[jdeegan]

Nobody can properly analyze 9000 bridge hands.

Madam, you are a fakir and a fraud. You are an idiot turned loose with a computer and a first year grad student's knowledge of statistics.

You present a double dummy playing analysis based on a potentially very biased sample.

[nige1]

Cthulhu D, on 2014-December-18, 22:55, said:

@nige1: yes, that is precisely my point. That is a weird/interesting artifact of the scoring system. He's overriding his judgement to open 1suit because of the way to the scoring works.

It depends on where you're coming from. You could equally argue that he is using his MP judgement to open 1N. Or you might just conclude that he adapts his strategy to the relevant scoring method and his tactics to the current competitive context.

[nige1]

jdeegan, on 2014-December-19, 03:40, said:

Nobody can properly analyze 9000 bridge hands. Madam, you are a fakir and a fraud. You are an idiot turned loose with a computer and a first year grad student's knowledge of statistics. You present a double dummy playing analysis based on a potentially very biased sample.

I find this topic interesting. Helene_t (who is unlikely to be a Fakir) went to some trouble to investigate it. We can disagree with her methods or conclusions but ad hominem attack sidetracks discussion and wastes our time. Better to present relevant data and constructive argument of our own.

[suokko]

jdeegan, on 2014-December-19, 03:40, said:

Nobody can properly analyze 9000 bridge hands.

Madam, you are a fakir and a fraud. You are an idiot turned loose with a computer and a first year grad student's knowledge of statistics.

You present a double dummy playing analysis based on a potentially very biased sample.

I can do a written analyze at 20 boards per hour speed. If that was my day job it would take only 3 months to analyze 9000 bridge hands.

But actual arguments for practical randomness in IMP vs MP games.

Normal winner score in MP is around 57-65% in competitions where skill level difference aren't big. If a tournament is about 40 boards then single board top-bottom difference is 2.5% in end results. But in practical play my opponents cannot score more than 2/3 of top with a good play. That translates to about 1.6% at stage each board which is 1/9 of score difference between winner and average. But even a good play can be often countered with a less important decision that still score a few points back to side that didn't have the more important decision to make.

At similar IMP competition winner score is +1.5-2.5 IMP per board per comparison. But a single vulnerable game swing is 10-14 IMPs per comparison. That translates to about 0.25-0.35 IMPs per board and those decision are pretty common with only one side having decisions that affect the make or not make situation. So even a normal game swing has a higher magnitude in IMPs than extreme maximum MP games offer. But if we factor in some extreme IMP swings with small slam their effect can be even twice the effect that games have. That makes IMP scoring more volatile because smaller number of important decision provide relatively higher magnitude changes to the results.

[Cthulhu D]

suokko, on 2014-December-19, 16:15, said:

I can do a written analyze at 20 boards per hour speed. If that was my day job it would take only 3 months to analyze 9000 bridge hands.

But actual arguments for practical randomness in IMP vs MP games.

Normal winner score in MP is around 57-65% in competitions where skill level difference aren't big. If a tournament is about 40 boards then single board top-bottom difference is 2.5% in end results. But in practical play my opponents cannot score more than 2/3 of top with a good play. That translates to about 1.6% at stage each board which is 1/9 of score difference between winner and average. But even a good play can be often countered with a less important decision that still score a few points back to side that didn't have the more important decision to make.

At similar IMP competition winner score is +1.5-2.5 IMP per board per comparison. But a single vulnerable game swing is 10-14 IMPs per comparison. That translates to about 0.25-0.35 IMPs per board and those decision are pretty common with only one side having decisions that affect the make or not make situation. So even a normal game swing has a higher magnitude in IMPs than extreme maximum MP games offer. But if we factor in some extreme IMP swings with small slam their effect can be even twice the effect that games have. That makes IMP scoring more volatile because smaller number of important decision provide relatively higher magnitude changes to the results.

I agree re: Variance. Jeff has done some analysis of this here: http://www.jeff-gold...rg/bridge/study

[PhilKing]

nige1, on 2014-December-19, 09:27, said:

I find this topic interesting. Helen_t (who is unlikely to be a Fakir) went to some trouble to investigate it. We can disagree with her methods or conclusions but ad hominem attack sidetracks discussion and wastes our time. Better to present relevant data and constructive argument of our own.

