Introducing SoccerSaber’s Team Ratings

The below was originally posted on the superb Shots on Target blog, although the version here has been modified a bit. The post introduces my own personal team ratings model based upon an amalgamation of various shot and chance data from Opta Stats. I have great confidence in this model and have been happy to watch as it foreshadowed, among other things:

 

  1. The rise of Southampton as a legitimate threat
  2. The improvement and underlying quality of Liverpool
  3. The overvaluation of Chelsea’s early season performance
  4. The putrid nature of Sunderland’s early season attack

I feel I was a bit ahead of the curve in identifying these constructs and have the model to thank (admittedly, I have missed on a couple as well). Going forward, I will publish model based weekly team ratings on the team ratings page and utilize these in my evaluations and predictions. Hopefully, you will grow to enjoy these ratings as much as I.

Thus without further ado, let me introduce SoccerSaber’s team ratings!

Football statistical modeling is in its infancy. While no doubt major clubs have squads of statisticians with proprietary algorithms defining player and team value, the general population has been left to look at goal records and the league table to determine the quality of player and team alike. That all changed with OptaStats. Now, everyone can see the story behind the game. We can look at the activities throughout the pitch and begin to ascertain which of these lead to goals. Further, we can begin to see which activities indicate innate ability and which are simply the luck of the draw. By combining these we can come up with forecasting models for both team and player, a Holy Grail for fantasy football managers.

While @shots_on_target has primarily focused on player evaluation, I have spent numerous hours over the past months hypothesizing on team value. I have worked to understand the underlying activities that drive goals scored and allowed, allowing me to construct team value models that I believe are far superior to the league tables. While still a work in progress, I am confident enough in these models to share them with you.

It has been known for some time that shots, more specifically shots on target, are great predictors of goals. Teams tend to score on about a third of shots on target on average. More importantly, teams that exceed or fail to achieve that rate one year tend to regress towards the mean the following (see James Grayson’s excellent blog for more discussion). As a result, we can use shots on target rather than goal scored as a better indicator of team performance, mainly due to the sample size issues in goals scored (logically, there are about 3 shots on target per goal scored). This helps us identify teams that maybe underrated or overrated based upon goals alone, especially early in the season when sample sizes are low.

But are shots on target enough? It certainly is a start and much better than plain old shots or goals scored as a forecaster. Yet, to me it seemed…wanting. I looked for some other factor that may do a better job of explaining things.

What I have found is that shots data in combination with “Big Chances” (BC) result in a stronger correlation than shots data alone. BCs are defined by OptaStats as follows:

Big Chance – A situation where a player should reasonably be expected to score usually in a one-on-one scenario or from very close range.

In my mind, BCs represent shots on steroids. Adding BCs as additional factor results in the following improvements in goals scored projections over the past three seasons:

2010 – 1.8%
2011 – 2.9%
2012 – 9.7%

While the improvement is not substantial, it is consistent enough for me to have some confidence in the model. For goals allowed, the improvement is even starker, although the data set available currently only goes back a single season:

2011 – 7.5%
2012 – 15.2%

It is clear the available information strongly suggests the SoccerSaber model is an improvement over a pure shot model, both for measuring team attack and defense.

So there you have it.  As mentioned, I will be referencing these ratings consistently moving forward.

Advertisements

9 thoughts on “Introducing SoccerSaber’s Team Ratings

  1. 2ndMan

    Looking forward to reading this blog when it really get’s going. I have some questions though, is the model essentially a regression model? Regressing goals on SoT and BC? If so could you elaborate on the model a bit more and whether any other variables are included? Thanks

    Reply
    1. soccersaber Post author

      Yes it is a regression model at the moment. And yes it is regressing SoT and BCs. I have numerous ideas to 1) test the predictive nature of the model and 2) tweak the model to include other factors. My wish list is extensive, but hopefully I have a good baseline to begin with.

      One of the things I will be most interested in is how much regression there is YoY for team’s the under or over perform the model’s expectations. Unfortunately, our data set is limited right now, so we will have to wait a few seasons before we really have a good handle on it.

      Thanks for reading!

      Reply
      1. 2ndMan

        Thanks, sounds good and looking forward to seeing the model develop.

        I’d be surprised if you could get a good analyse for teams performance spanning over years, wouldn’t changing managers and squad mean your not really looking at the same team and therefore wouldn’t expect a regression to expected levels over more years?

      2. soccersaber Post author

        Certainly although one would expect some “stickiness” in performance as the majority of the team and some of the tactics are surely consistent form year to year. The hope is that at some point we could identify how those changes will impact future performance, but that obviously is someway off.

        As it is, I believe the current model does a good job of identifying teams that have performed better/worse than the table suggests. I mention Southampton as an example. The model never thought they were a bottom three club. Instead, it placed them anywhere from 13th-16th pretty much all season. This differed quite a bit from the majority opinion and is indicative of the value of the model in analyzing performance.

      3. 2ndMan

        Yea I suppose your right, team’s don’t usually overhaul in just one summer. I’d be wary of things like a team who has been scoring fewer goals than you’d expect from their shots going and buying a clinical striker in the summer. Could you then attribute the stop in under-performance to regression back to expected levels, or an endogenous result of a new signing due to under-performing before.

        If that makes any sense at all?

      4. soccersaber Post author

        No it definitely makes sense. As you allude the real key is to separate individual performance from the team and identify how much that new and shiny striker will impact the team’s performance. At that point, you have something immensely valuable.

        We are a long way away from that as advanced data analysis of soccer is in its infancy. We will get there though.

      5. 2ndMan

        I admire your optimism, I suppose the biggest limit at the moment is the data collected and made available.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s