Chelsea: How to use football statistics when analyzing players

Why does the statistic say what it says?

This is where context comes in. Some players have certain statistics to their name that at first glance it seems outrageous, but if you look closer, you then realize “wait that’s not so impressive”. For instance, players in high possession sides would often make lots of short passes and would therefore average high pass completion rates amongst the central midfielders and central defenders.

When you know this, you would be better placed to evaluate the passing abilities of players in teams like that. Likewise a team that mainly builds up through long balls would generally have lower pass completion rates, this doesn’t necessarily make them bad passers of the ball either. Manchester City have up to 9 players on 90-percent pass completion and above, Brentford on the other hand, have no one on up to 90-percent,  they do have one player on 85-percent, two players on 80-percent and three players on 79-percent. How then should you try to evaluate the passing ability of a Brentford player compared to a Manchester City player?

It is here we look at more passing metrics, like long balls, but not limited to it. Long passes are infinitely more difficult to complete, so a player that completes a healthy amount at a healthy rate, should be of interest. Passing is just one example, as it is the easiest to investigate. Other passing statistics like key passes and progressive passes are also tied to total passes. This is also important regarding why some statistics say what they say. A player that makes more passes per game has more opportunities to therefore make more progressive passes, key passes and others. Therefore passes per key pass, or passes per progressive pass, may be a better indicator than just raw progressive passing statistics.

For instance Ziyech makes 1.6 key passes per game in the Premier League for Chelsea this season. Callum Hudson-Odoi makes 2.0 per game, however Ziyech averages 13.2 passes per game, compared to Hudson-Odoi’s 26.2, this means Ziyech makes a key pass every 8.3 passes, and Hudson-Odoi makes one every 13.1 passes. After context is applied, what looked like Hudson-Odoi creating more than Ziyech doesn’t look quite like that anymore.

Many things in football statistics are reliant on other things, and it is the responsibility of human beings to determine what those dependencies are. It is the responsibility of whoever is using the statistics to find and apply necessary context, so that players can be judged and evaluated fairly. Sometimes context is not easy to apply, as some data is more readily available than others.