The Player Analysis Model Part 2 – The Problem with Centre Backs (and the MLS)

We have added Centre Backs and the MLS to our new Player Analysis Model and we’ve found some gems but does it create more questions than it answers?

In my last post I introduced our new Player Analysis model based on the Tableau framework, and therein looked at a handful of fullbacks. A week on and I’m pleased to say that I’ve added Centre Backs (DCs for the purpose of the model) aswell as a few more FBs and players from the MLS.

Unfortunately, both Centre Backs and the MLS are pretty tricky subjects – I’ll talk about the MLS later but before we get down to some analysis, let’s talk about rating defenders. It’s bloody difficult. For example, the number of tackles a player makes doesn’t really tell us much. He may play for a team that is rarely in possession meaning he’ll have more opportunity to make tackles, but does that make him a better defender than a less burdened centre back playing for a high possession team?. There are also the intangibles – marking, closing down, positioning – all things that are, in theory at least, quantifiable but not with the free to use data that is currently available. I have attempted to take the sting out of this in a number of ways, including

  • By looking at the % of a teams defensive actions per game a player performs (which, if nothing else, is indicative of their work rate)
  • By throwning in Thiago Silva (regarded as one of the best defenders in the world) as a bench mark for the other players.
  • By Looking at Individual Tackle Success Rates

Of course, the other benefit of this model is that it looks at the players ability in all aspects of the game, and not just defending, and so we can form a holistic picture of that player but even so I would ask you to remain wary of the data for this one and we’ll try to find a way around it at a later date.

For the time being though, here’s the new and improved model. We’ll be discussing the centre backs in this article but feel free to look at the full backs if this is your first time experiencing the software.

I shirked from my analytic duties in the last piece by giving but a brief synopsis on a few of the FBs that I’d studied so this time round I’ll go a little more in depth with a few of the centrebacks.

Bruno Martins Indi

A good World Cup and links with incumbent Man United manager, Louis van Gaal, made the young Dutchman a regular part of the transfer gossip columns this summer. And gosh darnit, they got it right for once, well half-right, as the Feyenoord youngster signed for Porto on Tuesday (you’ll have to trust me when I say that I evaluated him before the transfer was announced) for Porto…for £6m. That’s daylight robbery at it’s most flagrant.

Indi had the second highest tackle success rate out of all the players we looked at (only Thiago Silva was more successful) and his European Tackle Success z-score (see the glossary for an explanation) was 1.2 above the average of the top 50 tacklers in Europe.

He’s a pretty good passer of the ball too. He had the 4th highest PS% and LB% Success out of the studied players and whilst both of his z-scores for these stats just fell below average, it should be kept in mind that this was quite a stringent test – looking at the top 50 passing centre backs over the Top 6 European leagues.

What is also impressive about Indi (and his miniscule transfer fee) is that he played left back for almost half of last season. And he did pretty well at it to boot. He created 13 chances, the majority of which came from the left side of the pitch…

…and (unsurprisingly) had the 2nd highest European z-score for Cross Completion.

I think though that there is a caveat to Indi. Whilst he had a high tackle success rate he also had the lowest % of Team Defensive Actions pG and the lowest Defensive Contributions pG out of the studied DCs. This of course is not conclusive evidence of a flaw in his skills – it may be because he spent half his season out wide, where presumably defensive actions are less common, or perhaps he played less of a pressing role and so was not called upon to make as many defensive actions. It does however call into question whether he can maintain his impressive tackle success rate if he is called upon to contribute more in Portugal, but at £6m I’d say it’s worth the risk.

Joel Veltman

I’m not going to beat around the bush here, I’m going to stick my neck out and say that Indi’s dutch teammate, Joel Veltman, is a better player. Perhaps even the best DC (with exception of Thiago Silva) out of the studied players.

Okay, he didn’t have the best Tackle Success %…

…but it was still above the European average. And he did have the highest % of Team Combined Definsive Actions.

Away from defence, he had the highest per Game Passing Contribution and was only one of two players to have an above average European z-score in terms of PS% and one of three for LB Success%. He even had the highest attacking contribution pG rate – dwarfing all his European counterparts – making him arguably the hardest working DC on the list.

Because his defensive contributions per game were lower than most of the studied players it could be argued that Veltman may be the kind of player that pulls the odd David Luiz and goes missing but I would counter with the fact that he still performed a huge percentage of Ajax’s pG Defensive Actions:

So, to sum up, young Veltman is the future, and with a current market value of just over £3m ( ) the future is looking quite affordable.

Andrew Farrell

I genuinely believe that if football now takes off in the U.S, like many are predicting, the MLS will become a great breeding ground for World talent, and so warrants a place in the study. But if I am 100% honest the only reason I initially added the American league to the model was to take a closer look at DeAndre Yedlin and the New England Revolution’s young centre back Andrew Farrell.

And at first glance, it seemed to be well worth the inclusion.

  • He was one of only two players to have an above average European Dribble Success z-score (yes, we know he didn’t play in a European league)
  • He had the highest European Key Pass z-score
  • He was the only player to have a positive Chances Created European z-score
  • And, as the graph below shows, he was only second to Joel Veltman in the busting-your-ass award having a phenomenal attacking contribution rate and a good defensive contribution rate:

It may also seem like a good thing to mention that he had the third highest European z-score for Tackle Success % but in fact this is where the trouble begins. Take a look at his European Tackle Success z-score compared to his Domestic Tackle Success z-score:

Something is very wrong here. Although his tackle success rate of 60% is well above the European average, it is significantly lower than the MLS average. This can mean one of two things:

  1. The MLS has an abundance of world class DCs or,
  2. Tackling is an easier action in the MLS than in the European leagues.

No offence to MLS intended but I think the latter is more likely. This is definitely something I want to investigate further but my initial thought is that there is a technical proficiency gap between the MLS and the top European leagues. A cursory piece of corroborating evidence would be the average PS% between leagues: the average PS% for teams in Serie A last season was 81.7, in the Premier League – 80, but in the MLS it was 76.9 (only one team had a PS% over 80). I’d therefore guess that defenders find it easier to do their job when their opponents are less technically apt when going forward.

This should necessarily cast a pretty dark shadow over Farrell’s potential outside of the MLS. His attacking proficiency is a plus but his ability to perform his primary function as a defender must be called in to doubt. He looks like he has potential, but it may still be a while before we start seeing an influx of young American’s in the higher echelons of European Football.

Leave a reply