Forum Discussion

Kahvijuoma's avatar
Kahvijuoma
Rising Newcomer
10 days ago

Three Eliminator statics, how do they work?

I have been wondering about this for a while.

What determines amount of points player gets?

At the "tournament leaderboard screen" (after the tournament), my friend usually gets a few more points than me. Even though he managed to more goals than me, I usually get a higher number of total  points due to having more assists.

What determines amount of RP we get per win?

Is it based on the combined current RP of our players compared to the opponent team's combined current RP? Or is it based on the rank (silver, gold, diamond) required RP or lowest/highest player RP from both team?

What determines amount of RP we lose? 

Same as above, but for lost.

What determines a team's rank position on the overall team leaderboards?

It seems to be more than just the points, as we see teams with fewer games played ranked higher than us. There must be a different math involved in team rank position.

Bonus question: What is different between local/global leaderboards?

Even when browsing "local" leaderboards, we see teams and players from the USA (just going by their ping of over 100ms in our matches), while our team players are from Europe (Finland to be exact). This suggests that the term "local" may not be "geographical".

 

  • All they want to/ are able to tell us is beating better teams = more RP and losing to worse teams = way less RP.

    I asked a few years back and it would appear only a certain few people on the team know the exact formula and if everyone did, there would be issues with users attempting to cheat/exploit the system so they choose not to make it public.

    RP is responsible for most of the leaderboard calculations - I assume overall wins make up most of the rest (sad to see teams ranked better just because they play more, even with vastly more losses).

    Local and regional is more fine tuned for NA where the team is situated - I believe teams/players outside the continent get grouped together using a slrt of hybrid local/regional model