Draw Fixing: An Official Study

The link below shows a talk by an Estionian researcher Katarina Pijetlovic, giving an exposition on draw fixing at a Corruption in Sport Symposium in Koln.


Katarina's talk starts at around 13 minutes:

http://www.livestream.com/playthegame_dshs/video?clipId=pla_44809e94-aa04-46c7-9f1e-35b212ba9d46


She examines the pattern of draws at slam tournaments (French Open was not part of the study) between 2007-2011, drawing the conclusion that ITF organised draw fixing on behalf of Nike seeing that Djokovic fell in Federer's half of the draw statistically virtually impossible 12 out of 12 times.

Roland Garros was not taken into the study as it showed a healthy 50/50 pattern.



Interesting facts, e.g. I didn't know that seeds 3 and 4 are drawn by hand unlike all the other seeds/players that are computer drawn.
Draws are apparently public and televised, but not really accessible anywhere on Youtube.

To me, the most blatant example of draw fixing was the Isner Mahut match played in the first round last year ON COURT 18, just like at the record breaking match the year before!!!


Katarina did the research hoping it would interest sports journalists and encourage them to contact the players and ITF.


So far nothing came out of it.

Have a look with an open mind and share your thoughts.

An amazing find NITB!Wel done! I always thought that draws could be rigged but one can never be sure as there is the possibility that though it's very unlikely that Djoko could constantly fall on Fed's side it coudl happen but the ESPN finds are simply sealing the case.

And Monte Carlo is simply pushing it even further. I was surprised they released the draw that late but clearly they were waiting for qualifiers to get through so they could place the weakest ones on the top seeds side.

Funny - I am called ridiculous and a conspirationist earlier today by Lydian and there again I am proven right. Lydian has no luck! However, it says again how influencial the sponsors are (as I believe the sponsor case this student chose is a good one) and that's something else I said for years. It's those same sponsors that pressurise indirectly the referees when needed to apply the rules. A bit what Tsonga was saying the other day.

I must say I often thought the draws were weird, but never realised that Novak fell into Fed's half on fast courts 12 out of 12 times!

I wonder what the slam distribution would've been like had the draws had healthier statistics like at FO.

Like you I was very cautious and often noticed the trend but could not quite confirm it for sure.

However the USO is probably the worse at that and in particular in 2010 as they wanted Nadal to win the USO desperately. It goes very much in hand with the slowing of the courts, bigger balls, providing Nadal with the best Saturday slot in 2010, etc...

There is a 50% probability of 1-3, 2-4 vs 1-4, 2-3. It does not imply that out of 12 times Federer-Djokovic played, it should have been Federer-Murray instead.

laverfan wrote:There is a 50% probability of 1-3, 2-4 vs 1-4, 2-3. It does not imply that out of 12 times Federer-Djokovic played, it should have been Federer-Murray instead.

yes, and it worked like that at the French ONLY. the other were 12 out of 12. No exception!

Tenez wrote:Like you I was very cautious and often noticed the trend but could not quite confirm it for sure.

However the USO is probably the worse at that and in particular in 2010 as they wanted Nadal to win the USO desperately. It goes very much in hand with the slowing of the courts, bigger balls, providing Nadal with the best Saturday slot in 2010, etc...

Remember Nole's (the defending champion) scheduling at AO 2009?

The ESPN study has nothing to do with the randomness of the 1-2-3-4.

The ESPN study talks about R1/R2 draws and being 'easy' on the top 4.

laverfan wrote:The ESPN study has nothing to do with the randomness of the 1-2-3-4.

The ESPN study talks about R1/R2 draws and being 'easy' on the top 4.

anyway, what are your thoughts on the 12/12 randomness?

noleisthebest wrote:

Tenez wrote:Like you I was very cautious and often noticed the trend but could not quite confirm it for sure.

However the USO is probably the worse at that and in particular in 2010 as they wanted Nadal to win the USO desperately. It goes very much in hand with the slowing of the courts, bigger balls, providing Nadal with the best Saturday slot in 2010, etc...

Remember Nole's (the defending champion) scheduling at AO 2009?

Yes...I keep reminding you!

But that was clearly an admitted request by Roddick's side.

