A football database (1 Viewer)

Philosorapter

Well-Known Member
I think this is the footballing equiverlant of the holy grail.

A while ago I tried to show how a computer chess engine works and the rationality behind it, not knowing if football is like chess, a zero-sum game. There is a lot of theory out there that states it is not. I disagree.

I believe chess and football works from the same principles even though the herustics are different.

However, the other part of what the chess world has been building this last decade or so is a games database.



There is though so much more to a game or players database when linked to an engine.



I don't understand why the football world hasn't got something like this.

I just thought it would be interesting to share it with everyone.
 

Last edited:

skyblue1991

Well-Known Member
6fccee56ece62575acd5f858fd5d73d5.jpg


Sent from my E5823 using Tapatalk
 

Nick

Administrator
Depends what you mean by varied. They all have a point on a field of play which can be measured.


Yes but a certain chess piece has the same weapons (moves they can do).

Not every right winger has the same, they might be fast but not be able to cross. Etc
 

Philosorapter

Well-Known Member
Yes but a certain chess piece has the same weapons (moves they can do).

Not every right winger has the same, they might be fast but not be able to cross. Etc

Exactly, advantages and disadvantages on a field of sport exist objectively and are conditioned by the situation on the pitch.
 

Philosorapter

Well-Known Member
Objects all have properties. These properties can be objectively measured. It doesn't matter how many properties each object has, as long as you can measure them against other objects which have the same properties but in varying degrees on a playing field.
 
Last edited:

Skybluefaz

Well-Known Member
The situations in a football field are so fluid though, I think something along these lines can be done although it's not as cut and dried as in something like baseball sabremetrics where you can measure lots of things like right handed batters v left handed pitchers performance.
 

Philosorapter

Well-Known Member
The situations in a football field are so fluid though, I think something along these lines can be done although it's not as cut and dried as in something like baseball sabremetrics where you can measure lots of things like right handed batters v left handed pitchers performance.

Going to repost these again and give a link to the book. The appreciation at the beginning of the book is especially worth a read.

http://www.simardartizanfarm.ca/pdf/-_Lasker_s_Manual_of_Chess.pdf

I believe this is one of the first books published that contains the ideas of positional play over combinative play.

"Pawns are the soul of chess" - Philidor

This old chess proverb for me holds the secret of the true nature of a lot of team sports linked with the creative skepticism of breaking down systems as explained in the appreciation at the beginning of the book.


IMG_0024.PNG IMG_0025.PNG IMG_0026.PNG
 

shmmeee

Well-Known Member
Could be an interesting machine learning problem. If you can get a birds eye view if enough games, split the pitch into a grid, assign each time step a state based on positions of players and ball, you could probably predict most effective positions and moves likely to lead to goals.

Getting the data would be the issue, I bet ProZone have it.
 

robbiekeane

Well-Known Member
So is it a database? Or some kind of model you mean?

Like against this starting 11 with x attributes in each position here are the most effective players and plays etc
 

Gazolba

Well-Known Member
Chess pieces never get tired, frustrated or injured. They also never age, don't have to train and have no concept of loyalty.
Anyone who thinks chess and football are similar is a first-class idiot.
Netball has more in common with football and even that isn't much.
 

Users who are viewing this thread

Top