USO 2015 Final - What do the stats say?

[falstaff78]

Below are stats about how our finalists have done in the US Open, and also on fast hard courts coming in. Main messages are:

Fast hard court form. (see below for list of "fast" courts)

• Federer is undefeated (30-0) and dominant (won 93% of sets) on fast hard courts post USO 2014
• Djokovic has not beaten Federer on a fast hard court in 4 years, and has not taken a set in 3 matches since USO 2014; historically Federer has dominated Djokovic on fast hard courts 13 to 4; (vice versa on slow hard courts 3-10)
• Since USO 2014 Federer has improved immensely on fast courts compared to the period post USO 2013 to USO 2014; whereas Djokovic has not improved; suggesting Fed's recent USO performances may not be a good form guide for tomorrow.

USO 2015 form. (updated from previous thread)

• Both players have been equally dominant in terms of winning sets, games and points
• Djokovic has been stronger on return, and Federer on serve; the two effects have more or less netted out
• However, Federer has played significantly stronger opponents
  
• Federer has played more first strike tennis (much better winners to unforced errors ratio); whereas Djokovic has elicited far more mistakes from his opponents; possibly because he makes them run like crazy on each point
• Federer has not exerted himself much and should be fresh for the final

I've seen enough. I'm calling it:

Federer to win USO 2015 and the Law of 13s to come true!

Please vote in the poll above. The fourth option has a typo. It should read: "I previously thought Federer, and after reading the thread I still think Federer"

(Dark blue means significantly better; light blue means slightly better)

P.S. From relevant tournaments I have considered the following to be "fast hard courts".
2000: USO
1000: Cincy, Shanghai
500: Dubai, Basel
250: Brisbane, Doha

There was some discussion in my other thread on a similiar topic. I had originally included WTF and Paris as "fast," but was persuaded to leave them out. If you include WTF as a fast hard court, the career "fast" H2H goes from 13-4 to 14-6, and "slow" head to head goes from 3-10 to 2-8

 

[Boom-Boom]

Good stuff, thanks for sharing the stats. Paris is definitely a fast hard court though.

 

[jm1980]

Djoko looks happier than Fed here

 

[falstaff78]

Good stuff, thanks for sharing the stats. Paris is definitely a fast hard court though.

welcome. yeah I know it's very subjective. if you follow a link to the other thread, you will find people with the opposite opinion. had to go one way or the other!!

 

[stringertom]

Djoko looks happier than Fed here

Why not? It was his first ever slam final and that piece of hardware became his second favorite trophy.

 

[Boom-Boom]

welcome. yeah I know it's very subjective. if you follow a link to the other thread, you will find people with the opposite opinion. had to go one way or the other!!

Yeah it's debatable was just ratiocinating

 

[stringertom]

I want to vote the fourth option but I can't accept the typo so I'm abstaining.

 

[Starfury]

Looking at those stats, it appears to me that the measure of the speed of a hardcourt seems to be whether Federer has beaten Djokovic on it (or not played him)

 

[Boom-Boom]

Looking at those stats, it appears to me that the measure of the speed of a hardcourt seems to be whether Federer has beaten Djokovic on it (or not played him)

What would you add as "fast hard court" tourney then?

 

[Boom-Boom]

I want to vote the fourth option but I can't accept the typo so I'm abstaining.

Hopefully the typo isn't a Freudian slip by OP

 

[zagor]

Don't know to be honest, those stats paint a pretty picture but Fed probably had better stats in 2014 and 2015 Wimbledon (than Novak) as well and still lost in the end.

I'd say what matters the most on Sunday is that 1st set, it's absolutely crucial for Fed.

 

[falstaff78]

Hopefully the typo isn't a Freudian slip by OP

yeah i was so annoyed.....copy pasting error

 

[Boom-Boom]

yeah i was so annoyed.....copy pasting error

Anyway happy to see the stats are backing RF who really seems on top these days. Hopefully he can manage tomorrow!!!

