AXR|Kroos

18 posts

AXR|Kroos banner
AXR|Kroos

AXR|Kroos

@tk8_bs

RPI 23 | ML Engineer NYU| FinGPT Brawl Stars CN ML Apr 25 — Monthly Top 8 (Player) Feb/Mar/Apr 26 — Champion (Coach) Coach/Manager at Ace Xero

انضم Ekim 2021
107 يتبع222 المتابعون
Bounty Hunters
Bounty Hunters@bhbountyhunters·
ESTAMOS NO GRUPO A DA BRAWL CUP BERLIN! 🔥 Qual é o "grupo da morte" pra vocês? 👀
Bounty Hunters tweet media
Português
15
3
143
11.3K
AXR|Kroos
AXR|Kroos@tk8_bs·
@bobbyBS_ TRUNK THE GOAT. See you in the Brawl Cup🥰
English
0
0
4
1.1K
bobby
bobby@bobbyBS_·
first time since july 2020 im first in NA
bobby tweet media
English
59
39
1.3K
35.9K
AXR|Kroos أُعيد تغريده
svegou
svegou@Svegou·
Special bundle for Ace Xero 🩵 #BrawlCup | #BSC26
svegou tweet media
English
2
2
53
1.7K
AXR|Kroos
AXR|Kroos@tk8_bs·
@Lobb_bs @fikssnewss It's single-elim and random. We don't have a lower/higher bracket in CN mainland. It's literally chosen by lot through a meeting... We qualified from the hardest group in Feb and March.
English
0
0
3
102
LobbBS
LobbBS@Lobb_bs·
@fikssnewss Did Ace qual from lower bracket or is the seeding random?
English
5
0
0
978
FiKsS' NEWS
FiKsS' NEWS@fikssnewss·
Toxic Lotus lost to Ace Xero again... Has Toxic Lotus lost its momentum? Or is Ace Xero truly the strongest link in the Chinese region?
FiKsS' NEWS tweet media
English
4
2
92
6.8K
AXR|Kroos
AXR|Kroos@tk8_bs·
@Noim156 Just won another one MF. I will post more on X this week so more people get to know about us. We want more connections to other teams.
English
1
1
8
1.2K
Noim
Noim@Noim156·
@tk8_bs I put it in this tier because it was 3-1 against TL.
English
1
0
1
970
Noim
Noim@Noim156·
チームランキングティア表【全地域ver】 ガチ目に1時間半考えた結果こうなった
Noim tweet media
日本語
42
21
1.5K
498.4K
AXR|Kroos
AXR|Kroos@tk8_bs·
Been cooking a side project: Predicting & simulating Brawl Stars drafts Data = BS API (scrims) + APK unpacking, meta by meta isolation Early ML models (MLP/Transformer) go beyond winrate stats — e.g. lookahead draft strategy Leave ur opinion below #BrawlStars #BrawlEsports #BSC
AXR|Kroos tweet mediaAXR|Kroos tweet mediaAXR|Kroos tweet mediaAXR|Kroos tweet media
English
1
0
12
891
AXR|Kroos
AXR|Kroos@tk8_bs·
@Ssugus__ nvm, my table is about the hit rate of predicting the next pick vs the actual pick. If your focus is more on overall match winrate, then that level of accuracy makes sense and can even go above 98%.
English
0
0
0
87
AXR|Kroos
AXR|Kroos@tk8_bs·
@Ssugus__ If it’s purely based on basic stats, the accuracy looks too good, unless the data overlaps with the real matches. I’m working on a small model with a Transformer backbone, where I use the real mechanics of each brawler as priors/weights. Would love to hear more about ur method.
AXR|Kroos tweet media
English
1
0
2
147
TRB Sugus
TRB Sugus@Ssugus__·
Basically what this does is getting wr% of each brawler vs each brawler and it gives the wr in total for one comp vs the the other. It will obsviouly depend on what data you have, in this case is scrims data from worlds teams
TRB Sugus@Ssugus__

I’m working on a new draft prediction model The goal is to compare drafts in real time and analyze past matches to prepare for future games. Here’s the first test: predictions vs real results from the EMEA Monthly Finals 👀 Do you agree with the predictions?🤔

English
2
0
10
2.5K
CR Tensai
CR Tensai@tensaibs·
@Arotan_bs 0-2と0-6連れてるんやばいな
日本語
2
1
93
15.2K
あろたん
あろたん@Arotan_bs·
赤い方やったホンモノ
あろたん tweet media
日本語
1
1
75
31.9K