Josh wrote:

---

It seems to me that they have made a conclusion then tried to find statistics that support it rather than looking at the issue objectively.

---

That was my conclusion too when I read about this last night, though at first sight it doesn't look like they got these results just by starting from a convenient year, since the year before they started, top seed Agassi drew Alex Kim (ranked 433 at the time of the draw) and 2nd seed Kuerten drew Wayne Arthurs (WR 101), who actually managed to beat him in 4 sets.

Without doing the calculations, the data does look like it would be very statistically significant, so there's no reason to doubt that they crunched the numbers correctly, but as we know from the freakish-looking runs of good and bad luck in Futures draws that we notice from time to time, if you take enough tournaments (and enough decades), you are bound find some odd streaks - for example, I doubt this data is more statistically significant than the number of times Muzz has ended up on the same side of the draw as Rafa in recent slams or the number of times Bally has avoided a seed in R1 of WTA draws when she has been unseeded herself over the last 12 months.

I can't see why they would feel the need to fix it like that anyway - there are very few players they would need the top two seeds to avoid to ensure that there was very little chance of them going out in R1, and drawing a qualifier in R1 (which is what those ranked outside the top 100 are likely to be) is often more dangerous than drawing a direct entrant in the 50-100 range.

Of course, none of these counter-arguments prove that there isn't something wrong (e.g. I wouldn't be shocked if they found a glitch in the software that does their random draw of unseeded players and they ought to check it), but the ESPN analysis is a long way off proving that there definitely is something wrong and even further off proving that the draws are deliberately fixed.

Josh wrote:

---

 

Women's Grand Slam	Percent of draw simula-tions as easy as actual draws
Australian	94.7%
French	99.2%
Wimbledon	30.7%
U.S. Open	0.0%

Surely that means that the Aus Open and Roland Garros are biased in giving the top 2 seeds a harder draw than expected?

 

---

Yes, those are quite strange ones too, since "as easy" presumably in almost all cases means "easier" since hardly any simulations would  come out with exactly the same average difficulty as over the last 10 years.

Clearly you most expect about 50%.  Now over just 10 years it's not too surprising that in actual fact they have tended to be a bit harder or easier on average and gone more to the margins of the similulation outcomes, but I must admit these have gone a bit more to the margins both ways then I would have expected they would  ( eg. 0.0 and 99.2 ! ).

What to make of it all if anything though, I'm not sure...