It really puts in perspective the validity of some slams' results. And it could explain such oddities as Davydenko having never met Nadal in a slam....even though some of it is Davydenko not going far enough. Davydenko still played 5 times Federer at slams including when Davydenko was highly ranked.

laverfan wrote:The ESPN study has nothing to do with the randomness of the 1-2-3-4.

The ESPN study talks about R1/R2 draws and being 'easy' on the top 4.

Yes I know. What it says is that it looks clearly rigged ...confirming that that they coudl also rig the top 4.

Tenez wrote:

noleisthebest wrote:

Tenez wrote:Like you I was very cautious and often noticed the trend but could not quite confirm it for sure.

However the USO is probably the worse at that and in particular in 2010 as they wanted Nadal to win the USO desperately. It goes very much in hand with the slowing of the courts, bigger balls, providing Nadal with the best Saturday slot in 2010, etc...

Remember Nole's (the defending champion) scheduling at AO 2009?

Yes...I keep reminding you!

But that was clearly an admitted request by Roddick's side.

It really puts in perspective the validity of some slams' results. And it could explain such oddities as Davydenko having never met Nadal in a slam....even though some of it is Davydenko not going far enough. Davydenko still played 5 times Federer at slams including when Davydenko was highly ranked.

Yes and Serena Williams coming back from the brink of defeat on the same day as Nole retired against Roddick (I think she was playing Kuznetsova) the play was stopped and the roof pulled closed for her, so she went on to win the match and the tournament.
I didn't even know they had a roof before that!

noleisthebest wrote:

laverfan wrote:There is a 50% probability of 1-3, 2-4 vs 1-4, 2-3. It does not imply that out of 12 times Federer-Djokovic played, it should have been Federer-Murray instead.

yes, and it worked like that at the French ONLY. the other were 12 out of 12. No exception!

The sample set is only 12 tosses of coins, so to speak. In 12 tosses, despite a 50% probability of a Heads or a Tails for each toss, you can get 12 Heads in a row, or 12 Tails in a row.

BTW, 2007-2011 is 20 slams (7,8,9,10,11) - 4 W and 4 FO. Why were the 4 Ws not considered in this discussion?

Projected Semi-finals (not actual)....

W 2007 - #1 vs #3 (Federer vs Roddick), #2 vs #4 (Nadal vs Djokovic)
W 2008 - #1 vs #3 (Federer vs Djokovic), #2 vs #4 (Nadal vs Davydenko)
W 2009 - #1 vs #3 (Nadal vs Murray), #2 vs #4 (Federer vs Djokovic)
W 2010 - #1 vs #3 (Federer vs Djokovic), #2 vs #4 (Nadal vs Murray)
W 2011 - #1 vs #4 (Nadal vs Murray), #2 vs #4 (Federer vs Djokovic)

So at W 2007-2011, we have #1 vs #3 (4 instances) and #1 vs #4 (1 instance).

Federer vs Djokovic - 4 instances (outside of HC slams) in 5 non-HC slams, which is 80% (=4/5).

noleisthebest wrote:Roland Garros was not taken into the study as it showed a healthy 50/50 pattern.

Neither was W 2007-2011.

Tenez wrote:

laverfan wrote:The ESPN study has nothing to do with the randomness of the 1-2-3-4.

The ESPN study talks about R1/R2 draws and being 'easy' on the top 4.

Yes I know. What it says is that it looks clearly rigged ...confirming that that they coudl also rig the top 4.

Poll rigging in Russia has nothing to do with slam rigging, does it?

laverfan wrote:
W 2007 - #1 vs #3 (Federer vs Roddick), #2 vs #4 (Nadal vs Djokovic)
W 2008 - #1 vs #3 (Federer vs Djokovic), #2 vs #4 (Nadal vs Davydenko)
W 2009 - #1 vs #3 (Nadal vs Murray), #2 vs #4 (Federer vs Djokovic)
W 2010 - #1 vs #3 (Federer vs Djokovic), #2 vs #4 (Nadal vs Murray)
W 2011 - #1 vs #4 (Nadal vs Murray), #2 vs #4 (Federer vs Djokovic)

LF - She is not talking about 13 24 v 14 23. In fact she said that what is really strange is that despite the ranking changing, Djoko still ended up in Fed's side. 12 in a row is really strange. A chance in 4096 or 0.02% whereas it shoudl tend typically tend to 50%.

