Ok this is my last straw this Vlr rating needs to go. How tf is skuba over reduxx in rating when reduxx carried them so hard.
https://www.vlr.gg/196161/m80-vs-oxygen-esports-challengers-league-north-america-w10
4
Robbery
posted in
General Discussion
Comments:
Threaded
Linear
Here are some advanced stats:
reduxx had 6 of his kills immediately traded compared to zero from skuba
reduxx had 10 deaths with even player differentials (most important fights) while skuba had 4
reduxx also took slightly more fights against eco rounds, and a higher percentage of skuba's deaths happened when there was a large player differential and had less impact on the round. Also 4 more assists for skuba
essentially, both players did really well, but a higher percentage of skuba's fights had a significant positive impact on their chances of winning the round.
KAST is widely considered to be little more than a gimmick stat.
The rating doesn't use the KAST number whatsoever. Things that affect KAST also affect the rating but in different ways. For instance, trading only affects the weights of the fights taken in order to determine how important that fight was to winning/losing the round.
I actually looked at correlation between rating and other stats, and the rating is most closely correlated to the kill and death +- (per round as all things are), and not as much K/D ratio, KAST, ACS, or other stats. The correlation to KAST is actually pretty flimsy (other stats have a stronger correlation with KAST). VLR's rating has a stronger correlation to kills and deaths than ACS or thespike's rating.
Thats not to say that there aren't significant upgrades that could be made to the VLR rating in future updates. It is currently balanced on stats from roughly 2021 to 2022, but patch 5.12 made the assist numbers of KAY/O and the controllers (who were already significantly on top of the assist charts) jump into the stratosphere. It didn't impact that specific OXY vs M80 map too much, but assist scaling has gotten a little wonky in 2023.
Discord