Nole's draws

You are just itching to go in that direction, are you not?

Well, what I am saying is kind of straightforward, no? Say Nole plays 40 tournaments. If he got the less favorable draw in 20 of them, that would not be interesting. If he got the less favorable draw in all 40, that would be for all practical purposes proof enough. But if he say got the less favorable draw in 32 out of them, that would be "interesting" - strange enough to make one wonder, but not strange enough to prove enything beyond the shadow of the doubt. And that is sort of where I think we are.

you mean where there's smoke there's fire...

Summerblues, you clearly don't know much about the laws of statistics, do you? I caould come up with at least 5 mainstreaim statistical tests to show you that getting 32/40 bad draws is highly statistically significant!.

Are there really such naive people around who still believe that draws are always made randomly (irrespective of Djokovic)? Do you people realise that tournament organisers have one priority only, and it is called profit? They always have a particular final in mind that they think will make them the most money and in the last 10 years, that final was Nadal-Federer, irrespective of what their rankings.

Thank you bladegunner !

bladegunner wrote:

I caould come up with at least 5 mainstreaim statistical tests to show you that getting 32/40 bad draws is highly statistically significant!.
----------------
Oh yes, it is highly statistically significant all right. But not enough to "prove anything beyond the shadow of the doubt" as I said above.

The likelihood of 32/40 is about "1-in-eleven thousand", which, for a normal distribution, would be roughly 3.7 standard deviations. That is surely interesting but just as surely it constitutes no proof.

"But not enough to "prove anything beyond the shadow of the doubt" as I said above."

and vice versa

Oh yes, it is highly statistically significant all right. But not enough to "prove anything beyond the shadow of the doubt" as I said above.
---------------------------------------------
It's not the stats that prove anything beyond the shadow of doubt....it's knowing human nature.

As Bladegunner say, It's knowing what the organisers want that closes the deal for me.

nitb wrote:

and vice versa
---------------------
Mhmm...if I understood what you mean here I might have an opinion.

Tenez wrote:

It's not the stats that prove anything beyond the shadow of doubt....it's knowing human nature.

As Bladegunner say, It's knowing what the organisers want that closes the deal for me.
------------------------------
This one is tricky.

On one hand the approach is reasonable. The stats themselves do not really address the question "what is the probability of draw being rigged". Stats answer the question "if draws were not rigged, what would be the probability of the outcome we got". There is no direct route from the latter to the former. However, it is the answer to the former question that we are interested in. In order to assess the probability of draws being rigged, given the data we have, we still need to have some concep of "a priory" probabilities. The way it sort of works then is that the higher the "a priory likelihood of draws being rigged" the less evidence one needs to feel they are really rigged. In this sense what you are saying is not unreasonable.

The flip side is that the "a priory likelihood" or whatever we call it is not very quantifiable so it introduces subjectivity to the exercise.

On balance, while I can see where you are coming from, I would say in practice we do not have enough evidence to make a conclusion "beyond reasonable doubt".

"Mhmm...if I understood what you mean here I might have an opinion."

I know you do, and no offence, but I think you think it's superior.

Guess what, so do I about mine