But do people remember what Nadal said way before those last 7 encounters? "Djokovic is the player I fear most!" (another bullet for us conspirationists! ). Cause clearly, chance could have made Djoko fall on Nadal's side 12 times in a row. How strange that odditity happened the other way around!

As I said many a time, nothing is left to chance for the player who doesn't leave to chance the way his bottles line up under his chair!

The ITF looks like a band of ignorants who are wearing NIKE fashion clothing. The drawing the 3rd an 4th seeded by hand in GS is definitely worse than polls in Russia. That is mockery

Tenez wrote:It really puts in perspective the validity of some slams' results. And it could explain such oddities as Davydenko having never met Nadal in a slam....even though some of it is Davydenko not going far enough. Davydenko still played 5 times Federer at slams including when Davydenko was highly ranked.

You really love to clutch that particular straw! Any odder than Federer never facing Ferrer in a slam?

It seems a bit odd to remove from the study a tournament where the data does not suit. That's precisely what you're not supposed to do with data.

Why would this be a Nike conspiracy too? Federer would surely be unhappy with such an agreement. Unless Nadal made him agree to it...

It should also be borne in mind that Federer was numero uno for some of this period, certainly in 2007 and 2008 until Wimbledon so to suggest the number two would somehow get preferential treatment seems a tad far-fetched.

Tenez wrote:In fact she said that what is really strange is that despite the ranking changing, Djoko still ended up in Fed's side. 12 in a row is really strange. A chance in 4096 or 0.02% whereas it shoudl tend typically tend to 50%.

Each draw is an independent event, not a cumulative event. This 0.02% chance is being calculated 1/((2)**12). That is not correct. Each slam has a 1/2 chance. Consecutive slams have nothing to do with each other, despite Federer and Djokovic being the two protagonists in discussion.

Tenez wrote:But do people remember what Nadal said way before those last 7 encounters? "Djokovic is the player I fear most!" (another bullet for us conspirationists! ). Cause clearly, chance could have made Djoko fall on Nadal's side 12 times in a row. How strange that oddity happened the other way around!

As I said many a time, nothing is left to chance for the player who doesn't leave to chance the way his bottles line up under his chair!

See 'Positively''s post.

Right! I am totally convinced that yet again that sneaky Rafa has been getting his own way. The draws are officially rigged. I have added proof as I have just thrown a coin three times and every time it came up heads...

But what I'm still not sure about is who has lost the most because of Rafa's persuasive ways. Who would be holding the slam titles that Nadal has so unjustly claimed. Federer? Djokovic? Andy quite rightly claimed that he would have 7 or 8 slam titles if only he had taken up boxing (or other players had taken up boxing... ). So how many Slam titles would Federer or Djokovic have if Nadal wasn't allowed to fix the draw?

Positively 4th Street wrote:You really love to clutch that particular straw! Any odder than Federer never facing Ferrer in a slam?

Yes those things can happen and this is why I don't say it;s a certainty...though Federer Ferrer woudl bear no particular interest for the top players, on the contrary, while Davydendo Nadal certainly would.

It seems a bit odd to remove from the study a tournament where the data does not suit. That's precisely what you're not supposed to do with data.

Not as odd as having 12 consecutive outcome in a series.

Why would this be a Nike conspiracy too? Federer would surely be unhappy with such an agreement. Unless Nadal made him agree to it...

Very simple cause Federer was able to deal with Djoko with more ease already then. And that increased the chances of having a Fedal final...what the people wanted and 2 Nike on both sides if the net!

It should also be borne in mind that Federer was numero uno for some of this period, certainly in 2007 and 2008 until Wimbledon so to suggest the number two would somehow get preferential treatment seems a tad far-fetched.

Again, Fed was handling Djoko ok.....up to 2009. For Nadal it was always a harder battle....as soon as Wimby 2006 (or was it 07?).

What is odd is that again, you try to rationalise a 1 in 4000 oddity.

laverfan wrote:
Each draw is an independent event, not a cumulative event. This 0.02% chance is being calculated 1/((2)**12). That is not correct. Each slam has a 1/2 chance. Consecutive slams have nothing to do with each other, despite Federer and Djokovic being the two protagonists in discussion.

