Hello, dear reader! Many of you may recall me as the guy that used to write both short, cryptic, exceedingly dumb devblog posts AND long, dramatic, still exceedingly dumb devblog posts!
But alas, I have not done so in a while, especially the latter. Indeed, Cass has taken the torch from me in hilarious yet informative devblog content (hi cass ur gr8 I enjoy all ur posts), where have I been?
Well, I would love to tell you all that I have been quietly biding my time. Waiting. Working on my Magnum Opus of a devblog post for months turning into years, wanting to wow you all.
That's entirely not what happened it's just most of what I've worked on is either not interesting enough for a post, or too spoilery for one. But still, we'll just pretend that's what's been happening.
ANYWAY! We are not here to talk about me, oh no. We are here to talk about the Grandest of Events. A Tournament like no other. Something so Amazing you'll hardly believe your eyes!
That's right. Today's devblog post, after much waiting and anticipation, is about...
not THE LABRADORRA GRAND TOURNAMENT OF B00TY!!!
(Seriously guys we posted about that on April Fools two and a half years ago, I don't think it needs more followup than that.)
"Wait, if it's not that... OMG! Are you talking about some kind of actual tournament you're running for us, the loyal fanbase of Reborn?" - I hear you ask?
Also no. And come on guys stop interrupting me I'm trying to be dramatic.
To explain what this post is truly about, we have to rewind back a bit. You see, as many of you may be aware from other devblog posts, the team (especially Cass and Perry I appreciate you both very much!) have been putting huge amounts of work in optimising a lot of our code, and one part of that was a lot of structural changes to AI to get everything in top working condition for E19. And that has been going wonderfully! But it comes with a problem.
How do you test such changes? There are so many ways the AI could go wrong, and the devs can only try so many things to try and see if it works.
Well, Perry had a simple yet brilliant answer to this - Whipping up some code that can make the AI trainers fight eachother for testing purposes. That way we can run a bunch of battles, look over the debuglogs, single out any dumb choices the AI made and fix those! So much more efficient than just hoping one of us stumbles on the AI doing something weird when we play.
But, let me explain how Perry and I are different people. You see, Perry is... I think the technical, scientific term is 'Productive'. This was a brilliant idea that solved a big testing issue!
But me? When I saw this, I knew that I could take this innovation and instead use it as an excuse to be deeply unproductive. As is my style, by taking it and using it to do something fun and interesting but also not particularly useful!
That's right. The Tournament in question was an impromptu Round Robin where every single trainer in the game battled every other one!
No I haven't seen that Elo World youtube video where someone did this for Pokemon Red & Blue I have no idea what you're talking about all my ideas are entirely original.
So buckle up! This is gonna be a fun and goofy ride but there's also gonna be probably more statistics and spreadsheets than you expect!!!
Now, to begin, lets explain the setup. We did three runs of this tournament, in three distinct formats (although the third is special for reasons that will be apparent). In ALL formats, all pokemon were set to Level 100, although movesets were retained. This specifically detriments that one random FEAR trainer in the Desert so he underperformed compared to how he should've, but also fuck that guy that's what he gets for running FEAR.
1) Everybody battles Everybody else in a Single Battle with NO Field.
2) Everybody battles Everybody else in a Double Battle with NO Field.
3) League Trainers only (That is: Gym Leaders, Elite 4 Members, & the Champion) fights Everybody but specifically on their preferred format (i.e. Julia fights everybody on Electric Terrain Single Battle, Shelly fights everybody on Forest Field Double Battle, you get the idea.)
The first was just the initial idea and attempt at this, just to see what happens. The second was an extension when I remembered double battles exist. And the third was very much more an interest piece of like 'Hey, lets see how the leaders perform in their best case scenario' because I thought it would be super interesting to compare and see how that looks!
Now, a few methodological points for anybody that is interested: I know I really should have run all of these two or three times each because there's rng involved and stuff, but also everybody fighting everybody else is in the ballpark of a million battles which, surprise surprise, actually takes quite a while to run! So you'll have to deal with the fact these were only run once so, as in any good pokemon tournament, there may have been some RNG-caused upsets!