 

[Starfury]

What would you add as "fast hard court" tourney then?

It's not about individual tournaments, it's more about how - if you listen to this board - the tournaments always seem to have been slowed down a lot or sped up from year to year. And that the performance of Federer seems to have by far the highest correlation with perceived court speed.

Like for example until 2 weeks ago, the US Open has steadily become slower and slower to the point where some people didn't even call it a fast hardcourt anymore.

 

[falstaff78]

Looking at those stats, it appears to me that the measure of the speed of a hardcourt seems to be whether Federer has beaten Djokovic on it (or not played him)

no sir. that's just the fast time you've seen the data!

 

[Down_the_line]

I honestly don't know what to think because the stats, although really compelling, mean nothing to me. Djokovic has simply found a way against Federer in the Grand Slams since Wimbledon 2012.

 

[Boom-Boom]

It's not about individual tournaments, it's more about how - if you listen to this board - the tournaments always seem to have been slowed down a lot or sped up from year to year. And that the performance of Federer seems to have by far the highest correlation with perceived court speed.

Apart Paris and (maybe) WTF, (as mentioned upfront by OP) don't see any other "fast hard court" currently on the ATP tour

 

[Carsomyr]

I don't know, the numbers look great for Fed, but I think the "Opp U/E" and "Opp gross winners" are probably very key stats. If Djokovic can lengthen significantly lengthen rallies through defense, the more likely Fed will become an U/E machine. We'll see.

 

[Starfury]

Apart Paris and (maybe) WTF, (as mentioned upfront by OP) don't see any other "fast hard court" currently on the ATP tour

Brisbane is the same surface as the AO, yet somehow it's counted as a fast hardcourt. Any reason in particular for that, apart from Federer winning there?

 

[MichaelNadal]

Man, so many threads and so many people all worked up in heated scientific discussions just to be disappointed by another Ultron 4 set win.

 

[Feather]

Nice work falstaff, much appreciated

I don't know. I am too scared to hope. I hope for the best

 

[falstaff78]

Brisbane is the same surface as the AO, yet somehow it's counted as a fast hardcourt. Any reason in particular for that, apart from Federer winning there?

brisbane is one of the fastest surfaces on the tour. these stats are only an approximation to surface speed. because they use ace rate as a proxy for surface speed. can't read too much into them, but I specifically included brisbane on the basis that it's right in the top tier.

http://www.tennisabstract.com/blog/category/surface-speed/

 

[falstaff78]

Man, so many threads and so many people all worked up in heated scientific discussions just to be disappointed by another Ultron 4 set win.

Djokovic has not beaten Federer on a fast hard court in 4 years, and has not taken a set in 3 matches since USO 2014; historically Federer has dominated Djokovic on fast hard courts 13 to 4; (vice versa on slow hard courts 3-10)

surely this has to count for something?

 

[MichaelNadal]

Djokovic has not beaten Federer on a fast hard court in 4 years, and has not taken a set in 3 matches since USO 2014; historically Federer has dominated Djokovic on fast hard courts 13 to 4; (vice versa on slow hard courts 3-10)

surely this has to count for something?

It counts for the fact that they haven't met at the USO since 2011. I would love nothing more than for Federer to carve him up like Razor Ramon, but I don't see it happening. Would love to be wrong, but all the hype is unreal. I see why Fed fans are so gutted after losses.

 

[Boom-Boom]

brisbane is one of the fastest surfaces on the tour. these stats are only an approximation to surface speed. because they use ace rate as a proxy for surface speed. can't read too much into them, but I specifically included brisbane on the basis that it's right in the top tier.

http://www.tennisabstract.com/blog/category/surface-speed/

Really interesting stats. Thanks for the link.