I thought you were a stat expert. Clearly you are missing the main part here! 12 out of 12 in a series if we decide to study that period. So one in 4096 is a perfectly calculated one and the same stat found by that clever sudent. You really come with some strange thinking at times!

hawkeye wrote:Right! I am totally convinced that yet again that sneaky Rafa has been getting his own way. The draws are officially rigged. I have added proof as I have just thrown a coin three times and every time it came up heads...

But what I'm still not sure about is who has lost the most because of Rafa's persuasive ways. Who would be holding the slam titles that Nadal has so unjustly claimed. Federer? Djokovic? Andy quite rightly claimed that he would have 7 or 8 slam titles if only he had taken up boxing (or other players had taken up boxing... ). So how many Slam titles would Federer or Djokovic have if Nadal wasn't allowed to fix the draw?

HE - Look at the MC draw again...and throw your coins a few more times. Tell me what you get.

Tenez wrote: You really come with some strange thinking at times!

It's called beating around the bush

noleisthebest wrote:

Tenez wrote: You really come with some strange thinking at times!

It's called beating around the bush

Yep. I am not saying this student's theory is waterproof but it's pretty close to be really. Another reason why there are so many conspirations around.

Tenez wrote:

laverfan wrote:
Each draw is an independent event, not a cumulative event. This 0.02% chance is being calculated 1/((2)**12). That is not correct. Each slam has a 1/2 chance. Consecutive slams have nothing to do with each other, despite Federer and Djokovic being the two protagonists in discussion.

I thought you were a stat expert. Clearly you are missing the main part here! 12 out of 12 in a series if we decide to study that period. So one in 4096 is a perfectly calculated one and the same stat found by that clever sudent. You really come with some strange thinking at times!

You have made a series out of discrete events. I already gave you an example of 12 Heads or 12 Tails in a row. There is nothing strange about it. Selectively discarding W and FO in the series, as Positively pointed out, is where you have already broken what you call a 'series'. It is no longer a continuous series from a statistical perspective, but a conspiracy series.

Managed to dig these out:

http://www.sportandeu.com/2011/10/sporteu-member-presents-incredible-statistics-about-grand-slam-tennis-tournaments/


http://leastthing.blogspot.co.uk/2011/10/sometimes-rare-event-is-just-rare.html

the key paragraph in the first link above:

"Combined with a study conducted by ESPN on the draws of unseeded players at US Open, which resulted in 1 in 250000 probability, it is ca. 32 BILLIONS TO 1. Media that has covered the Sport&EU member’s study thus far include Der Spiegel, Daily Mail, El Pais, Le Monde, and about hundred others. However, so far the organisers of the Grand Slams and the ITF are avoiding official response to the issue. "

NiTB...

Katarina Pijetlovic is a regular Sport&EU member and can be contacted through this link. - First link....

At PTG, I just saw a talk by Katarina Pijetlovic of Tallinn University of Technology, on alleged irregularities in ITF draws for Wimbledon and the US and Australian Opens. Second link...

Same person as the clip, nothing new.

laverfan wrote:

Tenez wrote:

laverfan wrote:
Each draw is an independent event, not a cumulative event. This 0.02% chance is being calculated 1/((2)**12). That is not correct. Each slam has a 1/2 chance. Consecutive slams have nothing to do with each other, despite Federer and Djokovic being the two protagonists in discussion.

I thought you were a stat expert. Clearly you are missing the main part here! 12 out of 12 in a series if we decide to study that period. So one in 4096 is a perfectly calculated one and the same stat found by that clever sudent. You really come with some strange thinking at times!

You have made a series out of discrete events. I already gave you an example of 12 Heads or 12 Tails in a row. There is nothing strange about it. Selectively discarding W and FO in the series, as Positively pointed out, is where you have already broken what you call a 'series'. It is no longer a continuous series from a statistical perspective, but a conspiracy series.

laverfan

That's all very well but how do you explain the fact that I just threw three heads in a row? You can't can you? The only rational explanation is that Rafa had something to do with it. I'm afraid this doesn't bode well for poor Roger and Nole's chances at RG...

