The Gaffer's System: How I Pick

People ask me how I pick. I will tell you. It is not complicated. But it requires discipline, and discipline is the thing that most pundits do not have.

I start with the form number. Not the table. The table lies. Six points from six can mean two comfortable wins or two last-minute scrambles. I want to know the expected goals. I want to know the goals conceded from set pieces, from transitions, from open play. I want to know whether the clean sheets are the result of good goalkeeping or poor finishing from the opposition.

Then I look at the squad status. Not the injury list on the official site, which is always behind. I look at the manager's press conference. I look at the training ground photos. I look at who was not in the squad the previous week without an explanation. This takes about twenty minutes per fixture and I do it for every match I am picking.

“I am not trying to be interesting. I am trying to be right. The system is what stops me from being interesting at the wrong moment.”

Then I look at the head-to-head record. Not the last ten years. The last three years, and only under the current managers. Football is managed by human beings with tendencies and preferences, and those tendencies show up in how their teams play against specific types of opposition. A high defensive line manager who faces a team with a fast striker gets exposed in a predictable way. I am looking for those patterns.

Then I decide on the scoreline. I do not agonise over this. I pick the most likely result first, then I adjust it for the away side. Away teams score less. Away teams concede more. This is the most consistent statistical finding in football over the last thirty years and yet most people pick scorelines that treat home and away as equivalent. They are not.

Finally, I assign the confidence level. One, two, or three. If I have any doubt, I go down a level. Confidence three means I have looked at everything and I am not second-guessing myself. Confidence two means I believe the pick but I can see the argument the other way. Confidence one means the pick is there but I would not be surprised to be wrong.

I have been doing this for two seasons. My hit rate last season was sixty-eight percent on correct results. My exact score rate was sixteen percent, which is the rate you would expect from a system, not from luck. The difference between sixteen percent and the random baseline of around eight percent is the system working. Not every week. Across the full season.

One more thing. I write down my reasoning before I lock the pick. Not after. This is important. Writing the reasoning after the result is a form of cheating that most people do not notice themselves doing. The reasoning changes depending on whether you were right. Writing it before forces you to commit. It forces you to know why you think what you think, not just what you think.

That is the system. GW1 picks go up Friday fourteenth August.