

 

12C2 also says that average plus and minus scores are "at most 40%" and "at least 60%" and "normally plus or minus 3 imps, but this may be varied as Law 86A allows". Law 86A tells us that at "When the Director chooses to award an artificial adjusted score of average plus or average minus in IMP play, that score is plus 3 IMPs or minus 3 IMPs respectively. Subject to approval by the Regulating Authority, this may be varied by the Tournament Organizer." In other words, at matchpoints you can award whatever artificial score seems appropriate, but at IMPs you can't. What on earth were they thinking when they wrote this?

 

12C2 also prevents our awarding a score like 90-10 in practice, because 12C2c says that you can't give a pair more than their average for the session, if that's above 60%.

[gordontd]

12C2 begins "When owing to an irregularity no result can be obtained [and see C1(d)]". Doesn't that mean that this applies both when the board is unplayable and when an adjustment is made under 12C1d?

That would have been my reading of it.

 

12C2 also says that average plus and minus scores are "at most 40%" and "at least 60%" and "normally plus or minus 3 imps, but this may be varied as Law 86A allows". Law 86A tells us that at "When the Director chooses to award an artificial adjusted score of average plus or average minus in IMP play, that score is plus 3 IMPs or minus 3 IMPs respectively. Subject to approval by the Regulating Authority, this may be varied by the Tournament Organizer." In other words, at matchpoints you can award whatever artificial score seems appropriate, but at IMPs you can't. What on earth were they thinking when they wrote this?

I think the "at most" and "at least" phrases are there to allow for what we are told to do when their session score is above 60% or below 40%. The flexibility offered with regard to IMP scores is often used for IMP-pairs where +-2 IMPs are commonly used (and this is usually increased for pairs whose session average is greater than 2IMP/board).

 

There's also a further possiblity offered by L86D.

[Trinidad]

12C2 begins "When owing to an irregularity no result can be obtained [and see C1(d)]". Doesn't that mean that this applies both when the board is unplayable and when an adjustment is made under 12C1d?

No.

 

Think of this particular case: A result was obtained (2♥-1). The reference in 12C2 to Law C1(d) is there for those (rare) cases that 12C1d applies (a result was obtained) and it is absolutely, completely impossible to say what the MP/IMP score would be without the infraction.

 

In this case, one could say that it is impossible to determine one accurate bridge score (2♦-4, 3♦-5, 4♦X-6, something else, varying from +100 NS to -1400 NS), but it is most likely not difficult at all to assign an accurate score in MPs: No matter what the bridge score would be NS would get a (near) bottom on the board (varying between 0 en 10% NS).

 

12C1d is there to prevent the need for arbitrary constructions of weighted scores. We could argue for days about the exact weighting of the set of possible results [2♦-4, 3♦-5, 4♦X-6, something else]. Law 12C1d gives us the power to say, e.g.:

90% will lead to NS going down in too high a diamond contract for a bottom for NS.

10% will lead to 'something else' -we don't know what, and then average would be the best guess.

90%x0% + 10%x50% = 5% for NS and 95% for EW."

 

This is entirely different from a situation where NS commit an infraction, get a complete top because of that, and we don't have a clue what would have happened without the infraction, neither in bridge score, nor in MP score. In that case, you give the offenders Ave- and the non-offender Ave+.

 

(Say, South opens 2♦ (weak, but explained as Benji Acol). As a result NS make a diamond partscore for +110, but EW can make game in hearts, spades and NT or slam in hearts (but not in spades) and NS can sacrifice in diamonds. All of these results -varying from +100 to -1430 in bridge scores for NS and in MPs from 5 to 95% for NS- would be possible outcomes.)

 

Rik

[blackshoe]

No. If Rik can say that, so can I.

