atas76 / openengine

Football match engine
MIT License
58 stars 4 forks source link

Very interesting project #16

Open xr1fab opened 8 months ago

xr1fab commented 8 months ago

Hello, first of all, congratulations for this very interesting project (very rare in this field...). I started watching the code but I'm a little bit lost with all the acronyms (e.g. mv, ttl, mpn. pnw, etc.). Is there a sort of docs to look at? Thanks in advance.

atas76 commented 8 months ago

Hi,

Thank you for your kind words. I understand that the acronyms are very esoteric. These are meant more as experiments. Essentially, these are the evolution of the same thing: selected football match event data. I am planning to write some kind of documentation that would explain more the rationale and structure of these files, once I have also made myself some conclusions about a relatively final format, and how to move forward with the actual simulation.

For the time being, I would suggest looking at the commits and particularly the latest ones. I am also putting the names of these acronyms in the initial commit for each file, hopefully making a bit more sense. However, I think what matters is just the latest 'file format' (specifically mpn), as the previous ones are kind of 'obsoleted'.

I consider this repository as an aggregate project of various approaches towards creating a football match engine from scratch, so there are a lot of experiments there. The current idea is to create a representation of a football match with event data at this point as a basis for the simulation. This will evolve in two ways: through adding new data (ideally in a 'backwards-compatible' way) and interventions with logic to make up for insufficient data points or the structure of the data itself being elementary.

On a secondary note, and as a side effect, the invention of these formats make for a more human-readable way of representing a football match, while also being relatively structured for some rudimentary analytics. Again, there is a trade-off between the two, and this is just a byproduct of forming input data for the simulation, but might be useful for football analysts. I don't know if there is an actual need for it out there, but maybe we don't know it yet :) One by-project would be for example, conversion from football analytics data to a human-readable format and vice versa.

A lot of things are just in my head, and a work in progress, so I don't expect to make perfect sense, but I am happy to answer any questions and discuss things.