Auto-translated
What specific points constitute the "correct direction of research"?
For me, it primarily consists of the following points:
1) Only the opinions of highly skilled players should be considered (i.e., the sample should include reports from the later stages of tournaments), or it should be a survey of opinions (even if private) of players who have demonstrated skill in the past.
2) For this specific question, the course of games on specific templates, under certain rules, should be studied separately for each template, with red rush allowed.
3) Specific games, specific game situations should be considered; no statistical studies should be used as arguments at all. Statistics can be a reason to investigate a question, but it should be investigated under a microscope with specific examples.
And I keep repeating – this is the path that Blizzard uses to balance its games. Their example is worth considering as a model.
Yes, they have Battle.Net, where they collect statistics on millions of games. And, according to their technical director, they collect all sorts of statistics, how many units and which ones are built, how the results of games depend on the units built, and they hope that this will help them someday, but when it comes to changing something in the balance, they carefully study the opinions of the best players in the world ranking and watch recordings of specific games. (!)