(Programming) Learnset Checking Project (new owner needed)

eric the espeon · May 2nd, 2011, 5:38:24 PM

On and off for the past year or so I've been working on a project to improve the accuracy of learnsets in various Pokédexes, and made over 2500 corrections to Smogon's learnsets for the first two generations. With my resigning as a C&C mod and greatly reduced activity it seems very unlikely I'll get to finishing it in any reasonable timescale. I have been given some Pokédexes and written macros to scrape others, together I've put together quite an impressive collection of information, and found massive, massive inconsistency between major Pokédexes. I believe it is important to correct this, and will be very disappointed if whoever takes over the project cares only for Smogon's own accuracy, though that of course will be their primary motivation.

Lets see.. what I have:

A 20,000ish line spreedsheet with all the moves that any source suggests gets a move in gen 3.
A 2100 line spreedsheet with all of the disagreements, sorted by how many sources agree.
A google document with a fair number of the total third gen disagreements on, and quite a lot of disagreements checked by wifiers/other people who have the games.
Large text documents in need of formatting with fourth gen learnset info for major dexes.
Maybe a formatted set of data for fourth gen, but I have not tracked it down yet. Seem to remember it being 30,000 lines. Or maybe 40,000.
Some stuff I used for previous first/second gen fixes to Smogon, could do with another look with regards to the work on NYPC I did here. And set stuff up for other sites to be fixed.
A method and practice with two generations.
The ability/willingness to talk you through what the things in the spreedsheets mean and how to generate new ones, assuming you will ask when you don't understand.

Now, I'm sure you all want to rush off and pick it up, but be aware this is not a small project, even if you move it to more of a group effort (which I strongly advise, but you'll need to oversee everything closely. With a large number of people helping it's too easy to let mistakes slip through otherwise.). Most of the data gathering and the majority of the most annoying text editing is done but there is a HELL OF A LOT of repetitive, boring, mind numbing text editing left. I spent ten hours in one sitting doing mindless text editing to scrape and clean just one dex worth of information. You will want to kill something. And it only gets worse in later generations. You need a high level of knowledge about where moves come from to avoid making a lot of mistakes, and to be seriously obsessive to stick with it long enough to make any impact. I'm no longer staff (or won't be once an admin gets around to demodding me), and probably won't be around to have a say it in, but anyone who can lead this project a significant way towards completion has has, in my opinion, done more than enough to be classed as a full contributor. And has my respect.

So, if you:

Are fairly obsessive/dedicated and don't have pressing time restraints
Care about the details. I don't want someone who thinks 98% accuracy is good enough, if there are errors they should be fixed.
Want to do something which will be helpful to huge numbers of users
Don't have a million other important projects to take care of already (looking at you here RD)
Are reasonably intelligent and either knowledgeable about oddities of learnsets, or willing to put quite a bit of time into learning/reading a lot of dexes/relevant things
Have a not-crap computer. Huge spreedsheets make my fairly new one struggle, they would kill an ancient one.
Want a shiny new badge (not guaranteed, but deserved)

Then post here/pm me/find me on irc/bully someone into giving you my msn and I'll talk to you then give the files you need.

Moo · Oct 30th, 2011, 10:19:51 AM

Just posting to say that this hasn't been forgotten. I've been plodding away at this over the months and am almost finished!

Here is the spreadsheet if anyone wants to see it. Also if anyone is interested in helping out, let me know, the more the merrier

Also credit to Nexus, EnKuChou and WildEep who also worked on the project

eric the espeon · Oct 30th, 2011, 11:48:37 AM

Great to see you're still making progress. Wish it was possible to give you an extra contrib badge for this, checking well over 1000 moves by hand is a gargantuan task.

