How does team average rank works exactly?

(DB Genome editor) #1

So recently they added the average rank of each team in the final scoreboard in Ranked, but how exactly is it calculated, and more importantly is it taken into account in the balancing of the teams? You would expect well balanced teams to have identical or very close averages…

I’ve seen several cases where the average seemed off and while in most cases it could be explained by the odd “unranked” who’s hidden rank could be really far from the visible ranks of their teammates, there are a few cases where all ranks were known and yet the average doesn’t seem to correspond.

The +/- figures are the delta with respect to the indicated team average. Now unless ranks below average have more impact on the average than ranks above, this team’s average should have been one rank higher at Silver Recruit.

Same idea here, I would have expect this team’s average to be at least Silver Agent (which would be the “natural” average) or even higher if we consider that a single low-rank player might not drag down they team that badly…

Note that there we no parties in these team, only 5 singles, so the average should not have been affected by that (and in any case I would expect parties to bump up the effective team average, not reduce it :wink:)

(Mustang) #2

Most likely:
Sum of teams Elo / Number of players on team = Average Elo per player -> The rank this corresponds to

It’s to give players an indication of the average team skill, but in reality because it’s abstracted it’s quite possible that they seemingly bear no relation.
Also I’d expect average team rank to not take into account group skill boosts, whereas obviously they would be used for balance, which only serves to exacerbate the gap.