What is Performance Analysis All About?

After reading this from the excellent Mark Upton I wrote the following reply.

Hi Mark, just spent my Friday evening reading this 3 part series and watching the videos. (Maybe need to get out more!!!). As usual you manage to convey so well a message that has been floating around in my head for ages.

Sport Science in general (but specifically talking about performance analysis) has played a big role in this reductionist philosophy. I would be interested in your thoughts as an ex-analyst.

My thoughts on analysis are thus. It can be an excellent tool for identifying what it takes to win. On base % for example is a great use of analysis/data identifying existing metrics and realising they are not fit for purpose i.e. There is a better way to measure success.

Baseball is an easy example in this case but in theory analysis could be used for similar purposes in other sports. Do you agree with that?

Where I believe Performance Analysis as a whole has got it completely wrong is in the day-to-day dealings with KPI’s.

The reason to collect match information such as pass success rate, or shooting accuracy is because as humans we have terrible recall, among other flaws. So it helps us remove (some) bias and gives us more objective data to work with.

However KPI’s for me should only be used to identify possible weaknesses and force us to ask questions about strategy but more importantly training.

For example a team has poor accuracy in front of goal (any sport). KPI’s would help us identify if this is because of poor shot selection, build up play, poor execution etc…

Unfortunately I feel this is where teams stop with the information. They just set targets based off this info ‘Come on lads, we must do better’. Next week we aim for xyz%

I imagine there are coaches conversations like this;

why is that happening?

it’s happening because we create chances that are too far from goal

why is that happening?

We don’t move the ball quickly enough in transition.

Ok let’s set a transition speed target for players.

This is where the investigation stops – let’s just find another metric to analyse. Instead of following this all the way back to the root of the issue.

How well have we coached (created an environment of learning) to help the players improve in these areas.

PA seems concerned with two things. Creating new metrics to analyse and showing players where they went wrong.

Some of that is driven by coaches but still feel analysts have contributed to that. It feels to me anyway that unless analysis is truly used to aid learning, instead of just monitoring, it is of little benefit.

Interested in your thoughts…



Rob Carroll. Founder of The Video Analyst.com Performance Analyst. Always learning.