League of Legends XXXII: Cupcakes and Kittens and Squirrels. Oh my!

[Zen Master]

Acromos...correct me if I'm wrong, but one of the questions you asked about your statistically generated champion selection algorithm was, "Would this work?"

You are entirely correct.

Therefore, it is PERFECTLY legitimate for people to respond by saying, "No, it will not work, because the statistics behind the model don't work out," and follow up with a statistical discussion.

Well - yes and no. If you decide to ignore the fact that I point out in my initial post that this is a thought experiment, and not something I'm seriously suggesting would be a great idea ..... then of course it's entirely valid to attack the premises.

However, if you do not decide to ignore key portions of the post, then attacking a hypothetical premis is pretty much the same as stating 'no, you cannot thought experiment that!'

If that's not the discussion you wanted to have, this probably isn't a model you should bother thinking about. And if you're so disinterested with respect to the quality of the idea, you probably shouldn't get defensive when people say it's a bad idea.

I say myself, in the initial post, that it isn't a good idea. How - please explain this to me - how is it defensive that I repeat what I already stated?

What matter of communication is open to me that would not be defensive?

To answer your questions myself...
-Would it work? Your later qualifiers defining "work" are "would it give better teamcomps statistically?" and "would it lead to higher win rates?" I think the human factor leads to the answer "no", for both questions. If you work off tournament games, you don't have enough data to have a working model, and misleading factors like the success of one particular team with a particular strat other players can't perform as well will dominate the data. If you work off a larger sample, you introduce people who don't know how to play (which is most of the playerbase), and their success with a given champion won't reflect the theoretical matchup, but rather the relative inability of the players. Nor will the theoretical matchup be a reliable indicator of what the player USING this tool can do in that matchup.

-Would it behave similarly to human-chosen bans and picks? Not really, because humans choose champions largely based on their personal expertise, and this isn't a factor in the proposed model.

It's actually no problem to account for personal expertise - though you're entirely right, my initial post said 'tournament games' which would not include that subset, unless of course you happen to be a tournament player.

I agree that a system like this wouldn't work. Just not for the entirely same reasons - but ... I feel that there is something more than statistics going on in a game like this. And it's not just 'player skill', which has been suggested - that is entirelty quantifiable.

One thing the model doesn't allow for is picking offensively. It would know how to counter, basically picking only defensively.

There is a quality of draft and ranked games I think the model would handle well, tho. Out of any (random) five people who put together a team of champions, there is almost always someone who picks whatever they like without consideration of anyone else - leaving the rest of the team to somehow fix the combo. If there are two of these types, that becomes impossible, often. A model like this would counter that.

However, such a model would also always try to follow the statistically most succesful meta - something which leaves it open to anyone who spots a weakness in that. It will be predictable, which is always undesirable.

Eh - nevermind =)