Canadian Arab wrote:

I actually thought we did have at least one person who works with the club on this stuff? Does Jaymes Monte (@ArabAnalytics on twitter) have some involvement? If he doesn't now, In think he did in the past.

My gut feeling is that the current level at which game and player data can be collected and analysed is pretty impressive. The thing about data is they show what happened, how often, and what the result of that thing was - i.e. the past. What is lacking, I think, is knowing how to use the data to change how we are going to play. Basically spending time during a training session focussing on one or two things (maybe small things) that the stats show worked for us or did not work for us. Believe it or not, training time is actually pretty limited and you need to maximise the benefits of everything that's done, and I can see what might be considered "minor" stuff like this being given a low priority.

I had a guy over here help me run a team years ago - he had a background in ice hockey. He kept stats on my players during games, on successful passes and whether or not they were forward passes, square, backwards or split a line, interceptions, challenges won and lost, shots on and off target etc. It was all very basic and honestly it didn't have much impact on how I did things or what I covered in my training sessions. But the one interesting thing he did talk about was how plays are developed in ice hockey. He said teams have maybe 3 or 4 different things that can happen when they have the puck in their own end. There are only 5 players on the ice and everyone has to be fully aware of where the other players are and what they're doing. Depending on where they win the puck, the near side winger either takes option 1 or option 2, and the other players are aware of his decision. They immediately make their moves in line with the choice he made. The puck then gets played to player A or player B. Depending on which player it goes to, the other players, who have already reacted to the first choice made of option 1 or option 2 then follow the plan for option 1.1 or 1.2, or option 2.1 or 2.2. The next pass is often the pass that is meant to create a chance, and depending on which of the two passing options is made, the other players then continue their movement to cover the likely outcomes of a saved shot or the puck going wide of the net on one side or the other. Within all that there's room for a skilled individual to do something unexpected, but it's not unexpected for his own team mates and if he decides to nutmeg a guy and drive at the net, the other players again know exactly where they need to be skating to.

His point was that our training sessions should have been run the same way that hockey training sessions were, for example the ball goes from the keeper out to a full back on the ground, and the rest of the team know exactly where to move to. The full back either plays it to the feet of the winger or inside to a CDM. Depending on which choice he makes, the rest of the team know exactly where they should move to because they already know what could happen next. If it takes 4 passes to get the ball into a position for a cross or a shot, it would have got there as a result of 4 consecutive "option 1 or option 2" decisions, i.e. 16 possible paths to get the ball to that point. It needs players to be able to think really fast and make the correct run or movement at each step of the move, depending on the choice made by the ball carrier. But you still need to leave room for a Ralph Milne, who decides to throw the rule book out the window and go on a mazy run.

If you watch the very best teams, I don't think there's much happens by chance. They move the ball around in pre-determined patterns, waiting for an opponent to fail to adjust their position correctly, and then they attack the error. But the entire team all have to recognise the point at which the positioning error occurs, and immediately make the correct runs or movements simultaneously. It's pre-scripted to a large degree.

I have no idea if that approach would work with players who are not among the best players in the world. Part of being a top player is processing speed and seeing what's going to happen before it does.

Yet another long post to say not very much. And I have no idea what xG means.

Edit: Ultimately I think this sort of an approach makes a team more likely to win and less enjoyable to watch.

 
I pretty much agree with most of what you said. My gut instinct is that the main benefit of the stats is about player recruitment i.e. finding a player who fits the system, who has thus far not been recognised. Sometimes a coach might be aware of an untapped market e.g. Ange Postacoglu’s knowledge of the Japanese market.

In terms of on the field performances, you really need granular data from a large sample size. Anecdotally, we all think we are experts. I mean, how many times did Davie Dodds actually score a goal with his head or foot as opposed to the ball hitting his knee, arse, shin, heal etc. But he was still in the right place. Or how many times did Paul Paton ever use his left foot? Did anyone ever witness this?