To what extent is data science used in Sport?

To what extent is data science used in Sport?

The short answer is “not to any great extent.”  But why? (and what can you do about it?)

I see the fundamental issues as firstly teams not understanding what value data can really offer them and then, when they have invested in data, knowing how to make that data useful (for their specific team.)

This is tricky because it’s not traditionally been done in sport (outside academia) so there is little evidence it will make any difference in practise.

Questions people in sport ask:

“How can you link HR to goals scored?”

“How can you link metres run to wins?”

“How can you link S&C to ‘performance’?”

“How can you link exercise load to injury risk?” … etc.

These questions don’t make sense.

They are not specific enough.

They’re speculative.

They are the wrong way round.

People start with the (often questionable) data they have and then try to make use it.

It is backwards.

Fine, you might get lucky, but it’s not a robust approach.

Sweeping statement, but in my experience, the people who run professional sports teams are rarely confident with maths and technology (and why would they be?)

Sparks rarely set fire to wet logs.

What happens in practise this is that some individual within a team who has some interest in the numbers and is experiencing some frustration at current process or a recent event, takes a spark of initiative and (bravely) tries to start to do something useful with the data.

Often due to their lack of experience, lack of time and lack of clarity of purpose, what they are able to produce can be easily undermined and brutally questioned:

“I get it Jim, but so what?” ….

Senior Manager at Sports Team’s (everywhere)

Unfortunately, this results in a real risk of data and stats becoming (thought of as) a distraction. A waste of time.

Worse, for that individual, it can be seen as them taking time and energy away from other more immediately impactful and more clearly valuable outcomes they could be delivering.

A bad experience for all that fuels (even more) suspicion about “data” and makes it easier to dismiss.

Being the best.

What people want, I believe is to be creating the best (or at lease a better) team, club or national governing body.

They’re competitive. They want to win. They want to invest time in activities that are clear and proven to help that aim.

They want to have (or create, or develop) some kind of competitive advantage for their team to help them win.

They don’t want to invest in any speculative, time wasting activities.

There is no time.

In fact, forget “AI-data-science-machine-learning-optimisation-voodoo” instead just being able to more rapidly and easily visualise long term trends and produce good looking, quick to understand summary reports would be a massive help to the majority.

The elephant in the room however remains the data.

“What secrets is the data hiding?”

“What could be done with it, in the right hands? “

“What if someone who know about data and how sports teams worked, what could they produce?”

It is a nagging feeling of doubt.

“My biggest worry is we’ll turn up at the Olympic games and see another nation doing something incredible we’d never thought of.”

Peter Bentley, British Sailing (at the time)

In my experience, having been delivering services around data exclusively to professional sports teams for nearly the last 10 years, what people need is help:

  • To get more clarity about how to make the numbers useful, to them, in their specific context.
  • To know how to go about using data / maths to help with their existing activities (training, strategy, recruitment, injury prevention etc) and help inform the decisions they make on same.
  • To understand what is possible (and not possible) with this kind if technology (again in their specific environment.)
  • To understand the pitfalls and where to focus (and where it doesn’t matter).
  • To determine how to progress from where they are now to some improved point in the future (with the minimum of risk and maximum of return.)
  • To implement something to gain confidence and then build on from there (with the right expectations and the right innovation process – for them.)

Clearly Pace Insights exists to help teams do just that. However, we’re expensive and our fees take away from more visible and tangible investments (say equipment, more staff etc).

It takes some clear and forward thinking senior management individuals to risk spending significant budget on only the potential of using data when all around them are doing fine without.

So what can you do at your sports team?

Start with this 5 Step approach

Before engaging with a consultancy like ours, try this simple step-by-step approach to get clarity of purpose and clarity of value.

Using this you can then try to progress it yourself or use it to start a conversation with someone (internally or externally) who might be able to help.

Grab a pen and paper and work through the following:

  1. Determine a decision you make on a regular basis (training programme, team selection etc)
  2. Write down what goes into you being able to decide one way or another (put down as much detail as you can – emotions, feelings are all valid – no-one’s going to see this so be honest with yourself.)
  3. What specific measurable objective data do you use to help with that decision (honestly! And ‘none’ is an acceptable answer.)
  4. If you could know something, anything, that would make your decision easier (or better), what would that be? (I mean anything you can think of. Be crazy here if you need too. Don’t be limited by what you think is possible. Put anything down that you know will give you an advantage making that decision. There might be more than one thing.)
  5. You’re done.

You see because at stage 4, you’ve just outlined a clear and valuable challenge that could (possibly) be answered (to some extent) by the use of “data.”

You also now have a challenge sheet to give to someone who might be able to turn it into something, given their expert knowledge on data and tech.

An Example

For example, I was at a football club recently. We were discussing this subject. The jokey suggestion at step 4 was:

“We’re playing xxxxx away. We’re down 1 nil with 30 minutes left. The decision is what, as a manager, do I do?

The options I’ve got are:

– keep as we are,

– change the formation,

– give the players a different strategic objective,

– make a substitution, or,

– some combination of all.

What I am trying to work out is what is the best thing to do.

What would help that is if I had in my hand some indication of what any of those potential changes would make to the final scoreline.

For example if I sub player a for player b, and change formation from x to y, we’re then guaranteed to win 3-1.”

Senior team member at Championship Football club

It might sound far fetched to him but in fact it is a fairly clear goal of something that would add meaningful value.

Whether it is possible or even whether it raises other questions about the “spirit of the game” etc are all besides the point.

The message is that, almost without realising it, this guy has identified how “data” would make a real different to their team.

As tech people, we’d then consider what data would be needed for that, what we data we had (could have) and then go from there.

Try it for yourself and perhaps data can start to become more useful to you.