Second, outside of some specifically interesting cases, I will be talking in terms of a percentage winrate instead of number of wins just because these tests were done as we were working on some postgame trainers so there were slightly different numbers of trainers in each format, but not enough that I feel like it skews the results enough to rerun things!
And that's that! We ran this, it spat out the result of every single fight, then I spent a few weeks messing around in a very large spreadsheet and got the data organised nicely and analysed!
So, some last bits of intro logistics to this concept and then I can finally like, actually start the main content of this post! I will be splitting this into two main sections. In Part One I'll talk about the general results, specifically in formats 1) and 2), pointing out anything notable or relevant in the top 10, 50 or 100 that I think stands out! In Part Two, I'll focus in specifically on the Gym Leaders and talk more in depth about their performance in all three formats, and give some analysis about what that means about their performance in general!
And also, you should probably familiarize yourself with two shorthands I am going to be using:
Illegal - This should be self explanatory. A trainer has an illegal team if it has something it would be impossible for the player to get. Things like illegal EVs via PULSE2, PULSE mons generally, shenanigans with illegal movesets, and just the complete and utter fuckery we have in store for you with the top secret postgame Dev fights because we made the game and we don't have to follow the rules.
Bullshit - A trainer's team is Bullshit if it is bullshit. Like, sure, running 6 Psychic Type Legendaries (because fuck there are so many of those) with a Tapu Lele lead that all have optimised statspreads and movesets designed specifically to fuck you up on Psychic Terrain is technically legal and the player could also do that, but the AI doing it is still bullshit.
These shorthands will be useful particularly in Part One of the post because surprise surprise the top 100 trainers in Formats 1 & 2 are pretty strongly dominated by things that are Illegal or Bullshit (although maybe not as much so as you'd think!)
So, without further ado...
Part One: Fieldless
So, lets talk about Singles first. 6 of the top 10 are dev fights. (Including me! Hell yeah. 7th place with a 97% winrate. Take that 8th place Cass, with your paltry also 97% winrate (but hers is a lower 97% I won like 2 more fights than her get DUNKED ON Cass).
In fact, 1st, 2nd, 4th, 7th, 8th & 10th places are all dev fights. Shoutout to former dev Lia in the #1 spot showcasing her fight's utter shenanigans with a 99% Winrate (also, spoilers, the top 50 looks totally different for doubles but she STILL took #1 there too, albeit with a slightly lower winrate. The rest of us devs need to step our game up, clearly).
Most of the remaining spots in the top 10 aren't too notable. There's a lot of Bullshit. Heck, the specific team I used as the example to define Bullshit is in there.
Oh, and Lin is in there too somewhere but don't worry about that.
What I want to talk about is 3rd place. Because, somehow, despite all the Illegal and/or Bullshit fights we threw into postgame, 3rd place is something that takes place in the Main Story. In E19, in fact.
Lets talk a little bit about PULSE-Clawitzer.
This is PULSE-Clawitzer. You fight them in E19. Apparently it's pretty hard because the fight has a 98% winrate here, which is higher than every single other trainer except the two most difficult dev fights. So, y'know, good luck with that.
On the bright side, E19 isn't all bad. You also fight PULSE-Mr Mime and that fight is allllllllll the way down at 552nd place with a paltry 39% winrate (as a point of comparison that's pretty close to like... Julia. The very first gym leader with a team of NFE mons with terrible statspreads and movesets. Sure, it was hard like an hour into the game, but...) So at least you know that when you play E19 you'll just be able to relax and that fight won't cause any problems at all! ^~^
The rest of the top 50 is about what you'd expect. Dev fights, postgame bullshit, some illegal shenanigans. A few victory road battles made it in which is neat!
Also lets all give a round of applause to Baby's First Illegal team which is only really barely Illegal, which is the E18 fight against Solaris in Agate City, with the PULSE2 on his Garchomp - You'd hope he performed well given Singles with No Field is the format he actually fights on, and he doesn't disappoint, all the way up at 32nd place, the highest ranking trainer that is already in the current version of the game!
