you are correct that, due to the nature of all stat based analysis, the results can be misleading
this analysis was very kill and trade focused so it favored duelists
and ofc it cant properly represent the complexity of the game; even when ignoring your point about the team not properly being considered (since its just gonna be impossible to separate a player from their team) it has huge issues
economy impact of kills, how much support util you were receiving during the kill (how good was it?), how many people were on the site, etc would all shift the weighting significantly - it just leads to an endless game of trying to get more aspects of the game into your stats... and in the end theyd probably still be wrong due to the whole team thing
your proposed solution wouldnt lead to a good result either, it would heavily favor players like ZmjjKK who have a risky and inconsistent playstyle, while a more consistent player like for example demon1 would look pretty mediocre
It also has the whole acs & kda already being linked problem that vlrs rating tries to tackle
At the end of the day the stats are great because he presents all his thoughts and reasoning for the stats, and its easy to figure out what aspects he overlooked and then decide how much you value them - no stats will ever beat the "eye test"