What statistics would be helpful for policy questions?

[Cathy]

I think we can all agree that many potential debates could do well with some data to reference. Although the pokemon usage statistics have made it possible to define OU objectively, as well as other applications, some more statistics might provide insights into some of the current and future topics on this forum (and elsewhere). The good news is that Shoddy Battle actually saves a log of every battle that takes place (indexed by date), and it is possible to glean other data from these logs than usage of pokemon. Since the logs for the past few months are many gigabytes big, I am not able to directly upload the "raw data" (which has been requested by some people).

Keeping in mind the limitation that data has to be collectible by parsing logs, what sort of statistics would be useful for these discussions? Once we decide on some (and I write the scripts), we can even find the statistics for all previous months of ladder battles, since the logs are still archived.

 

[X-Act]

One statistic I'd be interested in is what ability each Pokemon had when used in battle. For example, there are statistics for Arcanine, but some of those were of the Intimidate variety, while others were of the Flash Fire variety. Same for Bronzong (Levitate or Heat Proof), etc. This is important because, in my view, Levitate Bronzong and Heat Proof Bronzong, for example, are two completely different Pokemon.

 

[Jumpman16]

If viable, I'd like to know the actual most popular movesets for a given pokemon. When people say "oh nobody uses Chain Chomp anymore" or "Boah has really declined in usage", I'm as skeptical as I am when I hear "what nobody uses Celebi or Snorlax when in actuality they are #16 and #26 respectively. Without usage stats a single person would be hard-pressed to agree or disagree, and I think the same goes for popularity of movesets.

 

[Aeroblacktyl]

I would imagine it's hard to do 'exact movesets' as there are many variations of a moveset, but this is of course, obvious. It's probably easier, I'm assuming, to list the top most used moves per Pokemon. Which was a statistic that used to be kept on the bots back in the early RS days..

 

[Mekkah]

I think it would be much easier to determine movesets if anything actually logged the teams being used, but that doesn't seem to be the case. Shame, because I agree it would be really handy.

We might also need similar techniques when testing for the UU metagame if we ever get around to it.

 

[Aeroblacktyl]

Also has the problem of logging complete sets if you win while only using 2 of the moves or 3 of the moves etc etc but I'm just nitpicking

 

[Jumpman16]

Have you decided to consider compiling some or any of these statistics, Colin?

 

[Cathy]

We will be able to have statistics regarding abilities and items, probably with the usage statistics at the end of the month.

I didn't do anything about movesets yet, however.