top 20 teams

I attempted to make a worldwide mathematical ranking of teams

Rank ------ Name ------------- Elo -- unc - maps - score - oppo
1 Gambit Esports ---------- 559 - 108 - 49 - 76% - 350
2 Vision Strikers ------------ 549 - 150 - 22 - 77% - 376
3 Envy ---------------------------- 441 - 119 - 37 - 65% - 318
4 100 Thieves ---------------- 439 - 117 - 35 - 57% - 364
5 Acend ------------------------- 434 - 117 - 32 - 50% - 437
6 Sentinels -------------------- 427 - 113 - 35 - 60% - 361
7 G2 Esports ----------------- 410 - 114 - 38 - 61% - 335
8 SuperMassive Blaze -- 361 - 132 - 27 - 48% - 373
9 Team Liquid --------------- 345 - 207 - 10 - 60% - 301
10 Bren Esports ------------ 311 - 189 - 18 - 89% - 35
11 Rise -------------------------- 294 - 194 - 9 -- 44% - 328
12 TNL Esports ------------- 292 - 275 - 4 -- 50% - 347
13 XSET ------------------------ 278 - 205 - 9 -- 44% - 319
14 Natus Vincere ---------- 272 - 210 - 8 -- 38% - 338
15 Giants Gaming --------- 267 - 172 - 12 - 50% - 274
16 F4Q -------------------------- 238 - 132 - 29 - 31% - 384
17 FaZe Clan ---------------- 227 - 227 - 8 -- 38% - 318
18 Paper Rex ---------------- 225 - 142 - 26 - 50% - 231
19 Guild Esports ----------- 218 - 227 - 9 -- 56% - 192
20 Oxygen Esports ------- 195 - 229 - 7 -- 29% - 330

NOTE: the ratings are based ONLY on Berlin and Berlin qualifiers, NOT LCQ. That's why teams like Liquid are low.

Also, the rankings aren't perfect especially when datasets are small. For example, TSM probably shouldn't be in the top 20, but their only two results are a 2-1 defeat to Envy and a 2-1 defeat to Rise, so the program didn't know what to do and just put them right below rise (I manually removed teams that didn't win any matches, sorry TSM). The draw column is useless because valorant doesn't have draws (removed it nvm).

unc = uncertainty, oppo = average elo of the opponents they played

Actually cool take on ranking interesting to see too... Stats ranking wise has reasonable takes.

Krü? Giants? VK?

obviously some of the teams should be slightly shuffled, with some teams missing, but overall it's interesting that the list is relatively reasonable. neat stuff

Why are rise that high up when they finished 5-6th and xset that low when they had a second place in qualifiers and fourth place in Berlin qualis just barely missing out

what does it look like with minimum of 20 maps. or even 10, there are 7 teams with fewer than 10 maps in your model

Cool project, I like the idea. Of course the data is skewed cus of the time frame, but you mention that already. Quite interesting.

Surely Liquid > G2 and SMB?

this list is actually decently accurate, your formula is cracked

Keep up the good work mate

Lmao

This is really cool! However some of these teams don’t deserve to be as high as they are. Keep up the good work though!

Really interesting, thanks for posting this

Good work my friend