Official Smogon University Simulator Statistics — August 2012
Hope everyone's excited--we're just one month away from our next usage-based tier update!
I've been a busy bee the past few weeks, playing around with a bunch of new analyses, and it's my pleasure to bring back today metagame and moveset analyses! There's still more work that need to be done on both--on the metagame front, I hope to be able to define "semi-stall" vs. "full stall" vs. "offense" vs. "hyper offense" vs. "balanced" based on my new stalliness metric; with regards to moveset analyses, I'm hoping to add teammate and check/counter data--but I'm hoping the work that I've done already will be enough to satisfy you for this month.
A new version of "1337" stats are on my to-do list, but--I'll be honest--it's not a top priority.
Anyway, that's my progress report. Let's get to the data!
Link to previous month's stats
Tier Change Projections
Tiers will be updated next month based on three-month usage. Since the last time I posted these kinds of projections, I found this post by X-Act, who is kind of my role model in life, which says that, based on data from only two previous months, you can predict usage for the next month by weighting the previous months' stats 20:4. So if I do that to predict September's stats--rather than just assuming that they exactly mirror August's, which is what I did the last time--we get the following predictions.
Check me on these--it's a new script, and I could have easily made a typo.
*Abomasnow would indeed be allowed in UU, since the UU council banned Snow Warning and thus would have removed Abomasnow from BL.
In the interests of not wasting pastebin space on the single Delibird set that some idiot decided to use in OU, I'm only reporting the breakdowns for Pokemon in each tier that showed up at least 10 times and contributed at least 0.1% to the usage stats. The "most common sets" thing is kind of a crock, as any two sets that are even slightly different (Jolly vs. Adamant, Fire Blast vs. Overheat) don't get counted together, but I am working on an improvement that will involve cluster analysis.
Also, as I mentioned in the OP, I plan on incorporating teammate statistics into this data at some point, as well as lists of top counters/checks for each Pokemon (based on how often a Pokemon is KOed or switched out when the supposed counter is sent out).
OU Suspect Test
If you're interested in the methodology behind my metagame analyses, read here and here.