58 replies to this topic

[Fate 6]

I'd rather see some sort of chassis rating, though I'd still settle for weight class. The difference between my best and worst mechs is just too much variance for accurate matchmaking. Whatever algorithm PGI used to place people placed some people too low and some people too high. It placed me a few games shy of Tier 1, even though when I play for real I'm in the top 80 pilots that participated in MLMW - even the pilots on my team (QQ Ragequit Mercs) that I regularly outdid in competitive matches placed higher. I assume after some time it'll even out, but even after 5 games today I barely saw the bar move, so I have to assume that it'll take a few hundred drops for people to see significant progress in a direction.

I'm curious to know what algorithm was used because it would have been nice to see some weighting towards mechs that are actually regularly used by a pilot. I would have thought my top 10 Grid Iron play has more weight than the less than 10 Shadow Cat crapshoots that I put in or the absolutely abysmal winrate I have with the QKD-4G (my god that mech is simply unuseable but I need to basic it for my IV-Four).

Anyway, I think there is still room for improvement with the system. Back to only tryhard mechs for me until my PSR equalizes out to where it should have been if the algorithm was accurate. Time to dust off my SCRs and TBRs

[Wintersdark]

Depeñds where you spend the most time/matches. God knows I play random bull**** mechs and builds most of the time, and only drag out my tryhard builds in group queue matches, which is a small portion of my play.

[Bishop Steiner]

Krivvan, on 22 September 2015 - 04:20 PM, said:

A lot of people posting in the Reddit thread are Tier 1.
It's a different demographics from these forums though.

Hmmm...Reddit posting correlates to higher play...... I shall never become American Ninja Warrior then! Can't pay me to put up with the toxicity usually associated with MWO on Reddit.

[monk]

Maybe.

PGI should (maybe they have) look at the metrics for all players and see how much their current PSR rating would change and for how many people if we broke it down into the 4 classes. If it's a big difference (as many of us assume) then it probably would make sense to break it down like that for match making.

You could still have an overall PSR, though.

[Ghogiel]

It just dawned on me... but I very much doubt anyone has played 20k games since Jan.

[Dark DeLaurel]

Yes the PSR should be broken down by chassis. XVM does this for World of Tanks and gives you a good metric on if you're using that tank well or not. Even if we had weifht class broken into tiers (also separating Clan/Inner Sphere) this would help give people a better metric on where they need to bring up their game play (or be happyhappy with where they are and enjoy it).

[Karl Streiger]

They should not use more variables if there are not enough players in the pool so that after some time the MM starts to reduce the filter conditions.
back to the closed beta MM:
everybody but the Tier 5 (should be start) in the same pool
A assault in team blue = assault in team red and so on.

They could even throw 2-4men groups in PUG (not in T5) - maybe with a reduced income penalty
other group play is 8vs8 and 12vs12 - with increased income

[0bsidion]

I don't think that goes far enough, I think it should be broken up by chassis/variant. Probably too much to hope for though.

[Fate 6]

Looking at the PSR of my equal skilled comp teammates it seems like playing for funsies is pretty heavily punished by the system as I ended up basically a full Tier off. I think it's gonna take a bit more than weight class specific PSR. Hopefully this will change, but right now my PSR for a grasshopper is combined with my PSR for a Hellbringer. My HBR has 3+ KD and 600 average damage which is vastly different from Hopper stats. The rebalance will hopefully remove the need for separate PSRs but right now it just forces me to stop playing my fun mechs like my Nova and play the Clan Quadrinity (ACH/SCR/TBR/DWF) all day until my PSR gets up to that of my comp teammates

[L3mming2]

Bishop Steiner, on 22 September 2015 - 01:00 PM, said:

There are other factors being discussed in other threads about things people feel skew PSR stats, but here is a pretty obvious one, to me:

Some of us are specialists. Even though a guy like Audivo is still stinking deadly in anything, he is WAY deadlier in Lights, (or so it seems). Likewise, while I feel Tier 3 is probably about right fo rme, my Threat Factor as an opponent in a Hunchback or Centurion is markedly higher than if I am in an Assault or Light Mech.

