Matchmaker
#1
Posted 18 June 2018 - 07:11 PM
#2
Posted 18 June 2018 - 08:34 PM
#3
Posted 19 June 2018 - 03:12 AM
Edited by Monkey Lover, 19 June 2018 - 03:13 AM.
#4
Posted 19 June 2018 - 03:43 AM
Monkey Lover, on 19 June 2018 - 03:12 AM, said:
All that would really be accomplished with that is a bunch of games where the people who deserve to be in 3-5 have their torsos taken from them, violently. If a reset happens we need a slightly different skill system that rates people more aggressively. There are people that make me wonder if I even legitimately belong in tier 1 at my skill level, or if there should just be a tier 0 of sorts.
#5
Posted 19 June 2018 - 04:03 AM
Monkey Lover, on 19 June 2018 - 03:12 AM, said:
I suggest that new season = tier wipe, so puggernauts couldnt get tier 1 for 1 month, but toptier player can reach it with 20-30 700+ matchscore games farming potatoes with actual meta.
#6
Posted 19 June 2018 - 04:31 AM
Tiers need a revamp altogether rather than a wipe or we'll just end up having the same system later on down the road. (Assuming of course MWO is still around in that timeframe).
#7
Posted 19 June 2018 - 04:31 AM
And, is was deathly quiet on the coms. No one wanted to talk at all; and, I guess that's OK as well.
I'm pretty sure this is all a low population issue and if every more people start to play, MM will function better.
#8
Posted 19 June 2018 - 04:39 AM
I'm not saying matchmaking is good, but even perfect matchmaking wouldn't change the fact that the game is designed to snowball. Lopsided kill scores in themselves can't be regarded as evidence of bad matchmaking, they are simply the expected result of a high ttk environment where focus fire advantage feedbacks into itself.
Obviously there is other more valid evidence of bad matchmaking, such as looking at the historical stats of each team and seeing how they compare, and obviously mismatched teams will also lead to more predictable stomps in favor of the better team.
But what is absolutely false is any expectation of less lopsided end results when teams are more even, the only difference would that the winner couldn't be either team depending on who gets the advantage, the matches themselves would still snowball just as hard and often as the do now.
Karl Berg even suggested that his data indicated that the snowball effect increases the less standard deviation of skill there is within a team, in other words teams with internal large skill disparities leads to less lopsided results. If the goal is to reduce "stomps" we should let all tiers play together and just balance the total stats of each team, with the good players on both teams carrying the bad ones. Those matches wouldn't be better, but they would look better with more even kill scores.
#9
Posted 19 June 2018 - 06:41 AM
https://mwomercs.com...-tiers-and-psr/
Clarification One: I completely forgot why we had the Tier system introduced other than to separate new players from vets. I really...Really can't stress how much I would like to see a ranked system of some kind or a ranked que and a casual que.
Clarification Two: While this was great and all, the player base has stagnated, we're not seeing any influx of newcomers, while the rest are still making a slow and steady climb to Tier 3/2/1. What we have currently is that people are breaking this plateau from the sheer amount of games played, disregarding the W/L aspect.
That last little schtick of one trick ponies as well...I mean, you see it all the time in League of Legends where some dork will take only a handful of champions, play the ever living crap out of it until they're amazing, and then make the climb from division to division. The same can be applied to this game, where you could play only mechs that just dumpster dive people like the current metabringer.
Clarification Three: Great that this takes into account for mostly wins and somewhat of individual performance, but again, it looks like that overtime, people will just go to Tier 1 if they play enough.
Clarification Four: I created an alt for to test this, and while you can jump a tier, it will not be in one game. Unless you're pulling in like...an 800 match score and you are probably not using a trial mech.
Clarification Five:
Quote
Okay, no it doesn't. The PSR rewards you if you win the game or lose. You could be a complete derp doing an average of 100/150 damage a game with 10 assists, an average of 10-15 damage to individuals, so long as you got one or two other players pulling in 600 damage with a couple kills and the same assists. So long as that happens, your PSR will increase. Maybe not dramatically, but on average, you'll still go up a tier at some point if this is consistent enough.