Moo · Dec 4th, 2011, 11:41:14 AM

Hey, just an update. We're nearly done with this, but there's a section for the gamecube games and I don't have them :\

So if you want to help out and possibly earn a badge then contact me!

comatthew6 · Apr 28th, 2012, 4:37:02 PM

Ok. It's been almost four months since Nixhex and I finished the Pokemon xD Gale of Darkness moveset compilation and after Moo and co. finished the cartridge list. Can someone upload the google document, found below, please???

https://docs.google.com/spreadsheet/...COfZkfUP#gid=1

If this does not link to the document for xD, there is a tab on the bottom. If it doesn't work at all, please contact me.

eric the espeon · Apr 28th, 2012, 6:11:41 PM

the people with access need a list of things to change. one for moves to remove, one for moves to add. like the ones i did in this thread. you can make that (mostly) from what you've done as i understand it, but you'll want to compare it against a list of smogon's moves. talk to en. and talk to someone with access to smogon movepools (server level access i think? maybe some other tech admins.), you'll probably want an updated version.

comatthew6 · Apr 29th, 2012, 11:38:10 AM

Ok I will get on that soon.

May 2nd, 2011, 5:38:24 PM	#1
eric the espeon maybe I just misunderstood Join Date: Aug 2007 Posts: 3,695	Learnset Checking Project (new owner needed) On and off for the past year or so I've been working on a project to improve the accuracy of learnsets in various Pokédexes, and made over 2500 corrections to Smogon's learnsets for the first two generations. With my resigning as a C&C mod and greatly reduced activity it seems very unlikely I'll get to finishing it in any reasonable timescale. I have been given some Pokédexes and written macros to scrape others, together I've put together quite an impressive collection of information, and found massive, massive inconsistency between major Pokédexes. I believe it is important to correct this, and will be very disappointed if whoever takes over the project cares only for Smogon's own accuracy, though that of course will be their primary motivation. Lets see.. what I have: A 20,000ish line spreedsheet with all the moves that any source suggests gets a move in gen 3. A 2100 line spreedsheet with all of the disagreements, sorted by how many sources agree. A google document with a fair number of the total third gen disagreements on, and quite a lot of disagreements checked by wifiers/other people who have the games. Large text documents in need of formatting with fourth gen learnset info for major dexes. Maybe a formatted set of data for fourth gen, but I have not tracked it down yet. Seem to remember it being 30,000 lines. Or maybe 40,000. Some stuff I used for previous first/second gen fixes to Smogon, could do with another look with regards to the work on NYPC I did here. And set stuff up for other sites to be fixed. A method and practice with two generations. The ability/willingness to talk you through what the things in the spreedsheets mean and how to generate new ones, assuming you will ask when you don't understand. Now, I'm sure you all want to rush off and pick it up, but be aware this is not a small project, even if you move it to more of a group effort (which I strongly advise, but you'll need to oversee everything closely. With a large number of people helping it's too easy to let mistakes slip through otherwise.). Most of the data gathering and the majority of the most annoying text editing is done but there is a HELL OF A LOT of repetitive, boring, mind numbing text editing left. I spent ten hours in one sitting doing mindless text editing to scrape and clean just one dex worth of information. You will want to kill something. And it only gets worse in later generations. You need a high level of knowledge about where moves come from to avoid making a lot of mistakes, and to be seriously obsessive to stick with it long enough to make any impact. I'm no longer staff (or won't be once an admin gets around to demodding me), and probably won't be around to have a say it in, but anyone who can lead this project a significant way towards completion has has, in my opinion, done more than enough to be classed as a full contributor. And has my respect. So, if you: Are fairly obsessive/dedicated and don't have pressing time restraints Care about the details. I don't want someone who thinks 98% accuracy is good enough, if there are errors they should be fixed. Want to do something which will be helpful to huge numbers of users Don't have a million other important projects to take care of already (looking at you here RD) Are reasonably intelligent and either knowledgeable about oddities of learnsets, or willing to put quite a bit of time into learning/reading a lot of dexes/relevant things Have a not-crap computer. Huge spreedsheets make my fairly new one struggle, they would kill an ancient one. Want a shiny new badge (not guaranteed, but deserved) Then post here/pm me/find me on irc/bully someone into giving you my msn and I'll talk to you then give the files you need. __________________ For people who like storing things: The Box Reading and LC? LCF, LC Guide, LC Analyses Good channels: #littlecup, #C&C, #1v1, others And for SCMS editors: SCMS group ete on IRC. Goodbye Smogon. Good luck, was fun while it lasted. Last edited by Oglemi; Aug 18th, 2011 at 12:02:06 AM.