 

The purpose of 12C1{d} is to allow the TD to award an artificial adjusted score when awarding the normally required (because a result was obtained at the table) assigned adjusted score is too difficult because "the possibilities are too numerous or not obvious". How do we award an artificial adjusted score? We look to Law 12C2. Now the first time I read that law, I interpreted "at most 40%" and "at least 60%" as giving the director discretion to vary from "60/40" if in his judgment that is appropriate. After a lot of back and forth with various people I reluctantly accepted the alternate interpretation that the "discretion" doesn't exist unless the pertinent pair was already having a game greater than 60% or less than 40% - but I've never really believed that's what the law says, and I still don't. In fact these discussions about Law 12C1{d} have simply confirmed, in my mind, my original interpretation, because it is in 12C1{d} cases that the discretion is most likely to be apply, and there is nothing in the law itself that suggests that the TD does not have that discretion under Law 12C2.

 

There is another interesting aspect of Law 12C2: if both pairs are having, say, a 65% game, you cannot give the OS less than 40%, except in the ACBL, where the NOS will get their alloted 65%, and the NOS will get 35%. IOW, the ACBL has mandated that the scores must balance. That means that, in the ACBL, if you wish to give the NOS 90%, you must give the OS 10%. You have no choice. If you don't think this is reasonable for the OS, you'll give a different adjustment. In the rest of the world, of course, directors don't have this problem.

[Vampyr]

There is another interesting aspect of Law 12C2: if both pairs are having, say, a 65% game, you cannot give the OS less than 40%, except in the ACBL, where the NOS will get their alloted 65%, and the NOS will get 35%. IOW, the ACBL has mandated that the scores must balance. That means that, in the ACBL, if you wish to give the NOS 90%, you must give the OS 10%. You have no choice. If you don't think this is reasonable for the OS, you'll give a different adjustment. In the rest of the world, of course, directors don't have this problem.

 

How do the ACBL score, say, A/A+?

[blackshoe]

Heh. The restriction (that the scores must balance) applies only to A+/A-. Other than that, I would expect we would give 50%/60% (or more than 60% if that's appropriate. IAC, the computer handles it, so i doubt most directors have a clue, or that they even think about it.

 

I have to admit I was tempted to say "they don't". B-)

[axman]

No. If Rik can say that, so can I.

 

The purpose of 12C1{d} is to allow the TD to award an artificial adjusted score when awarding the normally required (because a result was obtained at the table) assigned adjusted score is too difficult because "the possibilities are too numerous or not obvious". How do we award an artificial adjusted score? We look to Law 12C2. Now the first time I read that law, I interpreted "at most 40%" and "at least 60%" as giving the director discretion to vary from "60/40" if in his judgment that is appropriate. After a lot of back and forth with various people I reluctantly accepted the alternate interpretation that the "discretion" doesn't exist unless the pertinent pair was already having a game greater than 60% or less than 40% - but I've never really believed that's what the law says, and I still don't. In fact these discussions about Law 12C1{d} have simply confirmed, in my mind, my original interpretation, because it is in 12C1{d} cases that the discretion is most likely to be apply, and there is nothing in the law itself that suggests that the TD does not have that discretion under Law 12C2.

 

 

 

 

I have spent some time parsing L12C2 and am quite fascinated. In part [for example] C2a provides that an art score for the NOS may be [since it must be at least 60%] ANY thing greater than 60% except [C2c] when his session score is more than 60% [in which case his score is EXACTLY his session percentage]. To put a fine point on it, the art score can be 200% or 236% or…. So long as the session score of the other boards is not greater than 60%.

 

I thus see no valid reason to believe the argument that convinced you otherwise.

[sailoranch]

I don't get it. Is there any situation where the possible results are too numerous or not obvious enough to award an assigned score, but obvious enough and not too numerous to conclude that the NOS would matchpoint exactly 90%?

 

Just pick a result or weight several of them and give an assigned adjusted score, which is essentially what's happening in your mind anyway.

[pran]

[...]

After a lot of back and forth with various people I reluctantly accepted the alternate interpretation that the "discretion" doesn't exist unless the pertinent pair was already having a game greater than 60% or less than 40% - but I've never really believed that's what the law says, and I still don't.