Would have thought the ace rate is a good proxy for surface speed indeed but then it's a bit confusing to see Cincy in Tier 3 and Basle and Dubai in Tier 4...I mean there are obviously other factors (balls, altitude, humidity) but still

 

[falstaff78]

It counts for the fact that they haven't met at the USO since 2011. I would love nothing more than for Federer to carve him up like Razor Ramon, but I don't see it happening. Would love to be wrong, but all the hype is unreal. I see why Fed fans are so gutted after losses.

I have been very circumspect in the lead up to previous major matches with djokovic. as my posting history will show.

forecasting something based on stats is similiar to astrology in one respect. you are trying to divine a future outcome based on disparate facts. which often contradict each other.

however, this is the first time that the stats suggest absolutely no reason djokovic will win. none. at all. try the following exercise. try to forget that the column headings are federer and djokovic. imagine they just say player A and player B.

after I did this exercise, i decided to put my nuts on the table. hence this thread. could he lose? sure! that's why we play, and not decide matches on paper. but i tell you man, HE'S GOT THIS!

 

[Feather]

It counts for the fact that they haven't met at the USO since 2011. I would love nothing more than for Federer to carve him up like Razor Ramon, but I don't see it happening. Would love to be wrong, but all the hype is unreal. I see why Fed fans are so gutted after losses.

Very good post. You hit the bulls eye. the only thing that matters is that they haven't met.

With all due respect to falstaff, Dubai, Cincinnati, Shanghai are all immaterial. They are mere wins. Bo5 is a different ball game.

That said, I again appreciate the work falstaff put into the thread.

 

[falstaff78]

Really interesting stats. Thanks for the link.

Would have thought the ace rate is a good proxy for surface speed indeed but then it's a bit confusing to see Cincy in Tier 3 and Basle and Dubai in Tier 4...

Agreed. That's why we can't mechanically rely on the stats. But it's a good sanity check that what we consider "fast" tournaments are all north of 1.00 and clay and slow tournaments are south of 1.00

Hence when the ace rates say Brisbane is tier 1, it gives me the confidence to put it in the "fast" bucket!

 

[ABCD]

Djokovic has not beaten Federer on a fast hard court in 4 years, and has not taken a set in 3 matches since USO 2014; historically Federer has dominated Djokovic on fast hard courts 13 to 4; (vice versa on slow hard courts 3-10) surely this has to count for something?

Absolutely. Based on your calculations it is clear that Federer will win. I don't see how he can lose. Hands down for your effort. However, there is one thing in your last statistical exercise that bothers me. How is that defensive Djokovic appears to have the most net approaches. I realised that this is probably due to being forced to go to the net (a lot of drop shots were probably played against him). Therefore, it would be great if you can look into the length of balls played by his opponents just prior to the net approach. If you can do this, it would be great. Also, when you compare two players, see whether obtained differences are statistically significant or not.

 

[falstaff78]

Therefore, it would be great if you can look into the length of balls played by his opponents just prior to the net approach. If you can do this, it would be great

we have a proxy for this. namely distance run per point. it would suggest that djokovic is rushing the net late in the points. potentially after he has run his opponents ragged into a position of weakness

Also, when you compare two players, see whether obtained differences are statistically significant or not.

wayyyy too lazy. you are more than welcome!

 

[falstaff78]

Djokovic has simply found a way against Federer in the Grand Slams since Wimbledon 2012.

Very good post. You hit the bulls eye. the only thing that matters is that they haven't met.

With all due respect to falstaff, Dubai, Cincinnati, Shanghai are all immaterial. They are mere wins. Bo5 is a different ball game.

That said, I again appreciate the work falstaff put into the thread.

you guys are making me double down and triple down on my call. but what the heck. I'll do it.

2015 fast hard court Federer is nothing like 2010-2014 fast hard court Federer. He is leagues better. Which is why I included his numbers from USO 2013 to 2014. please look at the improvement and tell me that past USO performances are a good predictor of form!!!

