Who wants to see something a little more grounded than pure theorycraft
also something different than talking about what an apparent deity Pika is? I know some peeps here are a big fan of numbers, so I figured something like this could catch a few fishies:
I'm not the most reliable source when it comes to statistics (it's been a while since I took that class lel), but if I were to break down a bit about what's going on in the article (Warning: May involve a few arbitrary math terms):
- What Mr. Author is doing here is taking the win-rates of characters pre-buffs/nerfs and post-buffs/nerfs and using them to calculate a value known as the
p-value. What this p-value is doing (I think) is checking to see if the data we have is significant/reliable enough to use as evidence to back up a claim we might have. In other words, if the p-value of two sets of data is 4%, then there's a 4% chance that our data isn't reliable enough to use to make correlations or connections. Naturally, we want there to be a pretty big chance that our data is useful, so the limit our p-value will usually be 5% in most scenarios.
To take an example from their data:
's change in WR between pre-patch and post-patch jumped from roughly 40.5% to a 45.14%! Since our p-value is well below 5% (and 1% for that matter), this suggests that our data is indeed useful/reliable enough to make the deduction that - yes, Falco benefited greatly from his buffs.
On the flip side,
's change in WR between pre-patch and post-patch dipped slightly from 40.6% to 40.4%. However, our p-value is well above 90%, which is more than a big enough indication that our data is too random/not significant enough to make the argument that Marth's buffs were useless and possibly detrimental. Instead, we're left wondering why Marth's win-rate slightly dipped; it could be from the buffs somehow, but it could also very well be that Marth players as a whole just got worse, or that his WR dipped after some of the better Marth players flocked over to Lucina, or any other potential reason the mind can think up.
You can even use the public player database to check to see if any characters you might be curious about benefited/got hurt from their buffs and/or nerfs respectively:
https://statsmash.io/blog/public-query-sql
I'm probably mincing what the article's saying horribly, but I did find it interesting, and could be a new avenue to take this thread in.