[Zelandakh]

HighLow21, on 2013-November-12, 11:00, said:

It depends on how you define mistake, doesn't it? Any time a pair fails to get its optimal result on the board, it could be argued that it's a mistake. Many such mistakes will be made routinely by even the best players in the world, but if you define it that way, swing = mistake.

OK, take a classic. Slam is 50%. One pair bids it and the other stays in game. A swing is guaranteed. Which pair has made the mistake? If it depends on the cards of the opponents then you are on fairly slippery ice with your definition.

[HighLow21]

Zelandakh, on 2013-November-12, 11:13, said:

OK, take a classic. Slam is 50%. One pair bids it and the other stays in game. A swing is guaranteed. Which pair has made the mistake? If it depends on the cards of the opponents then you are on fairly slippery ice with your definition.

I see your point, and I think you see mine. One is a mistake regardless of the result, and the other is a mistake given the result. The argument is starting to become semantic, and remember -- the original question revolved around empirical observations of IMP score variation per hand (or per set).

My point is that any factor which leads to more swings--different bidding systems in play, different levels of ability among the pairs, trickier hands, etc.--will obviously correlate to a higher standard deviation of IMP results per hand. Whether you define mistake as "egregious error" or "deviation from par result", at least a large part of the variation will be due to mistakes. And more mistakes will be made when you're playing against a bidding system you aren't familiar with or don't have an effective defense against.

[Trinidad]

awm, on 2013-November-12, 09:58, said:

It may be important to keep in mind that IMP expectations are not transitive (i.e. if pair A is +1 IMP/bd against B and B is +1 IMP/bd against C it does not follow that A is +2 IMP/bd against C). In fact due to methods and style you cannot cleanly rank pairs like this. Further, it is difficult to compare scores against a field when the field is very far off from the pairs being compared.

That is true.

However, within a field you can rank them and assign a strength: It would be the result if everybody would play the same, infinite amount of boards against everybody.

When you then take the top 10 pairs of that field (or any other subset) and do the same exercise on that smaller field, it is entirely possible that their ranking will be different from the original one.

But that doesn't mean that you can't rank within a given field.

Rik

[barmar]

Zelandakh, on 2013-November-12, 11:13, said:

OK, take a classic. Slam is 50%. One pair bids it and the other stays in game. A swing is guaranteed. Which pair has made the mistake? If it depends on the cards of the opponents then you are on fairly slippery ice with your definition.

Here's a different way to think about it.

At IMPs, you're supposed to bid 50% games. If game is 50%, but you don't bid it, you've probably made a mistake (perhaps your bidding methods are poor and don't allow you to determine the odds, or you had the information but used poor judgement in applying it). About half the time you won't be punished, because of the lie of the cards, but if you bid like this in general, your long term results will suffer because the scoring table awards bidding the games more than staying out of them.

The converse of this is that a bidding misunderstanding lands you in a horrible contract, but you find a lucky lie of the cards, or the opponents misdefend, so you make it. Regardless of the results, you still made a mistake during the bidding, although you may have compensated for this with your strong declarer play, or demeanor that didn't give away the show to the defenders (perhaps that's why they misdefended).

[Zelandakh]

In truth this is not a good example for this thread for a completely different reason, namely that the incidence of such hands is constant in the long run and can therefore be modelled. The times when a small system difference leads to a different end contract without any mistake having been made are a better example because that depends purely on how many pairs are playing alternate systems. What I am trying to show is that making a decision that does not lead to the "optimsl" contract does not mean a mistake has been made. Often enough better bidding is rewarded with a negative swing because of the distribution od defenders' cards. That is not a mistake, although it is often seen as such. Often enough that there is a word for it on these forums - resulting. Trying to give resulting a statistical basis and thereby some quasi-mathematical endorsement is something I think should be shot down quickly. There are some benefits to comparing long-term results against par but no benefit or justification whatsoever in calling good actions that happened to produce a negative swing mistakes.

[WellSpyder]

HighLow21, on 2013-November-12, 11:03, said:

Good explanation.

Yes indeed! I upvoted trinidad's post for the clarity of the statistical explanation rather than for the light-hearted touch at the end - though I enjoyed that, too.

[Fluffy]

My friends averaged +3.9 IMps/board between the 2 tables at a tournament this weekend, they scored more than 160 VPs from a maximum of 180 in the 20 VP escale 9 rounds (there was a bye, so actually the maximum was less)

http://www.aebridge....dos/buttler.pdf

The average hand was to win 4 IMPs or so lol.

[Zelandakh]

Sign them up to your team for a run at the next Bermuda Bowl, Fluffy!