Edit: As a huge point, it seems like I have made a mistake regarding the Troq vs Rook matchup - even tho I have crossreferenced it around 30 times throughout the weekend I wrote this, I ended up seeing that Rook had 27 vs 26 as a score vs Troq, while it was actually the other way around. I still hold my conclusions about Rook being a solid pick against the carousel.
@mysticjuicer has been hard at work for many years, adding results into my favorite spreadsheet: The Historical Matchup Chat
I have been looking at the numbers a little, and there are some very cool stuff in there! I did do one of these topics a little while back with some hysterically wrong assumptions, but I think these ones are actually correct this time, and not wrong.
So mysticjuicer has already pointed this out before, that the historical matchup chart isn’t a representation of what a perfect player making perfect plays, but a collection of all players with different skill levels making all types of plays ranging from beginner level ones to perfect ones. So the chart doesn’t tell us necessarily what is the optimal matchups, but it can give us a pointer of some misconceptions we might have about the metagame.
So, what is the meta in Yomi in this day and age? The 20XX of Yomi shows that DeGrey, Geiger, Troq and Zane are concidered the best characters in the game, they see the most play, and has the highest winratings in the game. They all go to the same circus, and ride the carousel where, supposedly, Zane wins every matchup except against Troq, Troq wins every matchup except against Geiger and Geiger wins every matchup except against Zane? I don’t really know where DeGrey fits into all of this, but I think he is just concidered overall very good, with low risk plays and high damage in general.
I hope that throughout looking at these numbers fairly carefully I am able to challenge some conceptions about what characters are good and what characters are bad.
[details=Playrate numbers!]1. zane…2424
2. troq…1930
3. sets…1863
4. arg…1805
5. degrey…1769
6. grave…1768
7. rook…1633
8. geiger…1547
9. valerie…1365
10. oni…1293
11. gwen …1190
12. ven…1182
13. gloria…1157
14. lum…1109
15. quince…1045
16. menelker…1039
17. midori…981
18. perse…952
19. bbb…949
20. jaina…807
[/details]
And, indeed we see Zane and Troq dominate. Zane represents with 500 more matches recorded then Troq on a solid second place. The two other culprits are respectively on 5th and 8th place in representation. Setsuki is only 67 matches away from surpassing Troq, and Arargarg is sitting swell on fourth. But of course, representation is not everything, winrate is more important.
[details=Winrate numbers!]1. zane…5.6
2. troq…5.5
3. geiger…5.3
4. arg…5.3
5. lum…5.2
6. sets…5.1
7. degrey…5.1
8. bbb…5.1
9. midori…4.9
10. rook…4.9
11. oni…4.9
12. grave…4.8
13. quince…4.8
14. gloria…4.8
15. valerie…4.6
16. menelker…4.6
17. ven…4.6
18. perse…4.5
19. gwen…4.5
20… jaina…4.4
[/details]
Firstly, I’d like to point out that the balance of the game is pretty decent! Of course this doesn’t point out some numbers that we will get to later, where we will see that some matchups are pretty bad. But overall, having these numbers be so close as they are, is pretty good. You could easily see similar games end up with a higher spread of numbers, ranging into the 7-3 spectrum. It is true that only 8 out of 20 characters can sport an actual winrate - so we are 2 characters away from having a 50/50 spread, which would be ideal.
So, on the surface, does this challange or confirm our idea of the carousel meta?
-
The 20XX: Zane and Troq are the most played characters, and they also have the highest winrates. Geiger shares the third place with Arargarg, but Arargarg has around 250 matches more played than Geiger. DeGrey has a positive winrate, but both Lum and Arargarg has a higher winrate - which to me is surprising. BBB also has a positive winrate, but is also considered one of the worst characters in Yomi, so how does that fit in? He has some of the fewest matches played, so possibly only used as a reactionary counterpick to Troq? Also a correlation between BBBs winrate and Rooks high playrate?
-
Oddballs? Arargarg and Setsuki sports a high winrate and a high playrate, but they aren’t concidered as decent options in the carousel. Lum is the oddest oddball, which sports the 5th highest winrate, but the 14th lowest playrate - but back in the day, he was concidered a fairly good option against Troq, and due to his damage out of combat, could ride the carousel pretty good. He does have a fairly high difficulty rating of playing in general, so that might explain it a bit. Midori is also one of the few characters that has less than 1000 matches played, but is among the 10 highest winrates in the game.
-
Profound sadness? Jaina has both the least representation and the lowest winrate. After that, Persephone and Menelker are the next on the list with the lowest numbers combined in both representation and winrate.
So, where to go next? My thinking is that we should delve into the individual matchups of the top 5 characters in both winrate and representation, and include Geiger and DeGrey, since they fall out of top 5 in one each - to see what is to find, if anything. Since it only includes 1 extra character, I’ll also include BBB into the mix.