Oct 30th, 2011, 10:19:51 AM	#2
Moo Professor Join Date: Jun 2010 Posts: 1,447	Just posting to say that this hasn't been forgotten. I've been plodding away at this over the months and am almost finished! Here is the spreadsheet if anyone wants to see it. Also if anyone is interested in helping out, let me know, the more the merrier Also credit to Nexus, EnKuChou and WildEep who also worked on the project __________________ GP / UU QC, VM for check VM for a UU rate Last edited by Moo; Oct 30th, 2011 at 10:39:21 AM.

Oct 30th, 2011, 11:48:37 AM	#3
eric the espeon maybe I just misunderstood Join Date: Aug 2007 Posts: 3,695	Great to see you're still making progress. Wish it was possible to give you an extra contrib badge for this, checking well over 1000 moves by hand is a gargantuan task. __________________ For people who like storing things: The Box Reading and LC? LCF, LC Guide, LC Analyses Good channels: #littlecup, #C&C, #1v1, others And for SCMS editors: SCMS group ete on IRC. Goodbye Smogon. Good luck, was fun while it lasted.

Dec 4th, 2011, 11:41:14 AM	#4
Moo Professor Join Date: Jun 2010 Posts: 1,447	Hey, just an update. We're nearly done with this, but there's a section for the gamecube games and I don't have them :\ So if you want to help out and possibly earn a badge then contact me! __________________ GP / UU QC, VM for check VM for a UU rate

Apr 28th, 2012, 4:37:02 PM	#5
comatthew6 # fucks I give = lim(x-->0)[cos(x) - 1]/x Join Date: Jan 2010 Posts: 1,585 New Jersey	Ok. It's been almost four months since Nixhex and I finished the Pokemon xD Gale of Darkness moveset compilation and after Moo and co. finished the cartridge list. Can someone upload the google document, found below, please??? https://docs.google.com/spreadsheet/...COfZkfUP#gid=1 If this does not link to the document for xD, there is a tab on the bottom. If it doesn't work at all, please contact me. __________________ * BarackObama gives channel operator status to comatt6 * V0x get down on knees # of facebook statuses user: jumpluff has liked: 7 My purpose is clear. My targets - doomed.

Apr 28th, 2012, 6:11:41 PM	#6
eric the espeon maybe I just misunderstood Join Date: Aug 2007 Posts: 3,695	the people with access need a list of things to change. one for moves to remove, one for moves to add. like the ones i did in this thread. you can make that (mostly) from what you've done as i understand it, but you'll want to compare it against a list of smogon's moves. talk to en. and talk to someone with access to smogon movepools (server level access i think? maybe some other tech admins.), you'll probably want an updated version. __________________ For people who like storing things: The Box Reading and LC? LCF, LC Guide, LC Analyses Good channels: #littlecup, #C&C, #1v1, others And for SCMS editors: SCMS group ete on IRC. Goodbye Smogon. Good luck, was fun while it lasted.

Apr 29th, 2012, 11:38:10 AM	#7
comatthew6 # fucks I give = lim(x-->0)[cos(x) - 1]/x Join Date: Jan 2010 Posts: 1,585 New Jersey	Ok I will get on that soon. __________________ * BarackObama gives channel operator status to comatt6 * V0x get down on knees # of facebook statuses user: jumpluff has liked: 7 My purpose is clear. My targets - doomed.