But still. Dev fights are very illegal. There's a lot of postgame bullshit here. Obviously PULSE-Clawitzer is illegal. Everything in the top 50 is Illegal or Bullshit or both.
Well, except for one trainer. The highest ranking 'Fair' (which I use as shorthand to mean neither Bullshit nor Illegal) trainer managed to make it to 46th place. This is pretty impressive - the next two Fair trainers aren't until 65th & 78th place.
With a 92% winrate in Fieldless Singles? I hope you're all ready for the gym battle against Saphira in E19!!!! Because she utterly destroyed the competition, miles ahead of the next highest winrate gym leader (surprisingly given it's singles and without a field, it's still Hardy at 85%!), and she even managed to beat the entire Elite 4 on winrate as well!
But we can talk more about that in Part Two.
Before that, though, lets look at how things are different in Fieldless Doubles!
Dev Team Lia still dominates at first place; her team is too good. The top 10 is once again 6 devs, although not the same 6 as before. (Alas I am the one that has been dunked on, in doubles I fell all the way down to 20th place and Cass kicked my butt and is now all the way up at 5th place)
Indeed, this theme is going to be apparent - While Doubles-centric teams can perform reasonably well still on Singles, a lot of the big hitters from Single battles seriously struggled battling in doubles. PULSE-Clawitzer fell all the way down to 21st place (below me! And I'm terrible at pokemon! Git Gud Clawitzer). PULSE-Mr Mime did slightly better than before but ultimately still very badly at 446th place.
Saphira really suffered hard though. She lost her spot as the highest winrate Fair trainer. That was taken by some randome generic Victory Road fight which isn't even a double battle and, honestly, I love that. This fact might be my favourite thing in E19, especially because that trainer jumped up like 50 places in the leaderboard from Singles. You gotta appreciate the hustle.
But alas, Saphira's bad day doesn't end there, oh no. She not only lost her spot as the best Fair trainer, in doubles she didn't even manage to be the best Gym Leader - Taking to the format like a fish to water, she was beaten handily by Hardy and, although by a slimmer margin, also by Adrienn!
Indeed, looking at gym leaders it is clear just how much harder doubles is if you don't have a team built for it. After those top three, the next highest ranking gym leaders are Charlotte & Radomus, beating the fights that came much later than them in game by merit of being built for the field. Even Aya did really well despite being earlier game because of her doubles team. As for the reamining Doubles leaders, Shelly still struggled because of how early on her fight is, and lets just not talk about Amaria until later because yikes.
Lin still comfortably sits in the top 20 but like, don't worry about it.
But, speaking of later, I think it's time to get into the reeeeaaaal fun. Have you ever wondered how well balanced the Gym Leaders are? Ever wondered which ones rely on their field most, and which ones least? Which ones are unusually strong or weak for their point in the game? Or do you just want to know how scared you should be of what's coming in E19? Well, then do I have the devblog post for you!
It's this one. This is the post. We're moving to part two now.
Part Two: League Members
So, before we begin, lets talk about the elephant in the room. You are going to notice a distinct lack of Corey and Kiki in the upcoming analysis.
Did I forget to run their fights? Yes.
Did we change the earlygame story a bit for E19 so they are no longer Gym Leaders because we all decided we really didn't like the contrived excuses to not give you a badge? Also yes.
With that cleared up, lets start from the top! There is going to be A LotTM here so I will be putting the individual analysis of each leader in spoiler tags, handily labelling them as we go along! Although this is just so this post is slightly easier to scroll through and look back over I will be writing all of these assuming you've read them all in order.
And with that, my dear readers, this most gargantuan of devblog posts comes to a close. I wish you all a fond farewell, and hope you are-
What's that, you say? I missed someone?
I have no idea what you're talking about. I already told you why we aren't including Corey and Kiki. So who else could you possibly-
Ah, right. I mean, I told you not to think about it but, I guess. It's not a big deal or anything but sure, whatever.