This I think can and will cause some disparities. And while they may "average" out over time, if for instance, I'm really a Tier 2 Medium jock (not saying that I am, but I do think I am pretty dang good in them) it's a little unfair to me, and my prey, to be counted as Tier 3 because of my other mechs, and thus, I drop in my HBK, and get to hunt Noobs and Scrubs.

Thoughts?

in a perfect world all chassi's would be (more or less) equal, so the difference in your effectivenes should be up to how well it fits your play style. in this perfect world i see no problem with 1psr for all chassi's

[0bsidion]

Fate 6, on 23 September 2015 - 08:04 AM, said:

Looking at the PSR of my equal skilled comp teammates it seems like playing for funsies is pretty heavily punished by the system as I ended up basically a full Tier off. I think it's gonna take a bit more than weight class specific PSR. Hopefully this will change, but right now my PSR for a grasshopper is combined with my PSR for a Hellbringer. My HBR has 3+ KD and 600 average damage which is vastly different from Hopper stats. The rebalance will hopefully remove the need for separate PSRs but right now it just forces me to stop playing my fun mechs like my Nova and play the Clan Quadrinity (ACH/SCR/TBR/DWF) all day until my PSR gets up to that of my comp teammates

I pretty much do the same thing. But if you think you're being punished you're looking at it wrong. The ELO system was messed up. I'm fairly sure that some people landed where they don't necessarily belong because of that, but things will balance out in the end. In the meantime, you can relax and play for fun. If you're better at the game than PSR placed you, eventually you'll rank up. If not, well, at least you're mostly matched against players that are at least in the same ballpark as you.

[Bishop Steiner]

L3mming2, on 23 September 2015 - 08:11 AM, said:

in a perfect world all chassi's would be (more or less) equal, so the difference in your effectivenes should be up to how well it fits your play style. in this perfect world i see no problem with 1psr for all chassi's

more or less equal in no way equals more or less equal fits for players. Locust is a lot different from an Atlas, no matter how "balanced" they are against each other. Some players are more comfortable fast, some slow and armored, some in between.

Global PSR ignores this, and punishes you for playing outside your comfort zone.

[Fate 6]

0bsidion, on 23 September 2015 - 08:21 AM, said:

I pretty much do the same thing. But if you think you're being punished you're looking at it wrong. The ELO system was messed up. I'm fairly sure that some people landed where they don't necessarily belong because of that, but things will balance out in the end. In the meantime, you can relax and play for fun. If you're better at the game than PSR placed you, eventually you'll rank up. If not, well, at least you're mostly matched against players that are at least in the same ballpark as you.

Well the plan is to go full tryhard for a while until I have a decent buffer in T1

[Raggedyman]

Bishop Steiner, on 22 September 2015 - 01:00 PM, said:

Thoughts?

Breakdown by Weight or Chassis works for me, as it stops people being worried that trying out a new mech or class could crash down their Tier ranking.

[Widowmaker1981]

Broken up be weight class, yes. absolutely.

Seperate solo and group queue, yes. Either that or completely remove the win/loss part of the equation so that it is possible to lose PSR in a win. Otherwise members of strong units all end up T1 maxed, pretty much regardless of individual skill.

[nehebkau]

NO.

Why? PSR is pilot skill and should reflect the pilot, regardless of the mech. Having said that i posted this: PSR Calc where i submit that PSR calculations should take into account the mech you are in, the damage it could possibly do as well as its weight and other things.

Paul has his big spreadsheet that he is using for balancing -- it should also be used to figure out a multiplier (or divisor) for PSR match scores depending on the mech you are in.

