Tennis Forum banner
1 - 20 of 20 Posts

·
Premium Member
Joined
·
3,323 Posts
Discussion Starter · #1 · (Edited)
On a lark, given the discussions about how "lucky" Emma was to win the USO, I became curious to check the Aggressive Margin stats for her main draw matches. For context, the AM is meant as a statistical indication of a player's level of play in a match, independent of the opponent's level. It's computed as W - UE + opponent's FE, often divided by the total number of games for scaling. Obviously there's some debate about it, but my understanding based on very little reading :) is that tennis stats folks broadly agree the AM/games is the best way to distill a player's level during a match to a single number. There was a user on TF who used to be very passionate about AM and he described it in this thread, including how AM stats apparently correlate to rankings. TLDR, anything in the vicinity of 1.5 is ~ Top 5, whereas an AM of 2 and above is quite rare, and correlates with Top 3 quality.

So Emma's AM by round: 1.70, 2.88, 2.53, 1.46, 1.47, 1.05, 1.57. With a tournament average of 1.80 (sorry, I was too lazy to search for stats for the qualifying rounds).

Pretty good! :)
 

·
Registered
Joined
·
1,375 Posts
Fantastic analysis @SmilingKnight. Inject this sort of material into my veins.

So, according to this system, she was playing top 3-level tennis against Zhang and Sorribes Tormo in Rounds 2 and 3, and around top 5-level tennis against Voegele, Rogers, Bencic and Fernandez, with an outlier against Sakkari.

For a player who was ranked 150 in the world and had played only a handful of professional tournaments, that's a really impressive level of consistency throughout.

It'll be interesting to see if she can re-capture this level of performance in future once she's settled on the tour.
 

·
Registered
Joined
·
3,488 Posts
On a lark, given the discussions about how "lucky" Emma was to win the USO, I became curious to check the Aggressive Margin stats for her main draw matches. For context, the AM is meant as a statistical indication of a player's level of play in a match, independent of the opponent's level. It's computed as W - UE + opponent's FE, often divided by the total number of games for scaling. Obviously there's some debate about it, but my understanding based on very little reading :) is that tennis stats folks broadly agree the AM/games is the best way to distill a player's level during a match to a single number. There was a user on TF who used to be very passionate about AM and he described it in this thread, including how AM stats apparently correlate to rankings. TLDR, anything in the vicinity of 1.5 is ~ Top 5, whereas an AM of 2 and above is quite rare, and correlates with Top 3 quality.

So Emma's AM by round: 1.70, 2.88, 2.53, 1.46, 1.47, 1.05, 1.57. With a tournament average of 1.80 (sorry, I was too lazy to search for stats for the qualifying rounds).

Pretty good! :)
Even against Sakkari she could have been playing top 5 but Sakkari made so many errors. So even if you're playing at a great level, if your opponent makes lots of errors to give you points, you can't earn those points yourself through the whole Aggressive Margin stat.

Not to mention long matches with lots of deuce games can help. Especially for two pushers playing each other that can't finish points and keep rallying. They don't make UEs and there's 100 deuces until there's a "forced error" or "winner". Really this stat helps pushers despite it being called AM.

The relative ability of your opponent also matters a lot. It's much easier to hit winners against some players than others. So then you're supposedly playing better tennis against lower-ranked players. Doesn't account for that incredible top-10 player who makes you hit an extra shot or puts lots of pressure on you to take the first strike before she does.
 

·
Registered
Joined
·
4,813 Posts
Emma played a fantastic tournament at US Open....including winning 3 matches in qualification. She deserved to win...outplayed every opponent put in front of her. That's all that matters. Nice stats though.(y) Wasn't aware of this.
 

·
Registered
Joined
·
3,488 Posts
Regardless of the flaws there are some insights. Especially when paired with watching the match and seeing how aggressive she was. So thanks for providing it! Only other thing. Having the AM of both players would be really good to assess the quality of the whole match itself.
 

·
Registered
Joined
·
229 Posts
Raducanu at USO:

RoundOpponentAM/GameOpponentFinal Set
1RVoegele1.710.471.33
2RZhang2.891.612.50
3RSorribes-Tormo2.540.462.29
4RRogers1.47-0.072.14
QFBencic1.471.051.50
SFSakkari1.06-0.121.00
FFernandez1.580.891.78
[Average]1.811.79

Selected others:

TournamentPlayerAM/Game [Average]Final Set [Average]Var [Final Set]
US OpenRaducanu1.811.790.30
US OpenFernandez0.931.240.78
US OpenSakkari0.991.270.68
US OpenSabalenka1.471.691.12
US OpenBencic1.701.601.41
OlympicsBencic1.281.500.21
WimbledonBarty1.401.760.37
Roland GarrosKrejcikova1.051.320.83
Australian OpenOsaka1.241.550.66

If you are Krejcikova with a low base level you want a high variance to put you over the top at times. If you're Sabalenka you really don't otherwise are in danger of totally busting out like she did in her match against Fernandez particularly in the final set when she was spewing errors by the bucket load.
 

