Yaya Toure is an immense footballer. I don’t think I’ve seen a player in my watching lifetime be as physically imposing – with his physical size and strength he’s a human bulldozer and he’s deceptively quick – but also have such technical ability to be able to act as a playmaker and dribbler. His qualities are pretty unique and he’s been a major asset for Manchester City since he signed for them.
However, I do have severe doubts about whether he is quite as good as the vast majority seem to think, and seriously suspect he will fail to meet many of the staggeringly high expectations set upon him after his fantastic 2013/14 campaign.
What is his role?
Everyone knows and accepts Toure’s role has changed since the move from Barcelona to Manchester City in 2010. The man who scored 20 Premier League goals is a far cry away from the man who played centre back for Barcelona in the 2009 Champions League final in Rome. But the way some talk you’d think the new found abilities going forward were just added to his already great qualities as a defensive midfielder, making him the ultimate central midfielder. After all, you still here the occasional person referring to him as a defensive midfielder and the common assumption is that he now plays as a box-to-box central midfielder. In reality, he has sacrificed a lot of his defensive qualities to become the player he is now.
Yaya 09/10 | Yaya 13/14 | |
Tackles per 90 |
2.36 |
1.67 |
Interceptions per 90 |
2.54 |
0.74 |
Dribbled past per 90 |
0.42 |
1.35 |
The 2009/10 season was an immense defensive profile, with Yaya tackling relatively frequently, intercepting like crazy and almost never getting dribbled past. The fall by 2013/14 is drastic. But the 13/14 version still probably compares well with average premier league players though, right? Not quite.
I decided to look at all the central midfielders who had scored at least 5 non penalty goals last season in the premier league. Allowing penalties would make for unfair comparisons as players could carry no goal threat during open play (allowing them to work more defensively) and still qualify. All these players I thought you could reasonably say were midfielders who carried a goal threat, like Yaya, last season and would probably be expected to play a similar role and share a similar defensive workload. In all there were 10 of them, and I compared their defensive stats per 90 last season.
In terms of tackles per 90 only Jonjo Shelvey completed less than Yaya last season. The likes of Aaron Ramsey, Steve Sidwell and Yaya’s partner in crime Fernandinho made twice as many tackles.
But lots of tackling isn’t always a sign of better defensive work. Tackling is often something of a risk, particularly in deeper positions. Players such as Xabi Alonso believe tackling is a last resort and something only to be done at a time of desperation. Measures such as interceptions can be better at showing how people are reading the game defensively and working hard in the shadows. Well Toure wasn’t exactly getting a lot of interceptions last season, in fact the trend gets worse when we look at them.
A combined tackles and interceptions graph shows how far Yaya falls behind some of his counterparts. While Fernandinho got a lot of praise for his unheralded work last season it is really easy to bypass how important he was generating twice the numbers of Toure and still contributing five goals. Special mentions to Aaron Ramsey who was the most prolific goalscorer among the central midfielders while still generating huge defensive numbers, and Yohan Cabaye, who fell 3rd on the list, marginally behind Toure, while being truly prolific defensively. Newcastle really did have a special player.
But using positive defensive metrics (i.e. occurrences) is debated as the best way to analyse defending. Maybe Toure doesn’t tackle or intercept because he doesn’t need to? A decent way to analyse this is to look at the dribbled past stats. Maybe people are getting past Toure less than the others?
While the trend improves for Toure, it’s still not brilliant, with Toure being a bit average. Remember, getting dribbled past, is bad, so the lower you are on the graph, the better.
Baring in mind that there is an inverse relationship with tackling and getting dribbled past – every tackle is a an opportunity to get dribbled – Toure’s ratio of tackles to getting dribbled past is actually rather poor, even compared to those who get dribbled past more. When you also bare in mind that you could never attempt to win the ball or stop the attacker and not ever get dribbled past it starts to undermine the stat in the case of Yaya.
Overall I actually think there’s a case to be made that he should play as the 10 in a 4-2-3-1, in between Silva and Nasri, with both Fernandinho and Fernando playing behind him. At least in the big games and Champions League. It will be interesting to see how Man City set-up tomorrow against Arsenal.