[...]

 

Yes, that is exactly what the laws say, and also what is intended!

 

When artificial adjusted scores are required a contestant not at fault shall not loose, and a contestant at fault shall not gain from being awarded AAS instead of real scores.

 

And in such cases loss and gain shall be relative to (based on) their respective averages over the session.

 

Thus, a contestant with an average of 65% over the session shall be guaranteed 65% on any AAS because of an irregularity during that session for which he is not at all at fault.

[pran]

[...]

There is another interesting aspect of Law 12C2: if both pairs are having, say, a 65% game, you cannot give the OS less than 40%, except in the ACBL, where the NOS will get their alloted 65%, and the NOS will get 35%. IOW, the ACBL has mandated that the scores must balance. [...]

I don't care what ACBL may have mandated, but Law 12C2c is clear that (initially) balanced artificial adjusted scores shall be modified for contestants not at fault having a session average higher than 60% and for contestants at fault having a session average lower than 40%.

 

So clearly artificial adjusted scores need not Balance.

[blackshoe]

I have spent some time parsing L12C2 and am quite fascinated. In part [for example] C2a provides that an art score for the NOS may be [since it must be at least 60%] ANY thing greater than 60% except [C2c] when his session score is more than 60% [in which case his score is EXACTLY his session percentage]. To put a fine point on it, the art score can be 200% or 236% or…. So long as the session score of the other boards is not greater than 60%.

 

I thus see no valid reason to believe the argument that convinced you otherwise.

I would never propose to carry adjustments to such ridiculous extremes, and I think reading the law as allowing it is nuts.

[blackshoe]

I don't care what ACBL may have mandated, but Law 12C2c is clear that (initially) balanced artificial adjusted scores shall be modified for contestants not at fault having a session average higher than 60% and for contestants at fault having a session average lower than 40%.

 

So clearly artificial adjusted scores need not Balance.

Yes, that's true - anywhere but in the ACBL. You may not care, since you don't have to rule here, but I do. I'm not going to make an illegal ruling here just because it's legal in the rest of the world.

[axman]

I would never propose to carry adjustments to such ridiculous extremes, and I think reading the law as allowing it is nuts.

 

Applying law as written is not ridiculous; conversely, writing such law is bad form.

[pran]

I don't care what ACBL may have mandated, but Law 12C2c is clear that (initially) balanced artificial adjusted scores shall be modified for contestants not at fault having a session average higher than 60% and for contestants at fault having a session average lower than 40%.

 

So clearly artificial adjusted scores need not Balance.

 

Yes, that's true - anywhere but in the ACBL. You may not care, since you don't have to rule here, but I do. I'm not going to make an illegal ruling here just because it's legal in the rest of the world.

 

I agree with you and appreciate your position.

 

An interesting question (although not really worth discussing) is what power ACBL claims to mandate a procedure that apparently violates Law 12C2c?

 

(The Powers given to Regulating Authorities in Law 12C1 do not extend to Law 12C2!)

[blackshoe]

I agree with you and appreciate your position.

 

An interesting question (although not really worth discussing) is what power ACBL claims to mandate a procedure that apparently violates Law 12C2c?

 

(The Powers given to Regulating Authorities in Law 12C1 do not extend to Law 12C2!)

The ACBL owns the copyright for the "Laws of Duplicate Bridge" in the Western Hemisphere. The organization has always maintained that it can do what it wants with its version of the laws. As far as the ACBL is concerned, the procedure doesn't violate Law 12C2c, it is Law 12C2c - although it isn't quite written that way. The change (that "when there is an offending and a non-offending side" the scores must balance), is in a footnote to 12C2c.

 

As for it not being worth discussing, you're right - certainly nothing we might say here is going to change anything. B-)

[Trinidad]

I don't get it. Is there any situation where the possible results are too numerous or not obvious enough to award an assigned score, but obvious enough and not too numerous to conclude that the NOS would matchpoint exactly 90%?

