Senario
Smash Ace
- Joined
- Jul 1, 2013
- Messages
- 699
In short, I should waste time on numeric analysis when it isn't my forte or field, gotcha. I'm just saying popularity and relative strength are as heavily correlated as you think. Strong characters attract more players, but the difference in balance is not accounted for. Meta Knight is vastly stronger than some of the rest of the cast and they have little chance of fighting back.A very cursory glance at the data shows that changing the view to filter to top 16 or top 8 has a minimal effect on the shape of the balance curve. It affects the particular a bit of which characters fall where, but it doesn't really make a big difference for either game how the distribution looks. Since the totals are so much bigger for "all" and therefore the data will have less randomness introduced by irrational overperforming individual actors, if the distributions are about the same, I feel like it's the best data set to use for this particular question of analysis. A sophisticated weighting algorithm would obviously be better, but that is pretty non-trivial to fabricate with the tools available to me. Capps has a team working hard on a data project for smash 4 that, if successfully integrated into the community, will allow us to conduct significantly superior analysis of the new metagame compared to what we can do for the old games, but I do think our current data is good enough to draw significant conclusions.
What you have done here is take my analysis of some data, say I'm wrong, and then tell me that reality is the complete opposite of what the data suggests. You then offer an anecdote about amsa and Axe who are only two players with Axe in particular clearly being a ridiculously huge outlier of the sort that data analysis always attempts to minimize. You then somehow insist that Brawl low tiers are amazingly bad and have no real hope without huge skill gaps despite this contradicting both all the data I've ever seen including what I've offered and the actual experiences of almost all real Brawl players. I truly don't even know where you get the last part since you do not seem to be a big Brawl player so it seems unlikely that you have extensive personal experience on this front, and you aren't presenting data either.
I'm sorry, but your argument here is not convincing at all. At best your experiences disagree with my data, but my experiences back up my data so it's a wash on the experiences with my data seeming to be the point that stands out. I do not want to bring even more Melee vs Brawl into the smash 4 forums than necessary, but this is a topic about Sakurai and his team's competence which is highly relevant to our view on the future for smash 4. My data may be imperfect, but you haven't really offered anything to counter it. If you have your own data, numeric analysis, or mathematical models that demonstrate that my analysis (which does not even suggest Melee is without merit; its balance curve does have some advantages) is wrong, I'd be very interested. Just telling me I'm wrong isn't going to do it; I need evidence.
I'd love to waste time on providing you statistics had I the resources or time. But what I can say is that your data for the differences honestly doesn't seem like any substantial difference. In my field that much difference is so little that effectively it should fall within the standard deviation of any normal dataset for popularity.