 

[Feather]

you guys are making me double down and triple down on my call. but what the heck. I'll do it.

2015 fast hard court Federer is nothing like 2010-2014 fast hard court Federer. He is leagues better. Which is why I included his numbers from USO 2013 to 2014. please look at the improvement and tell me that past USO performances are a good predictor of form!!!

I said I loved your thread. I will be only happy if Roger wins. Cheers

 

[Boom-Boom]

Agreed. That's why we can't mechanically rely on the stats. But it's a good sanity check that what we consider "fast" tournaments are all north of 1.00 and clay and slow tournaments are south of 1.00

Hence when the ace rates say Brisbane is tier 1, it gives me the confidence to put it in the "fast" bucket!

Yes good sanity check. The other thing a bit surprising is to see US Open exactly on par with Australian Open in terms of ace ratio. Seems as difficult to serve an ace in Flushing Meadows as in Melbourne. Or at least that was the case from 2010-2013...no wonder we had those Djokodal finals...hopefully they did speed up the courts a bit this year!!

 

[falstaff78]

Yes good sanity check. The other thing a bit surprising is to see US Open exactly on par with Australian Open in terms of ace ratio. Seems as difficult to serve an ace in Flushing Meadows as in Melbourne. Or at least that was the case from 2010-2013...no wonder we had those Djokodal finals...hopefully they did speed up the courts a bit this year!!

Yes. That is very odd. But even as late as 2010 it was dramatically faster than Melbourne. and I believe it has been sped up this year, from what i've been reading on the boards and on twitter. (which may explain why Fed has done so well.)

interestingly he himself referred to it as fast in his presser last night.

Q. Can you talk about your serve tonight and how valuable of a weapon it's become over the past couple of weeks in New York?

ROGER FEDERER: I mean, I always thought I had a decent serve with variation, to be quite honest. I definitely had better and worse days before, but now it seems like I'm able to, you know, just serve maybe better consistently maybe.

I'm not sure if that's the case, but definitely also talking about faster conditions in Cincinnati and also now here. I'm using conditions to my advantage, it seems like. I hope I can serve equally good, you know, come Sunday.​

will be interesting to see the data when Jeff compiles it at the end of the season!

 

[Sysyphus]

I'll remain a tragic optimist or something like that until Fed actually seals the deal on MP.

Fed definitely looks to have a realistic shot this time though. But we're talking Novak in a slam final here.

So for me, this is a real toss-up.

 

[ABCD]

you guys are making me double down and triple down on my call. but what the heck. I'll do it.

2015 fast hard court Federer is nothing like 2010-2014 fast hard court Federer. He is leagues better. Which is why I included his numbers from USO 2013 to 2014. please look at the improvement and tell me that past USO performances are a good predictor of form!!!

How do you mean that Federer is now better than in 2010? He is 5 years older now?

 

[MichaelNadal]

The weird thing is tennis-wise I feel like Fed should win this. I just can't see him getting past Novak's fitness, that's the problem.

 

[falstaff78]

How do you mean that Federer is now better than in 2010? He is 5 years older now?

it's quite simple. look at the stats. in 2014/2015 federer has improved his results on fast hard courts. this is because of his new racket and new aggressive approach.

2010 - 2013: 67-13; 5 titles 4 finals
2014 - 2015: 48-2; 7 titles 1 final

 

[Boom-Boom]

Yes. That is very odd. But even as late as 2010 it was dramatically faster than Melbourne. and I believe it has been sped up this year, from what i've been reading on the boards and on twitter. (which may explain why Fed has done so well.)

interestingly he himself referred to it as fast in his presser last night.

Q. Can you talk about your serve tonight and how valuable of a weapon it's become over the past couple of weeks in New York?

ROGER FEDERER: I mean, I always thought I had a decent serve with variation, to be quite honest. I definitely had better and worse days before, but now it seems like I'm able to, you know, just serve maybe better consistently maybe.