 

Just pick a result or weight several of them and give an assigned adjusted score, which is essentially what's happening in your mind anyway.

Take this particular case.

 

NS are the offending side. The table result (+100) is already a bad result for NS since the field will have reached a spade contract. Let's put a number on it and call it 30%. This score will only beat the NS pairs who went down in 4♠ when EW found the heart ruff (-100).

 

However, if EW would have gotten the correct information the result would have been even worse for NS. Perhaps the bridge result would have been 2♦-4 (-200), 3♦-5 (-250) or 4♦X-6 (-1400).

 

It is very difficult to say what the final contract and the bridge result would have been. It could have been either of the 3 above. We need to come up with weighting factors that are hard to determine.

 

However, it is easy to see what the MP score will be: All three bridge results will lead to a MP score of 0 for NS since none of them beats the score for 4♠-1.

 

So, in this case, though the possible bridge scores may be numerous, the MP scores aren't.

 

Rik

[Trinidad]

From a mathematical point of view I like the ACBL position of balancing the scores for Ave+ and Ave-.

 

Say that we have a club of average players who, against each other, will score 50%. One club night Helgemo-Helness pay a visit to this club. On the first 21 out of 22 boards they take one trick more than the rest of the field that consists of 6 tables. On each board the MP (in %) are:

- HH: 100%

- the rest: 40%

 

and

 

- HH's opponents: 0%

- the rest's opponents: 60%

 

So, if HH would do as expected on board 22 (take one more trick than the rest of the field) then they would win with a MP score of 100%.

The rest of the field would score 2x0%, 10x 60%, 10x40% for an average of 45.45%

 

However, on board 22, HH's opponents commit an infraction. Unfortunately, we have to assign an Artificial Adjusted Score (AAS):

HH: Ave+

Opponents: Ave-

 

For HH, this is easy: they will get their own 100% and score 100% over the evening.

For their opponents, it depends whether they are in the ACBL (when they get 0) or some where else (when they get 40%).

 

In the ACBL this would lead to a score of 45.45% for all other contestants. But in the rest of the world it would mean the offending pair scores: 10x 60%, 10x40%, 1x0% and 1x 40% for a score of 47.27%, beating all the other pairs!

 

The point is, of course, that if you are an average pair and play against a pair that scores over 60% then your expected score will be (100-the average score of your opponents), which is less than 40%.

 

Rik

[lamford]

WBFLC minutes 2003-11-09#2:

When there has been misinformation and a damaged side is to receive an adjusted score this should be assessed on the basis that the non-offending side is entitled to know the partnership understanding and to draw logical conclusions, given the information it received.

If given the correct information the partnership might or might not be aware that a misunderstanding had occurred, depending on the situation.

This thread has wandered off to discuss when and whether someone can get more than 60% if the possible auctions are too numerous. I would prefer to revisit the interpretation of the above, that one assigns a score assuming that one side only received correct information. If that were the case, why is there "might or might not" in the second paragraph? It seems to me that there is no right to know that a pair is having an misunderstanding, but if one learns that from the AI, this information is used in deciding what someone would bid.

 

This follows natural justice as well. It should not be beneficial for a pair to have a blank convention card. In this example, if NS had a card that showed that 2D was the majors, West looking at his hand and the card, would pretty much know that NS were having a misunderstanding. East would be in the same boat, and would pass it out. If NS had a blank convention card, West might still suspect something, but would probably still pass. East would not suspect anything and bid 2H. Therefore we should adjust in this case on the basis that "might be aware" applies not "might not be aware" applies. I think the WBFLC minute is being misinterpreted if you adjust on the basis that EW only received correct information, not both. And you provide an incentive to have blank convention cards.

 

In this example, when adjusting, if EW get a better score by passing out 2D, using all the AI, then this is what they should be deemed to do.