What point did you make? 12 heads and 12 tails happen...but one in 4k! Discarding FO draws can be done here if they choose to cause the French have a mind of their own....like starting a slam on Sunday...and maybe there, it could be explained by Nadal being more confident on his best surface didn;t need extra help and therefore if anything (should you follow the logic) it gives more credence to the studiant case.

She is talking for the 2008-2011 period that is 12 slams outside FO and in this period it's 100% Fed/Djoko! I am stunned that you cannot (or refuse to see it)!

you people do know that every single draw, and indeed, sequence of draws, has exactly the same probability of occuring right?

the French was discarded because it didn't suit their theory. Brilliant! I wish I could do that in my PhD. "Sorry but this case doesn't show what I want it to show, so I'll just ignore it". I might have a chance of success then

noleisthebest wrote:the key paragraph in the first link above:

"Combined with a study conducted by ESPN on the draws of unseeded players at US Open, which resulted in 1 in 250000 probability, it is ca. 32 BILLIONS TO 1. Media that has covered the Sport&EU member’s study thus far include Der Spiegel, Daily Mail, El Pais, Le Monde, and about hundred others. However, so far the organisers of the Grand Slams and the ITF are avoiding official response to the issue. "

That's exactly what sealed the theory for me. Cause you can have one in 4000 chance though extremely rare it can happen, but added to the 250k? as you say it's 1 in 32 billion! (though I checked and come up with one in a billion!

Mad for Chelsea wrote:you people do know that every single draw, and indeed, sequence of draws, has exactly the same probability of occuring right?

the French was discarded because it didn't suit their theory. Brilliant! I wish I could do that in my PhD. "Sorry but this case doesn't show what I want it to show, so I'll just ignore it". I might have a chance of success then

Another one who skipped his math lectures!

Tenez wrote:

noleisthebest wrote:the key paragraph in the first link above:

"Combined with a study conducted by ESPN on the draws of unseeded players at US Open, which resulted in 1 in 250000 probability, it is ca. 32 BILLIONS TO 1. Media that has covered the Sport&EU member’s study thus far include Der Spiegel, Daily Mail, El Pais, Le Monde, and about hundred others. However, so far the organisers of the Grand Slams and the ITF are avoiding official response to the issue. "

That's exactly what sealed the theory for me. Cause you can have one in 4000 chance though extremely rare it can happen, but added to the 250k? as you say it's 1 in 32 billion! (though I checked and come up with one in a billion!

I'm glad at least one person can see it, it's so obvious, I can't fathom how others can miss it!

I suppose the same people think Sepp Blatter is wonderful!

Tenez wrote:

Mad for Chelsea wrote:you people do know that every single draw, and indeed, sequence of draws, has exactly the same probability of occuring right?

the French was discarded because it didn't suit their theory. Brilliant! I wish I could do that in my PhD. "Sorry but this case doesn't show what I want it to show, so I'll just ignore it". I might have a chance of success then

Another one who skipped his math lectures!

sorry, please explain what I've said which is wrong? take a sequence of coin tosses, say of length 12. The probability of getting (say) HTHHTHTTTHHT is exactly the same as that of getting HHHHHHHHHHHH.

and seriously, publishing results which conveniently ignore the cases that don't suit you? you'd be laughed out of the lab...

noleisthebest wrote:the key paragraph in the first link above:

"Combined with a study conducted by ESPN on the draws of unseeded players at US Open, which resulted in 1 in 250000 probability, it is ca. 32 BILLIONS TO 1. Media that has covered the Sport&EU member’s study thus far include Der Spiegel, Daily Mail, El Pais, Le Monde, and about hundred others. However, so far the organisers of the Grand Slams and the ITF are avoiding official response to the issue. "

don't even understand what this means TBH, once again, all draws are equally (un)likely, and if you combine unlikely events you'll get an even unlikelier event. The problem with the general public is they confuse "unlikely" (the result of for instance, a sequence of coin tosses) with "impossible" (if you pick a random number uniformly between 0 and 1, it will never be equal to 1/2)

anyway here's some more fun:

right, even taking into account the (wrong) "12 in a row" thing, so that's roughly a 1 in 4000 chance. Now consider that there's say about 40 weeks per year when tournaments are actually held. So over the course of three years there would be 120 weeks worth of tournaments.