One should only be able to start a thread like if one has a phd in statistics.

Oh, wait ...

[cherdano]

So within a few hours, Helene (who knows more statistics than anyone on this forum), gets
- called "an idiot turned loose with a computer and a first year grad student's knowledge of statistics", and
- gets mansplained (in another thread) the most basic thing one could explain about standard variance.

I am sure one could explain this without referring to sexism, but I doubt it would lead to the easiest explanation.

[kuhchung]

cherdano, on 2014-December-19, 18:10, said:

So within a few hours, Helene (who knows more statistics than anyone on this forum), gets
- called "an idiot turned loose with a computer and a first year grad student's knowledge of statistics", and
- gets mansplained (in another thread) the most basic thing one could explain about standard variance.

I am sure one could explain this without referring to sexism, but I doubt it would lead to the easiest explanation.

ignorance and internet behavior

[PhilKing]

cherdano, on 2014-December-19, 18:10, said:

So within a few hours, Helene (who knows more statistics than anyone on this forum), gets
- called "an idiot turned loose with a computer and a first year grad student's knowledge of statistics", and
- gets mansplained (in another thread) the most basic thing one could explain about standard variance.

I am sure one could explain this without referring to sexism, but I doubt it would lead to the easiest explanation.

By an astounding coincidence, they are the two people on my ignore list - I only saw the posts because their idiocy was quoted for posterity ...

[gwnn]

I'm not sure how "first year grad student's" knowledge of statistics is supposed to be some kind of insult (it kind of is to Helene but I am talking in abstract). That would be probably something in the top percentile of even reasonably educated people and definitely above everyone on TV.

[PhantomSac]

<3 helene, just so you know I found this thread really interesting, didn't have anything to add to it for obvious reasons but would enjoy more posts/threads like this.

[whereagles]

yeah, but statistical services are expensive, ya know kthxbye

[nige1]

I'd be grateful if Helene_t would comment. Is the following a reasonable basis for an alternative approach to estimating the relative skill involved in MPs and cross-imps?

• Work out the average rank (as a percentile) of a selection of top pairs over lots of world-class MP pairs events.
• Also work out a their average percentiles at similar cross-imp pairs events.
• Then see how well these two over-all rankings correlate with their rankings in each individual event. Perhaps, also future events.
• The higher the correlation, the better the measure of skill? Especially if one form of scoring dominates the other as a predictor.

[mgoetze]

nige1, on 2014-December-20, 14:41, said:

Also work out a their average percentiles at similar cross-imp pairs events.

LOL.

[helene_t]

Yes, that sounds like a good plan.

But you need somehow to take into account that the fields could be more homogenous in one type of events than in the other. Suppose, for example, that the skill factor is really the same at both forms of scoring, and the IMP and MP tourneys are similar w.r.t. size and balance. However, IMP tourneys attract fewer weak pairs. So the relative ranking of the top pairs will be more stable at IMPs because there is less randomness associated with sitting in the same seat as a weak pair on a particular board.

So unless you have data where identical fields play some IMP tourneys and also some MP tourneys, the analysis is not trivial.

[nige1]

Clarification for mgoetze: I meant, for example, that if you finish 2nd out of 300 pairs then you are in the top percentile.

Thank you Helene_t.
Perhaps it would be better to use the ranks of the selected pairs relative to each other and ignore all other competitors?

[helene_t]

I think we have too few flame wars about statistics. I think I will start a poll: "Are Bayesian methods better than frequentist ones?"

[whereagles]

anyone can say "this is an absolutely obv 4S bid".

not so with statistics

[mgoetze]

Clarification for nige1: there is no such thing as "lots of world-class cross-IMP pair events".

[nige1]

mgoetze, on 2014-December-20, 17:04, said:

Clarification for nige1: there is no such thing as "lots of world-class cross-IMP pair events".

Fair enough. I thought some Calcutta and WBF events might qualify. Failing that, you'd have to settle for what is available.