Do you need statistics when you coach at amateur level

Statistics tables impress committees and reassure coaches. They have one flaw: between the possession percentage displayed on Monday and what happens in Tuesday's training, there is most often no link at all. A figure that changes no decision is decoration.

Why most figures are useless at amateur level

• The sample is too small. An amateur season has twenty to thirty matches, often with a squad that changes every week. At that volume, the gap between 48% and 54% possession is just noise.
• The measurement is unreliable. Automatic counting on amateur video mixes up shirts, loses players who are out of frame and misses throw-ins. The displayed precision is higher than the real precision, which is the worst of both worlds.
• The figure does not say what to do. Knowing you win 42% of duels does not tell you whether to work on positioning, timing or choosing which duel to take. Three different sessions, one figure.

The three counters worth keeping

A useful counter has three properties: it covers a behaviour you really train, it can be counted without hesitation, and it can move within a month. Three examples, to adapt to your sport and your theme of the moment:

1. The situations you are working on this week. If the theme is building from the back, count the build-ups attempted and the build-ups completed. Nothing else.
2. What you concede on one specific point. Set pieces conceded in the last thirty metres, turnovers in the first third, balls lost in transition. Just one.
3. One occurrence of collective behaviour. Out of ten turnovers, how many times the team reacts within six seconds. This is the counter with the most visible effect on an amateur group, because progress comes quickly there.

Counting without fooling yourself

Home-made counting has a known enemy: you find what you are looking for. Three rules are enough to keep it in check.

• Define first. Write what counts as an occurrence, in one sentence, before the first match. Otherwise the definition drifts over the season and the figures stop being comparable.
• Count afterwards, on the video. Live, you miss half the occurrences and remember the spectacular ones. This is the one real contribution video makes to counting.
• Always the same person. Two different observers produce two series that cannot be compared, even with the same definition.

The figure says where to look, the footage says what to fix

This is the right way to link the two, and it works at every level. The counter shows that you lose seven restarts out of twelve; it does not say why. The answer is in the seven sequences, which you have to watch again, and most of the time two of them are enough to reveal the common pattern.

Taken the other way round, it avoids the opposite trap: spending an hour watching footage without knowing what you are looking for. The viewing method is detailed in how to analyse a football match.

When to go further

The moment to add complexity is easy to recognise: when a decision depends on a figure you do not have. Until then, adding indicators adds data entry time, with no gain in understanding.

And if you want a buying criterion: a statistics tool earns its price the day it makes you change an exercise. A good-looking report does not earn it.