Thus, if you're looking for a sequence of 12 tournaments in three years, you have a choice of 120!/108!12! which is 10^16 roughly. Therefore, there will be on average 2.5*10^12 sequences of 12 tournaments in a three year cycle which produce a sequence of identical semi-final seeds.

MFC proves draws aren't rigged after all

Tenez wrote:She is talking for the 2008-2011 period that is 12 slams outside FO and in this period it's 100% Fed/Djoko! I am stunned that you cannot (or refuse to see it)!

4 FOs and 4 Ws.

So out of 20 slams from 2007 AO to 2011 USO, 4 FOs and 4Ws discarded (non-HC slams), 12 remaining. 8/20 is 40% of data. 12/20 is 60% of the data.

A theory based on 60% samples is a conspiracy theory. .

Can I discard 5 AOs since they are different surfaces and different sand amounts and just use 7 USOs (guess the ESPN study did that already)?

This is when Professor Mahalanobis would start laughing and say... 'You must not do that, dear child, you just broke the series.' Please read this - http://en.wikipedia.org/wiki/Mahalanobis_distance

Mad for Chelsea wrote:

noleisthebest wrote:the key paragraph in the first link above:

"Combined with a study conducted by ESPN on the draws of unseeded players at US Open, which resulted in 1 in 250000 probability, it is ca. 32 BILLIONS TO 1. Media that has covered the Sport&EU member’s study thus far include Der Spiegel, Daily Mail, El Pais, Le Monde, and about hundred others. However, so far the organisers of the Grand Slams and the ITF are avoiding official response to the issue. "

don't even understand what this means TBH, once again, all draws are equally (un)likely, and if you combine unlikely events you'll get an even unlikelier event. The problem with the general public is they confuse "unlikely" (the result of for instance, a sequence of coin tosses) with "impossible" (if you pick a random number uniformly between 0 and 1, it will never be equal to 1/2)

Read the whole top link a few post above. It should be obvious.

laverfan wrote:

Tenez wrote:She is talking for the 2008-2011 period that is 12 slams outside FO and in this period it's 100% Fed/Djoko! I am stunned that you cannot (or refuse to see it)!

4 FOs and 4 Ws.

So out of 20 slams from 2007 AO to 2011 USO, 4 FOs and 4Ws discarded (non-HC slams), 12 remaining. 8/20 is 40% of data. 12/20 is 60% of the data.

A theory based on 60% samples is a conspiracy theory. .

Can I discard 5 AOs since they are different surfaces and different sand amounts and just use 7 USOs (guess the ESPN study did that already)?

This is when Professor Mahalanobis would start laughing and say... 'You must not do that, dear child, you just broke the series.' Please read this - http://en.wikipedia.org/wiki/Mahalanobis_distance

Don't know why you are saying this again and again but it is not for 2007 to 2011... It is for 2008 to 2011 which is 16 slams not 20.

She is not discarding Wimbledon, just FO. That means she is considering 12 out of 16. On all 12 occasions, irrespective of Djoko's ranking (3 or 4), he ended up in Fed's half. That is the point.

Exactly M4C. If we look at how the draws work:

1. The draws are made on seeds, not names. This is important.
2. #1 is placed at the top, #2 placed at the bottom
3. #3 & #4 placed in a container, one number is withdrawn and put in the top half, the other goes in the bottom half
4. Seeds #5-#8 are placed in a box, first one drawn in the top half, second one drawn in the bottom half, third one drawn in the top half, remaining one in the bottom half.
5. This process is then repeated for seeds 9-16, 17-24, and 25-32
6. The rest of the 96 places are drawn randomly by computer.

As #1 is top, and #2 bottom...only #3 and #4 can interchange. Data shows over the last 14 majors back from end of 2011 that there were 6 occasions of seeds #1 & #3 in the top half and seeds #2 & #4 in bottom half, and 8 occasions of #1 & #4 in the top half and #2 & #3 in the bottom half. Or simpler...#3 is in top half 6 times, #3 in bottom half 8 times...so thats 6 vs 8, which is close to 50% overall...as you'd expect. i.e. the statistics of placement of seeds on either side of the draw matches fairly well with theoretically calculated probability values. Now....if the top #4 seeds had always had the same rankings across those slams and you got 12/12 draws the same then that would be more curious. But that wasnt the case.

