When dealing with a past problem, identify instances where the issue occurs in the data. Use the instances identified to create a training regimen to solve the problem. Carl Carpenter explains and provides examples for my favorite use case for data.

This is my favorite example from his post, “Using Tracking Data To Inform & Provide Blueprints for IDP Training Sessions”. He comes to this final conclusion after analyzing the data and video while working with a midfielder, trying to understand why they lack the ability to turn and face up the pitch when receiving passes from the center-backs:

The midfielder is left footed and it’s clear that they struggle to receive across their body (i.e. let the pass run across them so they can turn at the same time up the pitch): Rather than doing so, they almost take an unnecessary touch which closes open spaces as opponents recover and shut down lanes OR they simply don’t even try to scan and turn — choosing to play back to the CB who has just played the pass.

In this hypothetical scenario, the analyst can bring a video package of these 15 clips to both the player & coach and watch them together — 15 clips that provide infinitely more value to the 50+ clips that you could possibly show them. Since I added in filters for space the CM received in and where his teammates were positioned ahead of him at the same time, it’s impossible for them to misconstrue the message tracking data provided you. The IDP sessions one can create from this are endless, both in a vacuum and as a part of a coaching point for 11v11/small sided sessions with the rest of the group.

It is one action in a game, but when you fix that one action, the game opens up.

It takes a smart mind from the analyst to think of the ‘when’—to locate the area they need to focus on. That objective for the coach and player could not be created efficiently without this type of data analysis. It speeds up the process of finding all the needles within the haystack.

It is like climbing. It is one thing to say, ‘you have a problem receiving passes from the center-back’, and then we are forced to discover the path through trial and error. It is another thing to see the issue, and then immediately know the intended path because you can see what works and what doesn’t in the data.