I know many of the players who are showing themselves as Tier 1 on reddit -- most of them are always playing the same things, meta Timbers, hellbringers or SCRs. Is their PSR a function of their skill or that they are consistently playing in the best mechs in the game? (Logic says both and since PSR is supposed to be about the pilot we should find a way to REMOVE the mech from that equation -- not ADD more mech to it!)
--- EDIT ---

I just saw a problem with my idea. If a player always plays a meta-mech but is ranked lower because of it yet they perform like a higher tier in that mech -- then shouldn't they be in that tier? Perhaps my mistake was separating out mech and pilot. If I was wrong and you shouldn't separate mach and pilot then, realistically, there should be a PSR for EACH mech chassis in the game no?

To elaborate: If pilot a is really a tier 3 pilot adjusting for the fact that he is always in a Meta TBR but performs like a Tier 1 pilot in that TBR meta then he is, while piloting that specific mech a tier 1 pilot. Likewise, if that Tier 3 pilot drops into a locust 5V where he performs like a tier 5 pilot, then isn't he/she really a Tier 5 pilot. Along the same lines, if you drop that same Tier 3 pilot into an Cheetah where he performs like a tier 3 pilot, then isn't he/she a Tier 3 pilot at that instance? In the same breath that Tier 3 pilot may function like a Tier 4 pilot while in a direwhale but like a tier 2 pilot in a warhawk.

Thoughts?

Edited by nehebkau, 23 September 2015 - 08:57 AM.

[Torric]

nehebkau, on 23 September 2015 - 08:40 AM, said:

NO.

Why? PSR is pilot skill and should reflect the pilot, regardless of the mech. Having said that i posted this: PSR Calc where i submit that PSR calculations should take into account the mech you are in, the damage it could possibly do as well as its weight and other things.

Paul has his big spreadsheet that he is using for balancing -- it should also be used to figure out a multiplier (or divisor) for PSR match scores depending on the mech you are in.

I know many of the players who are showing themselves as Tier 1 on reddit -- most of them are always playing the same things, meta Timbers, hellbringers or SCRs. Is their PSR a function of their skill or that they are consistently playing in the best mechs in the game? (Logic says both and since PSR is supposed to be about the pilot we should find a way to REMOVE the mech from that equation -- not ADD more mech to it!)

Actually, making the PSR per mech variant is not really "adding" more mech... it is removing the variance that is induced by playing/having mechs that are so drastically more capable than others.

I like the idea of the per-mech-multiplier in theory, but then you also have to factor in the loadout, and possibly the modules as well.

By the way, judging from the current state of MWO and the PTS desaster, i would not put too much trust in "Paul's spreadsheet".

[nehebkau]

Torric, on 23 September 2015 - 08:59 AM, said:

Actually, making the PSR per mech variant is not really "adding" more mech... it is removing the variance that is induced by playing/having mechs that are so drastically more capable than others.

I am guessing this came in after you:

nehebkau, on 23 September 2015 - 08:40 AM, said:

--- EDIT ---

I just saw a problem with my idea. If a player always plays a meta-mech but is ranked lower because of it yet they perform like a higher tier in that mech -- then shouldn't they be in that tier? Perhaps my mistake was separating out mech and pilot. If I was wrong and you shouldn't separate mach and pilot then, realistically, there should be a PSR for EACH mech chassis in the game no?

To elaborate: If pilot a is really a tier 3 pilot adjusting for the fact that he is always in a Meta TBR but performs like a Tier 1 pilot in that TBR meta then he is, while piloting that specific mech a tier 1 pilot. Likewise, if that Tier 3 pilot drops into a locust 5V where he performs like a tier 5 pilot, then isn't he/she really a Tier 5 pilot. Along the same lines, if you drop that same Tier 3 pilot into an Cheetah where he performs like a tier 3 pilot, then isn't he/she a Tier 3 pilot at that instance? In the same breath that Tier 3 pilot may function like a Tier 4 pilot while in a direwhale but like a tier 2 pilot in a warhawk.

Thoughts?

[Joseph Mallan]

Ghogiel, on 22 September 2015 - 01:03 PM, said:

New players aside I really don't think there is much between T2-4 since the main factor is just the amount of games played. Just play loads and the bar will fill up to max T1 eventually.

Though this is true, there is little difference needed to be better.