·
Registered
Joined
·
1,375 Posts
So she had an outstanding tournament by the looks of it, statistically-speaking and otherwise, even compared to the likes of Osaka and Barty at the AO and Wimbledon respectively.

The peakiest of peaks or a sign of things to come in future? Tune in to find out.
 

·
Registered
Joined
·
208 Posts
Oh interesting. This really bears out the impression that she’s both really aggressive and consistent. If you have that combination then you can just play simple tennis and win. I wonder what the stats look like for her at IW. Probably negative. :(
 

·
Premium Member
Joined
·
3,323 Posts
Discussion Starter · #12 ·
Raducanu at USO:

RoundOpponentAM/GameOpponentFinal Set
1RVoegele1.710.471.33
2RZhang2.891.612.50
3RSorribes-Tormo2.540.462.29
4RRogers1.47-0.072.14
QFBencic1.471.051.50
SFSakkari1.06-0.121.00
FFernandez1.580.891.78
[Average]1.811.79

Selected others:

TournamentPlayerAM/Game [Average]Final Set [Average]Var [Final Set]
US OpenRaducanu1.811.790.30
US OpenFernandez0.931.240.78
US OpenSakkari0.991.270.68
US OpenSabalenka1.471.691.12
US OpenBencic1.701.601.41
OlympicsBencic1.281.500.21
WimbledonBarty1.401.760.37
Roland GarrosKrejcikova1.051.320.83
Australian OpenOsaka1.241.550.66

If you are Krejcikova with a low base level you want a high variance to put you over the top at times. If you're Sabalenka you really don't otherwise are in danger of totally busting out like she did in her match against Fernandez particularly in the final set when she was spewing errors by the bucket load.
Amazing, thanks! Another interesting datapoint would be Iga at RG 2020. I don’t know this for sure, but I’ve seen it mentioned that Iga has had matches with > 3 AM/games.
 

·
Premium Member
Joined
·
3,323 Posts
Discussion Starter · #13 ·
So she had an outstanding tournament by the looks of it, statistically-speaking and otherwise, even compared to the likes of Osaka and Barty at the AO and Wimbledon respectively.

The peakiest of peaks or a sign of things to come in future? Tune in to find out.
That’s the question, no? The thing that makes me optimistic about her is that it didn’t seem to me like she was redlining, going for the lines all the time. Instead, she often went for safe targets, and would pull the trigger after some point construction (or indeed, on receiving a weak second serve).

But who knows? Tennis is a confidence game and confidence can be easily lost.
 

·
Registered
Joined
·
1,375 Posts
That’s the question, no? The thing that makes me optimistic about her is that it didn’t seem to me like she was redlining, going for the lines all the time. Instead, she often went for safe targets, and would pull the trigger after some point construction (or indeed, on receiving a weak second serve).

But who knows? Tennis is a confidence game and confidence can be easily lost.
True. There were times when she was occasionally painting the lines in her matches, but a lot of her points came from forced errors because she was playing with SO much margin at times that her opponent was able to get a racquet on a lot of her shots.

We'll just have to wait and see. I wouldn't expect her to repeat this anytime soon, but if she's able to hit that sort of peak in a tournament, then her base level could be pretty good with more data required.
 

·
Registered
Joined
·
229 Posts
Raducanu at Wimbledon:

RoundOpponentAM/GameOpponentFinal Set
1RDiatchenko1.320.212.33
2RVondrousova1.170.390.90
3RCirstea2.331.432.17
4RTomljanovic0.150.85-2.00
[Average]1.361.18

She was already playing great tennis at Wimbledon before she blew up against Tomljanovic and her performance against Cirstea was as good as anything anyone has done all year.


There's only W/UE data on the official sites from previous years for Roland Garros but here's data for the the rest of the recent major finalists.

YearTournamentPlayerAM/Game [Average]Final Set [Average]Var [Final Set]
2021US OpenRaducanu1.811.790.30
2021US OpenFernandez0.931.240.78
2021OlympicsBencic1.281.500.21
2021OlympicsVondrousova1.141.040.55
2021WimbledonBarty1.401.760.37
2021WimbledonPliskova1.491.630.72
2021Roland GarrosKrejcikova1.051.320.83
2021Roland GarrosPavlyuchenkova1.161.050.05
2021Australian OpenOsaka1.241.550.66
2021Australian OpenBrady0.660.850.21
2020Roland GarrosSwiatek1.691.810.18
2020Roland GarrosKenin1.161.981.87
2019Roland GarrosBarty1.231.570.32
2019Roland GarrosVondrousova0.891.480.41
2018Roland GarrosHalep1.041.600.23
2018Roland GarrosStephens1.051.240.31

Swiatek was in Raducanu's league, just as air tight and consistent from match to match. Barty was more inconsistent within matches at Wimbledon this year but she could deliver her highest level when it mattered on a consistent basis just like them. Bencic, Halep and Barty at RG all had a top 5 level they could consistently find a way to reach even if they took a fair bit of time to find it.