Analyzing Yaya’s 20 goal season
So Yaya doesn’t do as much defending as he probably should and people think he does. But he still scored 20 goals last season which is more than enough to cover the fact he’s a bit lazy when his team doesn’t have the ball, right? Well in isolation I agree, which is why I’m in no denial about his fantastic season last year. But going forward his lack of defensive work worries me because I have little to no confidence of him scoring 20 goals in a season again. For a start 6 of his goals were penalties. Whether or not penalties should count towards goal rates and top scorer tallies is debated, but one thing for sure is penalties being won is not a consistent matter and Toure was granted the luxury for the first time in his City career last season, so you’d expect him to have an additional goal boost compared to his previous seasons and most other midfielders in the league.
But even him repeating his 14 non penalty goals next season is something I have high doubts about. Why, because last year Toure didn’t shoot more, he just converted a much higher proportion of his shots as goals. In fact, looking back over the last four seasons of Yaya in the premier league his shot generation was actually down marginally, 64 shots all season, compared to 68 and 65 the previous two years respectively, where he scored six goals each. In his time in the Premier League Yaya has shown decent repeatability with his shot generation, bar to an extent, his first season in England in 2010/11. But his 2013/14 goals tally is something he’s never shown a capability of before.
Why is this relevant? Well, because shot generation is actually a better indicator of future goalscoring than goals themselves. Alex Olshansky and Ben Pugsley reveal this brilliantly and show that, bar a few elite exceptions, goals, and assists for that matter, are, a bit random*. Sometimes players can shoot the same amount, from the same places, and just score a lot more.
Yaya’s conversion rate last season was an absurd 31.25%, easily the highest among any prolific goalscorer in the Premier League last season. It’s extremely unlikely to be repeatable in the long run. Granted, it’s unfair to hold Yaya’s penalty snatching against his overall tally and then let it contribute to his high conversion rate, so we have to exempt shots and goals that came from penalties. But that still leaves a conversion rate of 24.13, which is both higher than anything he’s achieved before, greater than other high class central midfielders across Europe have ever achieved since records began.
Below is a scatter graph of Toure, Vidal, Kroos and Modric for each of the last five seasons (since records began) bar Yaya in 2009/10 where his role at Barca was different. Each plot is a season by a player. It shows the number of non penalty shots and the number of non penalty goals against each other.*
The overall trend is strong, with an average conversion rate of 8.63% for 17 of the seasons, bar a couple of outliers. The first is Luka Modric’s last season at Tottenham, where he managed to take 83 shots but only scored four times, and the second, even greater outlier, is Yaya’s 58 shots, 14 goals season last year.
Could the dramatic shift in conversion rate be explained by means other than chance? Could it be that he was taking many long shots previously and is now using his great build and dribbling skills to enter the prime zones more? Could he just have dramatically improved his finishing along with his new found penalty and free kick ability? Quite possibly. I don’t have shot location data, meaning it’s hard to say if he’s shooting from better locations. But Yaya isn’t dribbling more than he used to or other’s do, so unless he’s suddenly dribbling in much more dangerous areas it’s unlikely to be a factor.
As for the finishing, I decided to compare him with some of the world’s best strikers. Lionel Messi, Sergio Aguero, Zlatan Ibrahimovic and Radamal Falcao all strike me as people who are renowned for being fantastic finishers and clinical goal scorers so I looked at their conversion rates for each of the last four seasons, except for Falcao’s 10/11 season, as I didn’t have data for the Portuguese league from back then. In total this gave us 19 different seasons, 4 of which where Yaya’s. Bare in mind these players all play as forwards, they are in an around the box more, meaning they tend to shoot from better areas, and their shot is almost always their main weapon, which is why they have much better conversion rates than Vidal, Kroos and Modric.
