Glossary

I received some feedback yesterday requesting some terminological clarification, so I thought I’d add a glossary page for some of the stats I tend to use.

First of all, if I have doubts about event definitions, I usually revert to this page from Opta, so that’s a good place to start if you’re confused about some of the more basic stuff. Now obviously, InStat is the source for passing/defensive stats, so optimally you’d want their definitions, but I haven’t found them online so have decided to stick with Opta.

A lot of my analysis is based on an Expected Goals (ExpG or ExG or xG for short) model I’ve developed. Expected Goals is an increasingly popular way of measuring chance quality, that basically takes every individual shot, and compares it to other shots with similar characteristics (location, header/foot etc.) and gives the historical rate at which that chance has been converted into a goal.


When it comes to shot based stats (including ExpG), I tend to talk about For, Against, Difference and Rate.

For = Production by the team in question (e.g. Shots For)

Against = Production against the team in question (e.g. Shots On Target Against)

Difference = Difference between For and Against (e.g. Goal Difference)

Rate = The percentage of production created by the team in question compared to its opposition (e.g. ExpG Rate/Total Shots Rate/Shots On Target Rate). This is calculated thusly: Production For/(Production For+Production Against).


ExpG under/overperformance = The amount of ExpG a team/player has accumulated compared to the amount of ‘real’ goals they/he has scored. If there are more goals than ExpG then that is considered an overperformance and vice versa. Generally speaking, goals tend to regress towards ExpG, meaning that if a team is overperforming ExpG, they are converting chances at an unsustainable rate (think loads of shots from distance) which means that you would expect their ‘real’ performance to cool down with time. The opposite is also true, naturally.

As an addition to the above point, it is worth mentioning that there is a measurable difference between players’ finishing ability – just that it requires more data than I have (or most anyone else for that matter), and that it mostly concerns elite players. My conservative estimate is that the difference in finishing ability between two Veikkausliiga players is negligible.


Possession adjusted defensive stats = I don’t use these a lot, but they pop up from time to time. Basically, since you can only do defensive actions when not in possession of the ball, a team with less possession is going to accumulate more defensive actions. By adjusting defensive stats for the rate of ball possession, you can get a slightly improved look at who does what and how often. There isn’t (to my knowledge) an industry standard for possession adjustments, but I do them on a game to game basis. You can read more about the concept on Statsbomb.

 

Advertisement