> I do have the search settings setting for winning and all of that positive stuff, but I still get paired up with people who don’t know when to stay back or who can’t get kills at all =( All while I get 20 kill’s and few deaths.
>
> This idea I think could work, I really think it should be tired out, it could always be removed if it wasn’t popular.
The thing is, as far as the search settings in Reach go, they aren’t restricted enough. Halo 3 ranked did a very good job at matching skilled players with skilled players and less skilled with less skilled. You see, the search restrictions of a playlist can be set to as high as you want. The problem is, Reach hasn’t them very high, even with your own restrictions on and thus you end up with less than good matches. For these reasons your idea is pointless, it unnecessarily restricts players away from the playlists without really doing anything to improve the quality of the matches (when you would still have the same, unrestrictive, crappy matchmaking system).
> And I could see how not being able to have a 4 player party would be bad, so I guess that would be a bad idea somewhat, but having a party with randoms going up against a 4 player party is unfair.
>
> Because the 4 player party are most likely a group of friends who player and would communicate with each other, a group of randoms doesn’t always communicate =(
The thing is, in a competitive playlist, it’s really nothing but your own fault if you get bad teammates when you don’t take the time to find a decent team to play with. So no, it’s not unfair in a playlist designed around matches that are supposed to be about winning, to have a full team going against a bunch of randoms. Lone Wolves only had the party restriction because it was FFA.
> Also I guess a more exact and better KD to go by would be KD per game.
>
> ps TrueSkill is good, but it does have flaws =( I gotta go though . .I have work this afternoon, I will check back on this thread later today.
You have to ditch the idea of K/D. There are much better statistics around, K/D is something that only shows a direction, but should not be used to determine is a player good. I’m not going to pretend TrueSkill is flawless, but it really is much less flawed than you think. After all, as a system that consistently evolves and uses probabilities, it becomes more and more accurate as you play. After hundreds of games, TrueSkill doesn’t lie. If you want to match players based on skill, TrueSkill is certainly the best choice. On average, in 4v4, TrueSkill can make a very accurate estimation on player skill only after 46 games. The only reason people don’t like TrueSkill is because it’s too honest. If you get locked at 35, that’s how good you are.
However, don’t think the Halo 3 version as an accurate presentation of TrueSkill. Bungie slowed the whole system down to be able to turn it to a progression system. In real TrueSkill, things are much more fluid and your real skill level is found much more quickly and your rank doesn’t lock in the sense that you can’t become better and progress your TrueSkill anymore, it only locks if you truly aren’t becoming any better.
So, TrueSkill truly is all you need to have a good matchmaking system. You just need to make it strict enough to not mix bad and good players.