I'm not sure if that's the case, but definitely also talking about faster conditions in Cincinnati and also now here. I'm using conditions to my advantage, it seems like. I hope I can serve equally good, you know, come Sunday.​

will be interesting to see the data when Jeff compiles it at the end of the season!

Yep think by conditions, he also means the roof structure that is taking out the wind of the equation. But by seeing the match last night and how Fed managed to steal time from Wawa, it definitly looked way faster this year!

Will be interesting to see the stats for 2015 indeed!

 

[Gazelle]

I'll remain a tragic optimist or something like that until Fed actually seals the deal on MP.

Fed definitely looks to have a realistic shot this time though. But we're talking Novak in a slam final here.

So for me, this is a real toss-up.

Novak in a slam final is nothing special. Novak in a slam final against Federer, is a different animal unfortunately.

 

[NatF]

I think Djokovic will win but I believe it will be tight. Middle of the day so the quicker conditions might help Federer.

 

[Sysyphus]

Novak in a slam final is nothing special. Novak in a slam final against Federer, is a different animal unfortunately.

Hehe, well, I think it's special in that it takes a very remarkable performance to take him out, as for instance Stan and Rafa have done in pretty recent times.

Fed can give such a special performance too, but my question is mostly if he can sustain it long enough at this age.

 

[Gazelle]

Hehe, well, I think it's special in that it takes a very remarkable performance to take him out, as for instance Stan and Rafa have done in pretty recent times.

Fed can give such a special performance too, but my question is mostly if he can sustain it long enough at this age.

I thought it took more a remarkably below par performance from Djokovic to defeat him. Peakovic is untouchable since 2011.

 

[Sysyphus]

Wilander just gave some thoughts on the final: Fed has a shot. However, the way Fed steamrolled a solid Murray at Wimby, and still didn't take any of the momentum after winning the third doesn't inspire confidence against Novak. Yes, he'll play very well for an hour--but can he keep it up for much longer?

Unfortunately, I think it's a pertinent point.

 

[World Beater]

In the last two usopen matchups between the two, djokovic has won his sets much easier.. Sometimes by double break.. Whereas fed has won his much tougher.. Thats is in tbs or single break..

Does anyone know the pt breakdown between the two?? I expect djoker to have won more!!

This match will depend on feds return imo.. The main reason he could not beat novak at w is because his return is not consistent enough

 

[tennis_pro]

Djokovic in 4, I'm not even gonna watch Mr Boring or Sir Like_me_please! win another Slam. He's building up his legacy by beating an Old Man, just because his name is Federer doesn't mean those victories mean squat. If Federer was the same age as Djokovic I'd buy some popcorn a coke and watch Fed demolish him in straight sets.

 

[zagor]

The weird thing is tennis-wise I feel like Fed should win this. I just can't see him getting past Novak's fitness, that's the problem.

There's a reason you don't see many 30+ age slam winners (especially the ones who played 1300 matches), fitness is a crucial element in a BO5 match. Just the way it is.

 

[zagor]

Wilander just gave some thoughts on the final: Fed has a shot. However, the way Fed steamrolled a solid Murray at Wimby, and still didn't take any of the momentum after winning the third doesn't inspire confidence against Novak. Yes, he'll play very well for an hour--but can he keep it up for much longer?

Unfortunately, I think it's a pertinent point.

Genius insight from Wilander, next he's gonna proudly announce that the sky is blue.

 

[The_18th_Slam]

How do you mean that Federer is now better than in 2010? He is 5 years older now?

He's better relative to the field. 2015 is a weak era, so it's easier to be better than the field [/weakeratheoristmode]

 

[Sysyphus]

Genius insight from Wilander, next he's gonna proudly announce that the sky is blue.

trolo, true, but it's actually very worth a reminder given that many think Fed is the favorite or close to it, being impressed by his stats up to the final.

It's easy to get carried away and lose sight of the obvious.