S.O.A.R

When it comes to sports statistics, some people enjoy stats from purely a research-driven view of the sport, and others just enjoy backing up their arguments with fellow fans.

For those that like sports stats for either reason, soccer can be a maddening game. 

There are very few indicators of player performance in a given match, other than goals, assists and bookings.  Given that these events are so rare in the context of a game, they don’t tell the whole picture.

Soccer right now appears to be in somewhat the same context as hockey was several years ago in relation to analyzing gameplay statistically.  Both are fluid games and tend to track progress by only the small end-driven stats like goals, rather than all the work that happens in between.

Baseball and football on the other hand have had a ton of research done to break down players by position and the contributions they make to team wins. 

The EPL is probably the most stat-laden soccer league I can find at the moment, with sites like Opta Sportsdata providing league leader data for Sky Sports.

That said, many of the stats I see present are count data like # of goals, # of passes or # of tackles.  While these are very valuable, they give very little insight into the efficiency of a player given the amount of playing time they are afforded.  In the case of discovering talent, this can be misleading since many young players don’t get a ton of time on the pitch.

The other type of stat that many soccer fans are familiar with is the “rating” the players receive from media outlets after the game.  While it’s easy for a casual fan to understand, it drives the data-obsessed few insane because these ratings are based on nothing more than someone’s opinion.  The Castrol Index seems to be a huge improvement on the old rating system, but I intuitively have my doubts about a system that sponsors certain players and in which I don’t know the weighting of the parameters they measure.

All of my rambling above was merely to give you some background on why I decided to create my own rating system.  At this point in time, most of the data available to the public are offensive statistics, so I’ve started with an offensive rating system.

The Sports Opinionated Attack Rating (SOAR) is my attempt to grade players on a goal per game worth for their attacking ability on the pitch.  That said, SOAR only measures a players ability to influence the game in the attacking third of the pitch, not defensive work or possession play, so please remember that when seeing the ratings.

SOAR currently incorporates 4 things: goals, assists, fouls drawn and time on the playing field.  I would like to include crossing stats soon.

I have also created a comparative stat called Expected SOAR (eSOAR) that evaluates the same impact but uses Shots on Goal instead of actual goals, showing the impact of a player with the same stats who converts goals at the league average rate.  This will give us a measure of either shooting talent or luck, but it will be hard to tell which one over a short period of time.

Goals: Obviously scoring a goal is the most vital of all offensive impact.  That said, I’ve removed any goals scored on penalty kicks, as the foul is the true play that created the chance, the PK is a specialty play that rarely represents a players true offensive impact.

Assists: Just as good as a goal in my mind.  If a player makes a pass that directly leads to a goal, he has clearly created a definitive offensive opportunity.

Fouls Drawn: This is something that we ignore all too often.  A striker who is so athletic, fast or shifty that he constantly draws unfair play from opposing defenders is a valuable commodity.  Free kicks that initiate in the offensive end are the equivalent (or close to) of a shot, not necessarily a SOG.  Last year in Serie A, a goal was scored on 1 out of every 7.98 shots.

Shots on Goal: Very few goals are created without a shot on target.  Last year in Serie A, 44.6% of all shots were on target and 1 out of every 3.56 SOGs resulted in a goal.

The Formula:  SOAR is a rate stat and I’ve converted it into a ratio that represents goals per 90 minutes.  All of a players true goals (minus PKs) and assists are added to the “potential goals” they’ve created with fouls drawn.  These totals are divided by the number of minutes played and multiplied by 90 to give a full game equivalent.  At some point I would like to add crossing statistics but don’t have a reliable source at the moment.

Comparing Players

The nice part about comparing players in full-season increments is that the quality of competition should be relatively equal for all players evaluated.  Each team plays all the other teams twice in a season, so there shouldn’t be any dramatic swings in competition faced.

Obviously players should be compared across positions.  It doesn’t make sense to compare a fullback’s SOAR to a striker’s SOAR.  That said, it may be interesting to compare a winger to a shadow striker in order to decide which one to deploy with the target man.

Remember that a player’s SOAR is essentially the evaluation of a player’s offensive impact in the system his team plays, while playing with those specific teammates.  At the moment, I don’t have a method for separating these factors.  So a team with incredible offense, should have multiple players with a high SOAR.

I have relied heavily on stats pages from the Guardian, ESPNSoccerNet and Football365, so thank you to those sites for making the data available.

Once I start adding players from other leagues, I will normalize the data based on the goals scored per game in each league, just to give a more comparable rating.

Below are some of the articles that include SOAR so far:

Evaluating Italy’s strikers in both Serie A and for the Azzurri

The Most Efficient Forwards at the 2010 World Cup

A Review of the Marouane Chamakh signing by Arsenal

Why Defoe should start over Heskey for England

The most in-form Forwards at the World Cup, After the Group Stage