How to Build a Shortlist from StatsBomb Data Without a Data Analyst

Event data does not give you a shortlist. It gives you a very big table. The skill is knowing what to ask it.

Most scouts who get access to StatsBomb data do the same thing in the first week. They sort by goals, or by xG, and get the same names every other club is already looking at. Then the data goes back in the drawer, because nobody has time to learn a new discipline on top of the job.

You do not need an analyst for this. You need a method. The one below works in a spreadsheet, and once you have done it once, a position takes an hour or two. Every step uses a left back as the example, and the numbers in the example are invented, so no real player is being rated here.

1. Write the role in one sentence

Do this before you open any data. Describe the player you need in one sentence, with a verb in it:

"A left back who gets us up the pitch on the ball and still defends his own box."

Two verbs, two jobs, so two things to measure. If you cannot write the sentence, the problem is not the data. It is that nobody has agreed what the club wants. Ask the coach three questions: What should this player do when we build? What should he do in the first five seconds after we lose the ball? What can he absolutely not be?

Their answers are your brief. Everything after this step is just checking players against it.

2. Pick four to six metrics, and map each one to the sentence

Every metric has to answer a part of the sentence. If it does not, drop it. A long list of stats feels rigorous and mostly produces noise.

For the left back sentence:

• Gets us up the pitch: progressive passes and progressive carries per 90 minutes.
• On the ball, under pressure: pass completion when under pressure. StatsBomb flags events that happen under pressure, which makes this possible.
• Delivery: passes into the box per 90.
• Defends his own box: tackles and interceptions per 90, and aerial duel win rate.
• Works without the ball: pressures per 90.

Two warnings. First, "progressive" is defined differently by different providers, so read the definition your data uses and stick to it. Second, event data counts actions. It does not tell you whether the action was the right decision. Keep that in mind for step 6.

If you want to practise before you touch your club's data, StatsBomb publishes free open data for selected competitions. It is enough to run every step in this article.

3. Filter before you rank

Ranking comes second. Filtering comes first, and it removes most of the nonsense.

• Minutes played. A common rule of thumb is a floor of around 900 minutes, roughly ten full matches. Per 90 numbers built on 200 minutes are noise, and one lucky cameo can put a player at the top of any table.
• Position actually played. Not the position on his profile. A winger who played eight games at left back is not a left back.
• Age, contract and level. Decide your age range and the leagues you can realistically sign from, so you do not fall in love with a player you cannot get.

Write the filters down. When someone asks in January why a player is not on the list, you will have an answer.

4. Compare like with like

This step separates a useful shortlist from a list of whoever plays for the best team.

A full back at a club with 65 percent possession will pile up progressive passes whatever his quality. The same player in a team that defends deep will look worse. And per 90 numbers from a lower level league are not comparable to numbers from a stronger one.

Two simple rules handle most of it:

• Rank inside the same league first. Compare players against others in the same competition and position. Then look across leagues by hand, for the top few only.
• Check the team around him. If a player's numbers are great and so are his whole team's, be careful. Look at how he compares with his own team mates.

You will not remove the bias completely. You only need to stop it from deciding the list for you.

5. Rank by percentile, then read the shape

For each metric, work out where each player sits compared with the others in the same position and league. That is a percentile: 90 means better than 90 percent of them.

Averaging the percentiles gives you a sorting order. It does not give you the answer, because an average hides shape. Here is an invented example:

• Player A: progression 91, delivery 84, defending 40, pressing 77. Average 73.
• Player B: progression 70, delivery 66, defending 74, pressing 71. Average 70.
• Player C: progression 88, delivery 60, defending 71, pressing 80. Average 75.

Player A looks great on the first half of the sentence and weak on the second. If the coach said defending the box is not negotiable, A is out whatever the average says. Player C fits the sentence best: strong at getting up the pitch, solid in defence, and the weaker delivery is not part of the brief. A table sorted by progression alone would have put A on top.

Do not build one magic score you cannot explain. If you cannot explain a number in one sentence, nobody in the building will trust it.

6. Take the top eight to video, and write down why

Data gives you candidates. It does not give you a decision.

Take the top eight and watch each of them in two or three full matches, not highlights. You are looking for what the numbers cannot show: positioning, decisions under pressure, how he reacts after a mistake, and whether the progressive pass was the right pass.

For every player, write one line on why he stays or goes. Then cut the list to three to five names and add, for each one: how he fits the role, two risks, and whether he is realistically available. Our post on how to write a scouting report that actually gets read covers the format. If your targets are free agents, the out of contract shortlist guide shows how to narrow the list by contract date.

The mistakes that ruin the list

• Sorting by one metric and calling it a shortlist.
• Ignoring minutes played.
• Comparing players across leagues without adjusting for level.
• Treating the data as the decision, instead of the reason to go and watch.
• Starting from zero every season, and losing the reasons behind every name.

Spreadsheet or Athleads

You can run this whole method in a spreadsheet, and it works. It gets slow when you repeat it for six positions, with three scouts, and the reports live somewhere else.

That is the part Athleads is built for. StatsBomb data is already inside the platform, so you can search players by position, age, nationality and club, compare them side by side, and move the ones you like onto a board. Reports sit next to each name, so the reason a player is on the list does not get lost, and AI helps you draft the reports and the shortlist so you do not start from a blank page. Same method, less copying between files.

If you want to see how a professional club works with it, read the FC Union Berlin case study. And if you are not sure you have outgrown your spreadsheet yet, five signs is a quick test.