Roland, on 15 November 2013 - 10:47 PM, said:
Honestly, here's a better system than what they are currently doing:
1) Find a group of people that are currently trying to play, who have Elo rankings as close to each other as possible.
2) Then break up that group into two teams, as best you can, respecting the groupings of the players of course.
This will result in a match where one team may have a higher Elo rank.. But it will at least have a bunch of players who are as closely matched as possible. For high Elo games, everyone will at least understand how to playing the ******* game to some minimum level of competence. They will at least understand how to move their mechs and target enemies.
The idea that you can take a ****** player and a great player and equal two mediocre players if fundamentally flawed. That is not how the game works. It is terrible, and results in a terrible experience for everyone involved. Stop doing that.
That's the point of Grits N Gravys idea, and mine. Here's the problem though -
Are you willing to give up weight matching to get that? How much are you willing to give up on that?
Still need to split premade/pug Elo so those Elo scores are more accurate.
So make the primary matchmaking criteria Elo band though? How much tonnage mismatch are you willing to take to get that? I'm in 100% agreement with you by the way - absolutely no question. Personally I'd put up with 200 tons to get ~10% variation in Elo across both teams. If I can at least trust that everyone on my team isn't likely to have to disconnect suddenly because they've poked themselves in the eye while slamming their face on the keyboard to pilot their mech I can make plans around tonnage mismatches. Besides, the tonnage mismatch is as likely to be for me as against me and certainly not all mechs of a weight are equal. People will still pilot Awesomes, Battlemasters, Dragons and Locusts.
I'd even settle for 20% variance on Elo score, drop the Golden Master Race Elo score folks in with whatever match is showing the highest average and ignore their impact on the team 'total'. That way the odd 2600 Elo player can drop in a match with people playing at 2,000 and not have to pull some window-licker from the Underhive to gimp his team. I'm all fine with ignoring the Elo impact of highly ranked players, they're few and far between anyway. That and the difference between a really good player and a great player isn't nearly as profound as a good player and a poor player.
Tonnage mismatch I'll take in return for better Elo seating. Tighten that, loosen tonnage, split premade/pug Elo so Elo scores are far more accurate.
Would you be good with that?