Reviewing the 2008-09 Actim Index

Nicolas Anelka is a better striker than Fernando Torres.

How much better, you ask? 1.93 times better, that’s how much. In fact, Anelka is the best player in the entire Premier League, and the margin isn’t even close between him and the second best player: Dirk Kuyt. Also, Manchester United’s best player is Dimitar Berbatov, Lampard’s better than Gerrard any day of the week, and Jamie Carragher was the best defender in the league last season.

Sound like a parallel universe you want to live in? According to the folks who run the Actim Index, you’re already there. Welcome to the wild world of soccer statistics, where what you see on the pitch may be lying to cover up what’s happening off the ball.

American sports are covered in statistics. If you missed a baseball game, you can still read a box score to get a reasonable idea of what went on; miss a gridiron football game and you’ll still want to know what your quarterback’s rating was and how many yards the running back got; if you’re really clever, you’ll also want to know how many carries it took to get him there to get some idea of how well he played. Too many carries? They were stuffing him at the line. Too few? He must’ve gotten everything in one long run.

The issue that soccer stats will always have to contend with is that there are no isolated incidents. It’s the polar opposite of the statistical goldmine that baseball provides; instead of being strictly a contest between one man (the hitter) and another (the pitcher), soccer incorporates all eleven players in such minute ways that it’s almost impossible to represent it statistically. For example, the main event that triggered a goal might not be the assist; it might be the tackle three plays back that a midfielder made in space, or a long clearance to a winger who was already running. How do you quantify what you’re seeing in a meaningful way that allows you to assign a value to it?

Actually, let’s step back even further: why would someone even want to assign a value to the movements on the pitch?

One reason is tactical. Charles Reep is a good example of someone looking at stats to try and divine some sort of tactical advantage. Back in 1950 Reep, an accountant by trade, came to the conclusion after watching several games that most goals resulted from three passes or less. Neverminding the possible reasons for this (that goals are often scored on the break, or that goals are easily scored when a defender gives up possession cheaply in his own half), Reep decided that the best, most efficient style of play possible was a direct, “long ball” tactic. This tactic became the primary tactic of the Norwegian and English national team for a good portion of the seventies and eighties.

There are fewer Charles Reeps out there than there are Billy Beanes anymore. Beane, the protagonist of “Moneyball”, is best known for taking the Oakland A’s to prominence in spite of a lack of funding. He did this by exploiting the market and placing value on things that actually win games (like an ability to take walks) instead of things that don’t (anything aesthetic, like a player’s weight).

Let’s put that into soccer terms for a second. There’s a midfielder that two teams want; let’s call him, I don’t know, Seymour. Two teams are analyzing Seymour; one of them looks at the stats on the Barclay’s Premier League website, finds out that he had two goals and seven assists in twenty nine  games, and decides not to put a bid on him. The second team, using a more comprehensive skill set, realizes that said midfielder makes a ton of tackles in midfield. They also recognize that the team’s overall performance dropped in the nine games he missed. Without competition for his services, they’re free to sign him.

That’s what the Actim Index is trying to chart: over the course of the season, how many plays does a player make, and how does the team do when he’s out there? According to their website, there are six calculations:

1. Assesses a player’s contribution to a winning team, based on points won by the team when he appeared.
2. Assesses a player’s performance in each game, by allocating points for actions that positively contribute to a winning performance such as shots, tackles, clearances and saves. It also takes points away from players for negative actions such as yellow/red cards and shots off target.
3. Allocates points based on time on the pitch.
4. Allocates points for goal scorers.
5. Allocates points for assists.
6. Allocates points for clean sheets.

After calculating all of that data, the top 100 player list is then released. They do this weekly (the end of the season is available here); it’s strictly a “gross points” stat, so players who play more games will inevitably be ranked higher.

Simple, huh?

Sadly, the Actim Index has some issues. For one, there’s a tendency to rate players who play forward too high; while Anelka and Kuyt are good players, they aren’t the first and second best players in the league. They both play in a similar way, though: they’re unselfish, provide the ball to other players, have tons of stamina, and (most importantly) score goals for teams that also happen to have good defenses. That last part is crucial; if Anelka had the exact same stat line but instead played for Newcastle, he wouldn’t be anywhere near as highly rated because the team wouldn’t have picked up points while he was playing. Conversely, the opposite is true: a goalkeeper who keeps a clean sheet is punished if his offense is unable to score and pick up points.

Another issue is that the Actim index focuses on a player’s actual output rather than their potential output. Take the Anelka/Torres example above; Anelka accumulated 751 points to Torres’s 389, but he did that with thirteen more games. If we’re grading potential, though, than the number that’s really needed is how many points per appearance a player earned. That tells you what a player would be capable of in a perfect season (incidentally, the player who picked up the most points per game was Ryan Giggs; in only 15 appearances, Giggs accumulated more points than Fulham’s Danny Murphy earned in a full 38 game slate).

Since I’m a stats geek, I put that whole image file into a spreadsheet, added in positions and player appearances, and came up with some interesting info.

• Statistically speaking, Dimitar Berbatov was more helpful to Manchester United than Cristiano Ronaldo. Berbatov was actually United’s highest ranked player; he was ranked fourth overall in total points, third overall in points per game (Actim points, not goals).
• Last week I wrote a piece on my five most overrated players; according to the Actim index, I’m an idiot. According to Actim, Jamie Carragher is a top-ten defender (in both total points, where he’s first, and points per game, where he’s sixth), while Gareth Barry’s the seventh best player in the league. Robinho’s also on the list as the 24th best player. Tevez isn’t on the list, but his lack of minutes could account for that.
• Rafa Benitez’s transfer policy has been highly critiqued, and the Actim index doesn’t answer all of the questions out there. Alvaro Arbeloa, who’s set to be replaced by unranked Glen Johnson, was one of Actim’s top defenders in the league. The other major transfer – the one that didn’t happen – was Gareth Barry to Liverpool with Xabi Alonso going elsewhere to fund the deal. According to Actim, Benitez was 100% right about this.
• Manchester United has a complete eleven-man side on the top 100 list.
• Players I was surprised to not see listed: Michael Essien, Javier Mascherano, Robin van Persie, Luka Modric, Andrey Arshavin (who I suspect has a really high Points Per Appearance score and just not enough total points), Ricardo Fuller, and Carlos Tevez.

Does this tell us anything? I’m not sure, actually; the index has some flaws, and the lack of data on the remaining three hundred players makes it hard to use effectively. Is it interesting? If you’re a stats geek, yes, it is.

Assesses a player’s contribution to a winning team, based on points won by the team when he appeared.