Alternatively, if you look at top half/bottom half split for Nole (i.e. number of times that Nole was drawn in the top half vs bottom half) it is 4 out of 11. For this you have to only count tournaments where Nole was seeded #3 or #4 (for seeds #1 & #2, the placement is not random). An occurrence of 4 out of 11 from a small sample, where the theoretical expectation is 50% doesn't sound like conspiracy at all. Its not significant - its not like he was drawn in the top half 11/11.

Finally, out of the top 4 seeds, the only relevant random event applies to the #3 seed...i.e. 50/50 chance as it determines into which half of the draw #3 goes into. #4 is then automatically put in the opposite half to #3 - remembering that seeds #1 & #2 are automatically put at the very top and very bottom of the draw respectively. So....as Federer and Djokovic were in the top 4 for so long, yet kept changing seeding positions without ever having both been in the top 2 or 3-4 at the same time, this explains how they could be drawn in the same half so often.

Either way, there is nothing to support 'rigging'.

spuranik, yes but that's LF and my point. you CAN'T discard a whole bunch of results (in this case the FO) just because they don't suit your theory. What these students have done is basically twisted evidence to suit their theory, which is probably the worst thing you can do in science. If they were my students I'd have given them a swift kick up le derriere and made them write out 100 times "I must not ignore results just because I don't like them".

NITB, thanks, have briefly looked into the article you mentioned, and I am happy to concede there may be something in the US Open continuously handing out easy first round draws to the top seeds (I would have to read through things more carefully to make a judgment one way or another at this point - Tenez, please note that it's over the last ten years, so predates Nadal). In fact the ESPN study is A LOT more thorough and much better conducted than the "12 in 12" nonsense.

Mad for Chelsea wrote:sorry, please explain what I've said which is wrong? take a sequence of coin tosses, say of length 12. The probability of getting (say) HTHHTHTTTHHT is exactly the same as that of getting HHHHHHHHHHHH.

Yes, that is correct...but I am sure you can find out what makes the 12 T exceptional.

and seriously, publishing results which conveniently ignore the cases that don't suit you? you'd be laughed out of the lab...

I think you are the one ignoring the case that 1 in a billion is an oddity which clearly shows you have not quite grasped how weird are those events.

lydian wrote:Exactly M4C. If we look at how the draws work:

1. The draws are made on seeds, not names. This is important.
2. #1 is placed at the top, #2 placed at the bottom
3. #3 & #4 placed in a container, one number is withdrawn and put in the top half, the other goes in the bottom half
4. Seeds #5-#8 are placed in a box, first one drawn in the top half, second one drawn in the bottom half, third one drawn in the top half, remaining one in the bottom half.
5. This process is then repeated for seeds 9-16, 17-24, and 25-32
6. The rest of the 96 places are drawn randomly by computer.

As #1 is top, and #2 bottom...only #3 and #4 can interchange. Data shows over the last 14 majors back from end of 2011 that there were 6 occasions of seeds #1 & #3 in the top half and seeds #2 & #4 in bottom half, and 8 occasions of #1 & #4 in the top half and #2 & #3 in the bottom half. Or simpler...#3 is in top half 6 times, #3 in bottom half 8 times...so thats 6 vs 8, which is close to 50% overall...as you'd expect. i.e. the statistics of placement of seeds on either side of the draw matches fairly well with theoretically calculated probability values. Now....if the top #4 seeds had always had the same rankings across those slams and you got 12/12 draws the same then that would be more curious. But that wasnt the case.

Alternatively, if you look at top half/bottom half split for Nole (i.e. number of times that Nole was drawn in the top half vs bottom half) it is 4 out of 11. For this you have to only count tournaments where Nole was seeded #3 or #4 (for seeds #1 & #2, the placement is not random). An occurrence of 4 out of 11 from a small sample, where the theoretical expectation is 50% doesn't sound like conspiracy at all. Its not significant - its not like he was drawn in the top half 11/11.