Rank | Player – Season | Conversion rate | Rank | Player – Season | Conversion rate |
1 |
Messi – 12/13 |
26.42 |
11 |
Aguero 10/11 |
15.32 |
2 |
Yaya – 13/14 |
24.13 |
12 |
Messi – 13/14 |
14.29 |
3 |
Messi – 11/12 |
20.83 |
13 |
Falcao – 13/14 |
14.00 |
4 |
Aguero – 13/14 |
19.77 |
14 |
Ibra 11/12 |
13.74 |
5 |
Messi – 10/11 |
18.49 |
15 |
Ibra – 13/14 |
12.5 |
6 |
Falcao – 12/13 |
17.39 |
16 |
Aguero 12/13 |
11.63 |
7 |
Ibra – 12/13 |
16.56 |
17 |
Ibra – 10/11 |
9.73 |
8 |
Yaya – 10/11 |
16.32 |
18 |
Yaya – 11/12 |
9.23 |
9 |
Falcao 11/12 |
15.83 |
19 |
Yaya – 12/13 |
8.82 |
10 |
Aguero 11/12 |
15.75 |
Only Lionel Messi’s greatest season ever manages to beat Yaya last year, with none of the other seasons getting particularly close either. Conclusions? Yaya either had a bit of a lucky season with his conversion rate and won’t be able to repeat it, or he suddenly turned into the best finisher we’ve seen in recent time in one highly productive off season. Given he’s the wrong side of 30 and hadn’t shown any signs of a potential improvement in his trend, I’m much more inclined to go with the former.
A more clinical creator than Özil and Silva?
A similar pattern can be found with Yaya’s assists and key passes last year. His key pass rate was again down on the previous two seasons, but his assist tally more than doubled. Toure completed nine assists last season, the same number as David Silva and Mesut Özil, but he only created half as many chances overall – 40 compared to 86 and 76 respectively. This can mean either two things. Toure’s chances created could just be much more clearcut than Silva and Özil’s or, due to factors beyond his and the other two’s control, people just finished his chances better. Assuming the former, realistically, relies upon us believing Yaya is a better playmaker than two of the best chance creators Europe has seen since records began, who is only creating less chances because he is playing deeper. I’m skeptical, and like with his conversion rates, he’d previously shown no similar trends. Indeed, his key pass to assist conversion rate more than doubled anything he’d ever done before. 22.5% last season compared to 8.89%, 5.56% and 10.81% the previous three seasons.
I also compared his key pass to assist plot with Vidal, Kroos and Modric again, and just like with his finishing he stuck out as the most clinical claimer of assists seen recently among players seen as the elite midfielders.
Whether or not a key pass leads to an assist is pretty random. You can quite easily create a clearcut one on one chance for a striker that they can miss, and equally you can play a two yard sideways pass to someone who dribbles past three players and smashes it into the top corner. It’s random and highly unlikely that Toure was anything more than lucky with his chance creation last year.
Conclusions?
I’m not really trying to detract from Yaya’s last season. He probably got a bit lucky yes, but that kind of thing evens out and he was probably unfortunate with conversions in an earlier seasons. But looking towards the future it is highly unlikely Yaya will be able to keep such conversion rates up, so unless his shot generation and chance creation suddenly shoots up, something it hasn’t shown a potential trend of yet, then expect his goal and assist tally to fall, potentially quite dramatically. He’s currently viewed among the very top of central midfielders and the world’s best footballers overall, but I’m wary that people’s expectations of him are unrealistic. If he gets 10 non pen goals and 5 assists this season he’ll have had another great year, and, rather than having a post birthdaygate strop, will have just regressed to the mean. He’s also 31. If anything you’d expect his shot generation and chance creation to fall off, not rise.
Three games into this season he’s had 9 shots and 1 key pass. By last season’s levels, you’d expect him to have at least two goals by now. But it’s simply impossible to keep such levels up. Add in that he’s probably not as well rounded as people like to think and you realise he’s perhaps, contrary to his own opinion, actually overrated, not underrated. Like I said earlier, still a phenomenal players, just not quite as good as the masses think.
* * *
1) Colin Trainor does great stuff on this type of thing, and he’s done loads of stuff that talks further about shot locations and expected goals if you haven’t come across this before and want to read more: http://statsbomb.com/2013/09/where-have-all-the-goals-gone/ http://statsbomb.com/2014/01/why-suarez-cant-keep-up-his-scoring-rate/ http://statsbomb.com/2014/05/liverpool-an-analytical-look-at-201314-a-missed-opportunity/
2)This does work on the flawed assumption that all penalties are scored. I only have data for penalties scored, not penalties missed, and it’s too much work to go through every match individually looking for missed penalties. The difference would only be one or two shots less for a single season maximum though, so it’s almost negligible.
All stats are sourced from whoscored, except for minutes played and penalty goals, which come from transfermarket. Picture © BBC