63 replies to this topic

[Why Run]

WarHippy, on 16 September 2014 - 12:26 PM, said:

If I was just to look at that pic and use it to measure skill and nothing else the only conclusion I could come to is that your entire team sucked including yourself. You have one mech on your team that could be a trial mech as far as I can see as does the other team so blaming it on "new players" seems rather silly. You are just making excuses and blaming what you assume to be new players for performing in a way that was only slightly more terrible than how you performed.

Dude, my entire premise is that they are new, and the MM algorthim to match them up is faulty. Most of us "sucked" when we were new. This game has an absurdly steep learning curve, as well as a pretty steep tech curve. It takes quite a while to get elited mechs. Well more than 25 games, more like 200 games. I sure did play like crap. I'm also currently "cleating" (easy cowboys, not literally), the match makers 1:1 w/l "rule," so perhaps I'm being "punished" back to the rule.

But the idea that the MM attempts to pick one team, and the match them using averages (not more appropriate median ELOs) is simply illogical. First, a statistical average is useless, it reveals nothing about the composition of the group as averages are so easily skewed by outliers. I think perhaps the problem is the game lacks enough of a pool to pick players of similar ELO, so it's almost immediately widening the search parameters. That's not really addressable by changing much though. However, I still see the problem going further than that. For the same reason using average tonage would be useless (would you rather have 12 - 60ton mechs or 6 100 tons mechs and 6 20 ton mechs?), using average ELO based off of one team, and then building a second team is not producing much by way of balance, especially among the lower tiers.

Certainly Paul can come back and say all those guys with less than 30 damage have played 2x as many games as I and it was just an "off game", but the numbers are so divergent between the team and between the two teams, that I don't believe it. (He certainly doesn't need to, and I apologize I'm not disparaging them, I'm suggesting they are new, and at a pretty significant disadvantage to get rolled by team FS and 5 kills - Guessing that guy is a little above average ELO huh?)

Edited by Why Run, 16 September 2014 - 02:07 PM.

[Why Run]

It's also my understand of ELO that it's used to match individual skill 1v1 to figure out their appropriate level of competition and group tournaments of like players (read, CHESS). However, here, it's been adapted to a 12 person group setting, with a huge band range to grab players to fill holes if (3 minutes?) passes without being filled, which entirely defeats the purpose of the system...

Furthermore, it arbitrarial pegs 25 match new players near mid-elo (recently adjusted slightly south of mid?), rather than at the bottom. Which increases the likelihood that a huge volume of new players gets tossed into the meat grinder instantly, which cannot be fun....

Edited by Why Run, 16 September 2014 - 02:15 PM.

[Boris The Spider]

New players in Miserys, Jesters and Direwolves?

[WarHippy]

Why Run, on 16 September 2014 - 02:06 PM, said:

Dude, my entire premise is that they are new, and the MM algorthim to match them up is faulty. Most of us "sucked" when we were new. This game has an absurdly steep learning curve, as well as a pretty steep tech curve. It takes quite a while to get elited mechs. Well more than 25 games, more like 200 games. I sure did play like crap. I'm also currently "cleating" (easy cowboys, not literally), the match makers 1:1 w/l "rule," so perhaps I'm being "punished" back to the rule.

But the idea that the MM attempts to pick one team, and the match them using averages (not more appropriate median ELOs) is simply illogical. First, a statistical average is useless, it reveals nothing about the composition of the group as averages are so easily skewed by outliers. I think perhaps the problem is the game lacks enough of a pool to pick players of similar ELO, so it's almost immediately widening the search parameters. That's not really addressable by changing much though. However, I still see the problem going further than that. For the same reason using average tonage would be useless (would you rather have 12 - 60ton mechs or 6 100 tons mechs and 6 20 ton mechs?), using average ELO based off of one team, and then building a second team is not producing much by way of balance, especially among the lower tiers.

Certainly Paul can come back and say all those guys with less than 30 damage have played 2x as many games as I and it was just an "off game", but the numbers are so divergent between the team and between the two teams, that I don't believe it. (He certainly doesn't need to, and I apologize I'm not disparaging them, I'm suggesting they are new, and at a pretty significant disadvantage to get rolled by team FS and 5 kills - Guessing that guy is a little above average ELO huh?)

Like I said you are making excuses. Paul explained as have others, but you choose to not believe it because the only explanation you can process in your brain is that it must have been a bunch of new players when that doesn't look to be the case. The MM is not the problem.