Data Official Smogon University Usage Statistics Discussion Thread, mk.3

Status
Not open for further replies.
Something similar to what I did with OU should have happened with the non-usage-based tiers which were affected by Pokebank (that is, Anything Goes, Ubers, Doubles [OU / Ubers] and LC), but I don't see that reflected, so I'll have to figure out what happened there.
So... it looks like there were no non-bank ladders for Ubers / Doubles / LC? Is that correct?
 
Is it possible to get usage stats on teams? Like most common teams? I'd imagine the higher you get, the more you start seeing similar teams. I'd mostly be interested in the 1825, 17xx (for non ou) ranks
 
MagikaripIsOP no, I don't have any intention of providing that level of detail. It's not competitively useful (most common will still only be like 1 in 1000 at most), so it's just an opportunity to get lazy with teambuilding.
 
Since pidgeotite and steelixite have been released yesterday they will technically fall under this month's usage stats, just with a pathetic amount of usage. Will this be used as an opportunity to have them immediatly be passed down to the lower tiers (where we all know they'll end up) or are they going to be kept in OU for another month until we have a full month of usage stats for them?
 
So FYI, processing this month's logs is going super slowly, and I'm not exactly sure why. I'm currently copying the whole month of logs off of the Smogon sim server and onto my local machine, and I'm going to try to process them locally, which should be a bit faster. I'll keep everyone posted with regards to my progress.
 

erisia

Innovative new design!
is a Site Content Manager Alumnusis a Forum Moderator Alumnusis a Community Contributor Alumnusis a Top Contributor Alumnusis a Smogon Media Contributor Alumnus
What I think might be a good idea is making the code for analysing the stats available to a small group of people (and general principles available to the public) so that we can try and decentralise the stats process a little bit. I sincerely appreciate the constant effort and excellent work of Antar's that allows us to even do this usage based tiering approach in the first place, but I do wonder if it's a bigger task than we should reasonably expect one person to do in their spare time for free. If nothing else, if we had say a member of each tier calculate their relevant stats, they could use their computers to work on each tier in parallel and they could make stats available as soon as possible. The stats page has grown bigger and bigger over the past few years with official OMs, legacy tiers, and even random metas like CC1v1 getting full stats now, so I reckon it's as good a time as any to look into this process further and make it more transparent in general.

There's also the fact that we have literally no contingency plan if Antar has to take some time off doing stats for personal reasons lol.
 
Last edited:
I know the data is a lot and so processing it takes very long. However, tier changes occurring after 10 days into the month isn't very effective.
Oh, it's not going to be 10 days. They're STILL transferring and I STILL can't run the stats directly on the server.

Stats are just not going to happen this month until Zarel can fix whatever's going on with PS that's making disk performance so poor.
 

Disjunction

Everything I waste gets recycled
is a Site Content Manager Alumnusis a Top Social Media Contributor Alumnusis a Community Leader Alumnusis a Community Contributor Alumnusis a Tiering Contributor Alumnusis a Contributor Alumnus
Oh, it's not going to be 10 days. They're STILL transferring and I STILL can't run the stats directly on the server.

Stats are just not going to happen this month until Zarel can fix whatever's going on with PS that's making disk performance so poor.
Is it at all possible to simply calculate just usage stats for the only three tiers (ou, uu, ru) that are actually reliant on said usage stats? Like, I doubt anyone is going to be upset if we stopped calculating battle spot/gen 1 random battle/cc1v1 or even gen 7 lc/dou/ubers if it meant our resources were going toward keeping our schedule on track for our lower tier development.
 
Is it at all possible to simply calculate just usage stats for the only three tiers (ou, uu, ru) that are actually reliant on said usage stats?
I don't think you have a sense for how bad the situation is. If I try to process OU logs on PS, those scripts get killed for exhausting system resources. In copying logs to my own machine, I'm maybe 10% complete, after running this for a week.

Excluding randbats (the only non-usage-based tier that sees a significant enough number of battles to make a difference) might allow me to get the usage stats out a week or two sooner, but as it stands right now, even if I focused just on OU, we wouldn't be seeing usage stats this month.
 
I know this is a stupid question and not really the place to ask it. But could we, like, let Beedrilite be used in UU already? It's been released for some time now and it would be expected to drop turn this month.
 
I know this is a stupid question and not really the place to ask it. But could we, like, let Beedrilite be used in UU already? It's been released for some time now and it would be expected to drop turn this month.
Soon as usage stats come out, it'll probably be dropped.
 

Zarel

Not a Yuyuko fan
is a Site Content Manageris a Battle Simulator Administratoris a Programmeris a Pokemon Researcheris an Administrator
Creator of PS
I don't even know whose job it is to decide such things. I assume unless someone says otherwise, Antar will use the sampled stats and I'll make a tier out of it?
 
Status
Not open for further replies.

Users Who Are Viewing This Thread (Users: 1, Guests: 0)

Top