[Trinidad]

I think the "might" and "might not" be aware of the misunderstanding is very simple:

 

If the NOS learns about the misunderstanding only because of the misinformation, we adjust on the basis that they wouldn't have been aware of the misunderstanding. But if they learn about the misunderstanding from something else, e.g. one of the opponents making an impossible (or highly unlikely) bid, we adjust on the basis that they are aware of the misunderstanding.

 

Rik

[sailoranch]

Take this particular case.

 

NS are the offending side. The table result (+100) is already a bad result for NS since the field will have reached a spade contract. Let's put a number on it and call it 30%. This score will only beat the NS pairs who went down in 4♠ when EW found the heart ruff (-100).

 

However, if EW would have gotten the correct information the result would have been even worse for NS. Perhaps the bridge result would have been 2♦-4 (-200), 3♦-5 (-250) or 4♦X-6 (-1400).

 

It is very difficult to say what the final contract and the bridge result would have been. It could have been either of the 3 above. We need to come up with weighting factors that are hard to determine.

 

However, it is easy to see what the MP score will be: All three bridge results will lead to a MP score of 0 for NS since none of them beats the score for 4♠-1.

 

So, in this case, though the possible bridge scores may be numerous, the MP scores aren't.

 

Rik

 

I don't see why you wouldn't award -1400 in the ACBL or a weighted score elsewhere. Of course the weighting factors are hard to determine because they are always somewhat arbitrary.

 

The MP scores would be super hard to be confident about if the board is yet to be played at another table. On the other hand, if the event is over and the weights don't affect the MP score at all, then why the need to resort to an artificial score at all when you could just commit to some weights and be done with it.

[sailoranch]

This thread has wandered off to discuss when and whether someone can get more than 60% if the possible auctions are too numerous. I would prefer to revisit the interpretation of the above, that one assigns a score assuming that one side only received correct information. If that were the case, why is there "might or might not" in the second paragraph? It seems to me that there is no right to know that a pair is having an misunderstanding, but if one learns that from the AI, this information is used in deciding what someone would bid.

 

This follows natural justice as well. It should not be beneficial for a pair to have a blank convention card. In this example, if NS had a card that showed that 2D was the majors, West looking at his hand and the card, would pretty much know that NS were having a misunderstanding. East would be in the same boat, and would pass it out. If NS had a blank convention card, West might still suspect something, but would probably still pass. East would not suspect anything and bid 2H. Therefore we should adjust in this case on the basis that "might be aware" applies not "might not be aware" applies. I think the WBFLC minute is being misinterpreted if you adjust on the basis that EW only received correct information, not both. And you provide an incentive to have blank convention cards.

 

In this example, when adjusting, if EW get a better score by passing out 2D, using all the AI, then this is what they should be deemed to do.

 

This was discussed extensively in the topic I linked to earlier. I was taking your position, but the accepted view seems to be that the irregularity is the original instance of MI rather than the call based on MI. So in adjusting, the entire auction has to be reconstructed with the correct explanation rather than from a later point more beneficial to the NOS.

[barmar]

Applying law as written is not ridiculous; conversely, writing such law is bad form.

Is it really necessary for the Laws to go into detail, explicitly prohibiting all forms of nonsense? Just because the Law may allow something, it doesn't mean that it's appropriate; it's meant to be applied by people with capacity for reason, and an appreciation for the history of the game and the spirit of the Laws.

[gnasher]

This thread has wandered off to discuss when and whether someone can get more than 60% if the possible auctions are too numerous. I would prefer to revisit the interpretation of the above, that one assigns a score assuming that one side only received correct information. If that were the case, why is there "might or might not" in the second paragraph? It seems to me that there is no right to know that a pair is having an misunderstanding, but if one learns that from the AI, this information is used in deciding what someone would bid.

I think you should read the more recent minute posted by RSliwinski in post no 42. The word "only" in that minute makes the WBFLC's views uncharacteristically clear.