For the sake of comparison, here's what Swiatek/Kenin did following up:

PlayerAM/Game [Average]Final Set [Average]Var [Final Set]
Swiatek1.231.300.43
Kenin0.880.890.60

Kenin's ballbashing isn't the problem oddly enough, it's that she can't ever get in sync and find the wild level she was getting at times at Roland Garros the previous year which when it clicked allowed her to plough through her draw. Swiatek dropped off but she still has the same level of consistency to go with it, if Raducanu drops off to a similar extent she'll have similar runs where she'll cruise to QF with no exceptional performances and then bomb against the first player good enough to take advantage.

Swiatek's level these days is closer to what her real level probably always was and it'll likely be the same with Raducanu. She's going to need time to develop her game to point where she can deliver her highest level on a consistent basis even when everyone knows what to expect and that's going to take years even if she is a GOAT in waiting.


Don't read too much into Osaka's numbers, the level of play at this years Australian Open was pretty mediocre:

PlayerAM/Game [Average]Final Set [Average]Var [Final Set]
Osaka1.241.550.66
Brady0.660.850.21
Serena1.041.150.98
Muchova0.771.070.51

She was still the class of the field by a distance and still added another slam while not really having to get out of second gear. I'd imagine her peak levels at the first slams she won were at least on Raducanu's level if not higher and if she ever figures out her issues she'll be back delivering that level of play.


Individual match performances of the finalists above 2.00:

PlayerYearTournamentRoundOpponentAM/GameOpponent
Raducanu2021US Open3RSorribes-Tormo2.540.46
Raducanu2021US Open2RZhang2.891.61
Bencic2021Olympics1RPegula2.221.22
Pliskova2021WimbledonQFGolubic2.250.63
Raducanu2021Wimbledon3RCirstea2.331.43
Pliskova2021Wimbledon2RVekic2.440.94
Osaka2021Australian Open1RPavlyuchenkova2.200.47
Swiatek2020Roland Garros4RHalep2.070.53
Kenin2020Roland Garros3RBara2.360.71
Swiatek2020Roland Garros2RHseih2.061.00
Halep2018Roland Garros2RTownsend2.310.88

Crazy performances really aren't that common and her performances in the early rounds should have deservedly been bragged about. It wouldn't have put her fave for the tournament as both Bencic and Sabalenka had a 2.50+ performances in the early rounds at USO and that's just of the players I've checked but it would still mark her out as one to watch. Her blow up against Tomljanovic should have kept people nervous about her ability to make it all the way through but after she beat Bencic people should have known what was going to happen in the proceeding matches.
 

·
Registered
Joined
·
1,375 Posts
Oustanding job Celeste. (y)

Makes you wonder how well she might have done at Wimbledon had she not had breathing difficulties. From memory, she was starting to suffer towards the end of the first set against Tomljanovic. That Cirstea performance is an eye-opener, because she's no mug and based on her statistics, was playing well herself.

Another eye-opener were those statistics for the Zhang match at the USO. I've watched the match a couple of times now, and thought Zhang didn't play as well as she did in San Jose 6 weeks prior. Perhaps she did, but Raducanu was just that much better on the day, and that much poorer in San Jose.
 

·
Premium Member
Joined
·
3,323 Posts
Discussion Starter · #18 ·
That Cirstea performance is an eye-opener, because she's no mug and based on her statistics, was playing well herself.
Yes, my impression at the time was that Cîrstea was very unlucky to run into Raducanu, because she herself was playing very well. The stats for that match confirm it.
 

·
Registered
Joined
·
89 Posts
On a lark, given the discussions about how "lucky" Emma was to win the USO, I became curious to check the Aggressive Margin stats for her main draw matches. For context, the AM is meant as a statistical indication of a player's level of play in a match, independent of the opponent's level. It's computed as W - UE + opponent's FE, often divided by the total number of games for scaling. Obviously there's some debate about it, but my understanding based on very little reading :) is that tennis stats folks broadly agree the AM/games is the best way to distill a player's level during a match to a single number. There was a user on TF who used to be very passionate about AM and he described it in this thread, including how AM stats apparently correlate to rankings. TLDR, anything in the vicinity of 1.5 is ~ Top 5, whereas an AM of 2 and above is quite rare, and correlates with Top 3 quality.

So Emma's AM by round: 1.70, 2.88, 2.53, 1.46, 1.47, 1.05, 1.57. With a tournament average of 1.80 (sorry, I was too lazy to search for stats for the qualifying rounds).

Pretty good! :)
The 2.88 is probably why Zhang was shaking her head in disbelief at times during the match !

SST got off lightly !
 

·
Registered
Joined
·
89 Posts
That’s the question, no? The thing that makes me optimistic about her is that it didn’t seem to me like she was redlining, going for the lines all the time. Instead, she often went for safe targets, and would pull the trigger after some point construction (or indeed, on receiving a weak second serve).

But who knows? Tennis is a confidence game and confidence can be easily lost.
She hit a lot of lines.
 
1 - 20 of 20 Posts
Top