Finally, out of the top 4 seeds, the only relevant random event applies to the #3 seed...i.e. 50/50 chance as it determines into which half of the draw #3 goes into. #4 is then automatically put in the opposite half to #3 - remembering that seeds #1 & #2 are automatically put at the very top and very bottom of the draw respectively. So....as Federer and Djokovic were in the top 4 for so long, yet kept changing seeding positions without ever having both been in the top 2 or 3-4 at the same time, this explains how they could be drawn in the same half so often.

Either way, there is nothing to support 'rigging'.

We knew Nadal fans were more likely to be ladies. Now we know they are blonds!

spuranik wrote:

laverfan wrote:

Tenez wrote:She is talking for the 2008-2011 period that is 12 slams outside FO and in this period it's 100% Fed/Djoko! I am stunned that you cannot (or refuse to see it)!

4 FOs and 4 Ws.

So out of 20 slams from 2007 AO to 2011 USO, 4 FOs and 4Ws discarded (non-HC slams), 12 remaining. 8/20 is 40% of data. 12/20 is 60% of the data.

A theory based on 60% samples is a conspiracy theory. .

Can I discard 5 AOs since they are different surfaces and different sand amounts and just use 7 USOs (guess the ESPN study did that already)?

This is when Professor Mahalanobis would start laughing and say... 'You must not do that, dear child, you just broke the series.' Please read this - http://en.wikipedia.org/wiki/Mahalanobis_distance

Don't know why you are saying this again and again but it is not for 2007 to 2011... It is for 2008 to 2011 which is 16 slams not 20.

She is not discarding Wimbledon, just FO. That means she is considering 12 out of 16. On all 12 occasions, irrespective of Djoko's ranking (3 or 4), he ended up in Fed's half. That is the point.

Good luck Spuranik. LF is known to go in circle for days...at least this time she takes position which is good cause usually she argues against herself.

Tenez wrote:

Positively 4th Street wrote:You really love to clutch that particular straw! Any odder than Federer never facing Ferrer in a slam?

Yes those things can happen and this is why I don't say it;s a certainty...though Federer Ferrer woudl bear no particular interest for the top players, on the contrary, while Davydendo Nadal certainly would.

That was just an example off the top of my head to put the Davydenko-Nadal 'oddity' in perspective.

Tenez wrote:

Positively 4th Street wrote:It seems a bit odd to remove from the study a tournament where the data does not suit. That's precisely what you're not supposed to do with data.

Not as odd as having 12 consecutive outcome in a series.

No serious study would do such a thing. Not odd, but malpractice!

Tenez wrote:What is odd is that again, you try to rationalise a 1 in 4000 oddity.

I just raised some, for me, valid points. Unlike you who seems to have closed the case already, perhaps as it echoes your thoughts. I'll take being rational every time.

Like M4C I am more intrigued by the early draw stuff at the US Open. Remember a few threads on this last year and wanted to delve deeper myself but, alas, work prevented me from doing so.

Tenez wrote:

Mad for Chelsea wrote:sorry, please explain what I've said which is wrong? take a sequence of coin tosses, say of length 12. The probability of getting (say) HTHHTHTTTHHT is exactly the same as that of getting HHHHHHHHHHHH.

Yes, that is correct...but I am sure you can find out what makes the 12 T exceptional.

To the normal person's eye, yes. To the mathematician's eye, there should be no difference. These people are Maths students, they should be ashamed of themselves. 7

Tenez wrote:

and seriously, publishing results which conveniently ignore the cases that don't suit you? you'd be laughed out of the lab...

I think you are the one ignoring the case that 1 in a billion is an oddity which clearly shows you have not quite grasped how weird are those events.

1 in a billion? That's the one that came from combining lots of results (not really linked - other than they all had to do with tennis draws) right? I mean, I'm ready to concede the ESPN results which seem to at least stem from some reasonably serious research (though I'd need to read through it in more detail). However, these so-called PhD students' claims are so utterly devoid of any mathematical or scientific rigour it's just not funny. Seriously, they'd be laughed at by a scientific review commitee, it's a wonder they even got their results published IMO.

Tenez, clearly my post went straight over your head. Yes the US draw thing is interesting...but the 12/12 thing is completely discountable. You make the mistake of assuming 6 consequetive occurences where the top 4 seeds and their placement in top/bottom half can vary is the same as 6 same tosses of a coin. It isnt.