

Elo Is Coming: What To Expect
#41
Posted 17 February 2013 - 12:52 AM
#42
Posted 17 February 2013 - 01:12 AM
I wonder how the games will be for me after the next patch. I have nightmares of games consisting only of 3Ls, A1s, K2s and DDCs.
#43
Posted 17 February 2013 - 01:16 AM
#44
Posted 17 February 2013 - 01:16 AM
#45
Posted 17 February 2013 - 01:34 AM
#47
Posted 17 February 2013 - 01:48 AM
#48
Posted 17 February 2013 - 01:59 AM


#49
Posted 17 February 2013 - 02:09 AM
It prevents perverse incentives from arising (if damage is rewarded excessively, people will tend to boat LRMs for damage rather than sniper builds, for instance). As long as kills, k/d, damage and assists are contributing to the overall goal of winning (which generally they are, cutting up disconnected 'Mechs aside), a player with better 'stats' is going to have a higher ELO than a player with lower stats. An added bonus is that non-damage-based contributions to the team (distraction, tanking, scouting) are also allowed for, which are very difficult to otherwise quantify.
My worry is that MWO's matchmaker will suffer from a phenomenon known as ELO hell. In essence, it is a situation where an overly conservative ELO implementation prevents players from reaching their 'true' ELO. This is particularly common in team-based games such as LoL, but even Starcraft 2 suffered from a form of ELO hell in 1v1 because the developers had not properly modelled the effect of their implementation on the bottom 10% or so of players.
In a similar vein, Halo 3 had a famous group of players who went 120+/1 W/L from fresh accounts in a ranked team playlist but could not move beyond the low 20s in a skill ladder of 50 (so just below average) because the matchmaker would not allow them to be matched against players with significantly higher ELO from whom they could extract significant skill points. In that instance the matchmaker was fine for individuals, but not for teams.
#51
Posted 17 February 2013 - 05:59 AM
I do hope Elo pans out like your predictions, this is pretty much what I'm expecting as well.
#55
Posted 17 February 2013 - 06:07 AM
#58
Posted 17 February 2013 - 06:13 AM
Vassago Rain, on 17 February 2013 - 06:10 AM, said:
It needs to measure some kind of success, that you can improve/maintain. What it measures isn't important, just that it has a way to tell who's good from who's bad.
Damage should be a better measurement of individual performance than win/loss that depends on the performance of the team.
Edited by Mackensen, 17 February 2013 - 06:13 AM.
#60
Posted 17 February 2013 - 06:28 AM
Mackensen, on 17 February 2013 - 06:13 AM, said:
Damage should be a better measurement of individual performance than win/loss that depends on the performance of the team.
Yeah lets put loads of really really bad players in with a few people who headshot a stationary mech every time they see it......Great thing to base matchmaking on!
14 user(s) are reading this topic
0 members, 14 guests, 0 anonymous users