Automated data gathering - an idea

[rotfront1918]

Well, a lot of untested hypothesis floating around:
- how to break red army: Arm, AP or manpower
- Blitzkrieg: encircle vs. spearhead
- winter strategy
- importance of leningrad and moscow
etc.

You could run some serious statistical analysis to test this stuff, however the data is too scarce and unmethodically collected. But there is some turn logging (save folder), from which you could automatically extract information into a excel spreadsheet. Some data is missing there (OOB ie.), but its still nice stuff - especially the production data. You could for example infer the displaced/destroyed factories from soviet data, and then look how this influences the probability that the axis wins or looses (controlling for other variables) if you have this data from multiple games.

Does anybody know how to write a litte program that automatically extracts the juicy parts of the turn logs into a spreadsheet? (like python does for websites...) everybody playing a GC could easily run the program and get the numbers (massive time-series data base!), and share them within the forum.
Statistical modelling could then answer a lot of questions... [dont worry, there will be still enough room for debate[;)]]