A good team or game, you'll see some damage spread with players that pull 350-500 damage max, with about a 300-350 match score.
Quote
Lets face it. In Tier 4/5, you're either getting new players, vets with new accounts or you have some of the worst pilots ever. Tier 3, is more or less the same with the exception that trial mechs are nowhere to be found. Tier 2 and 1 are folks like myself that have been around for a solid minute. Even the "small adjustments and macros" focus more on the individual rather than a team.
Perfect examples of team play is more akin to Community Derpfare, assuming you're dropping with a unit, you'll have a drop caller who wants the team/lances to go to specific grids or to complete a certain objective. IE: Take down A gate, leave B gate alone, Jump over C gate and hit their stupid LRM boats in the back.
In quickplay, you're lucky to get anyone, no matter which tier, to drop command, and more often than not, folks ignore that drop commander because they're too busy throwing down rain on cover or bird watching off into the far distance. Teamwork is almost non-existent in QP, and you're at the mercy of 23 other individuals of whether you win or not.
The reason why I quoted all of this is because while I get the intention behind it, it's more or less a temporary solution that's beginning to unravel at a gradual rate.
Edited by DrtyDshSoap, 19 June 2018 - 06:43 AM.
#10
Posted 19 June 2018 - 06:50 AM
-and the way you guys in QP really want it to work, at it's core.
right?
many (to the point of almost all) matches could be A LOT closer with even a tiny taste of comms and teamwork, yet everytime it goes the established QP-way. "ignore what that guy says and do the opposite" to "you're not my real dad" or "if you 11 guys rotate to the right - eff you, I go left".
so... 12:3. yes, I'm annoyed by it - but I've come to accept that in QP, a good match is a rare diamond. best way around it is to just. not. do. QP.
Edited by Teenage Mutant Ninja Urbie, 19 June 2018 - 06:55 AM.
#11
Posted 19 June 2018 - 11:23 AM
Sjorpha, on 19 June 2018 - 04:39 AM, said:
I'm not saying matchmaking is good, but even perfect matchmaking wouldn't change the fact that the game is designed to snowball. Lopsided kill scores in themselves can't be regarded as evidence of bad matchmaking, they are simply the expected result of a high ttk environment where focus fire advantage feedbacks into itself.
Obviously there is other more valid evidence of bad matchmaking, such as looking at the historical stats of each team and seeing how they compare, and obviously mismatched teams will also lead to more predictable stomps in favor of the better team.
But what is absolutely false is any expectation of less lopsided end results when teams are more even, the only difference would that the winner couldn't be either team depending on who gets the advantage, the matches themselves would still snowball just as hard and often as the do now.
Karl Berg even suggested that his data indicated that the snowball effect increases the less standard deviation of skill there is within a team, in other words teams with internal large skill disparities leads to less lopsided results. If the goal is to reduce "stomps" we should let all tiers play together and just balance the total stats of each team, with the good players on both teams carrying the bad ones. Those matches wouldn't be better, but they would look better with more even kill scores.
https://en.wikipedia...hester%27s_laws
Lanchesters laws of combat, Particularly his exponential law apply here.
If you want less stomps there is little that would help other then respawns.
Edited by Gierling, 19 June 2018 - 11:23 AM.
#12
Posted 19 June 2018 - 01:26 PM
Gierling, on 19 June 2018 - 11:23 AM, said:
https://en.wikipedia...hester%27s_laws
Lanchesters laws of combat, Particularly his exponential law apply here.
If you want less stomps there is little that would help other then respawns.
MW:O itself prooves that respawns do NOT help with stomps, at all . Doesnt change a thing re "being stomped" whether you die quickly once to a superior Team, or 4 times in quick succession .
10 user(s) are reading this topic
0 members, 10 guests, 0 anonymous users