53 replies to this topic

[SolarCleric]

Fenrisulvyn, on 26 September 2015 - 02:19 PM, said:

Problem I am seeing is that since match score is weighted heavily towards damage, non-brawly lights are screwed. The reward system under-represents spotting, flanking, luring, curling, capping, etc. which exacerbates the negative effects of PSR.

Today I bought Timberwolves. Its easy to rack up damage and kills, and significantly mitigates taking a loss. In fact I had several matches where we lost but I actually went UP in PSR soley because of the extra damage I could throw out.

My Raven 3L is running 3 med pulse.
My Timber is a 4 ERML / 24 Cstreak.

I work my butt off in my Raven and get negative arrows
I am a clueless noob in my Timber and get positive arrows.

That's very frustrating. If I want to raise my PSR, I have to ditch my fav mech.

There is the reason why the HEAVY and ASSAULT ques are the longest.....

Edited by SolarCleric, 26 September 2015 - 03:20 PM.

[Aggravated Assault Mech]

k05h3lk1n, on 26 September 2015 - 05:04 AM, said:

The best indicator would be to collect your average performance(match score) and comparing that with the rest of the player base. Then you can divide the playerbase into tiers according to this average performance.(Top 10% = Tier 1, following 20% = Tier 2, following 40% = Tier 3, following 20% = Tier 4, Tier 5 = worst 10%)

Winning is the ultimate indicator of skill because that's the objective of the game.

Making PSR score based benefits those who play to farm score, rather than those that play to win. That's a system that makes high tier players fight against their own team rather than with them- using allies as meat shields, damage farming off torsos and legs, letting allies die in favor of farming etc.

Win rate is the best metric because it rewards people for playing at the most organized and competitive level, and does not reward selfish or toxic tactics that are products of any score-based system.

The only pitfall of making it win rate based is if group and solo win rates aren't separated by the system.

[Light-Speed]

[VinJade]

@Night
That is just stupid, if you did that well you shouldn't have lost any ranking.

[Samedi Wretch]

I just want to put this out there.

A good player WINS.

The job of the PSR is NOT to tell you you are awesome (read: needs better hobbies) but to construct competitive matches. That's what it was designed for and its an essential element of this game. The only reliable metric, let me repeat that, THE ONLY RELIABLE METRIC for predicting future winners is measuring past winners! That's it. Deal with it.

[Bluttrunken]

samadhiVOID, on 27 September 2015 - 07:59 PM, said:

A good player WINS.

While I wouldn't say I'm a good player myself I have to say no: A good team wins not the good player. In the solo queue your team is a completely random collection of 12 individuals. Making your own progress dependable on the effort of 11 other people diminishes any use of the PSR as an indicator of *personal* skill. Sometimes you have matches which are won by a hairs breadth and the good effort of both teams will only awarded to one of them.

In the end it doesn't matter, though, because what we have aren't Tiers but more a visualization of how long we effectively played the game(more or less). Real Tiers would always keep their size at a certain percentage of the population and as soon as your performance drops you drop out of the Tier as well. Tier 1 players should have to struggle to keep their Tier 1, which really isn't the case with the current system. The complete population of MWO could, theoretically, be Tier 1.

[Randomerlight]

The problem with much of the above discussion is that it focuses on extremes--the either really terrible or really amazing "god" players. The middling players are left to the randomness applied by the PSR's team based weight.

I question a ranking system where I have to play like a god to regularly ascend, and any other match performance (above average, average, middling) is treated relatively randomly--my personal score is based entirely on my team unless I am so amazing I can counter the random abilities of 11 other players. Inouye's own post references that the new system is built around team play, and actually encourages it--hence the focus on team W/L in the OP's posted graph.

The problem with this is PUGs are rarely team based, and if they exhibit any kind of team work, it's the most basic "rally point" or "enemy sighted" chat directions. So from what I can tell, the biggest problem with the new tiers is that they are heavily biased against PUGs. Anyone of middling skill (I average about 200-250 dmg per round, only play lights and meds) who plays PUGs is then forced to deal with the reality that their PSR is completely based on the randomness of the teams they are assigned.

Either PSR should be based on *your* match points (as the OP suggests) across the game, or there should be some changes made to differentiate between PUGs (where players are going to be more dependent on their individual skill level) and true team based play (where the current PSR system makes more sense).

[Light-Speed]

BTW, I want to mention something I thought of because of a recent conversation.

It doesn't make sense to reward the most long-term stably-supporting players with Queue Time Hell.

So I support the system where T1 players have to struggle to keep their position, because there are bound to be T1 players who troll their teams by actively trying to suck (to lower their tier, by, say, dying in the least amount of time possible) while taking away from the team a T1 player.


Depressing conclusion not for the pessimists:

Spoiler

Doesn't make it depressing if the fact that you might fall into Queue Time Hell the better you do and the longer you stick around hangs over you while you strive to learn, improve, and win?

Edited by Nightingale27, 10 October 2015 - 04:28 PM.

[Tsar Bomba]

Nightingale27, on 27 September 2015 - 06:28 PM, said:

ROFL, I love how you protect the names of your fellow losers. That's a bro move if I've ever seen one. Thanks for the laugh, I needed it!

Edited by Tsar Bomba, 10 October 2015 - 05:21 PM.

[TexAce]

vnlk65n, on 26 September 2015 - 02:52 AM, said:

Strictly speaking win rate is the best indicator of skill over infinite number of games, if you were to seperate solo from group matches. A good player will win more matches than the average, and a bad player the opposite.

wrong. Because the MM is set up in a way to always keep you on a W/L ratio of 1. As long as you only pug that is.

If you only pug your W/L will reach 1 after a while and never move again. Its just like that.

Thats why PSR cant work with such a big influence on winrate

[Randomerlight]

KisumiKitsune, on 10 October 2015 - 05:02 PM, said:

Change or remove the graph, please. You're spreading misinformation. The bar moves the same amount whether you get a 99999 match score win or a 100 match score win.

Um, that's Inouye's own graph. http://mwomercs.com/...-tiers-and-psr/

[TexAce]

KisumiKitsune, on 10 October 2015 - 05:02 PM, said:

Change or remove the graph, please. You're spreading misinformation. The bar moves the same amount whether you get a 99999 match score win or a 100 match score win.

There IS however, a direct corrolation to people of higher skill and win / loss ratio. All pilots are not equal, and some pilots do consistently win games for their team. Some are carried. Higher skilled players can, and do, win entire games for their team that would have been lost otherwise. As it is, the only way you can go down is by doing bad on a loss. You have to be pretty dang bad to actually go down tiers, because you have to contribute nothing or very little on wins, and do even worse on losses to not stay where you are.

This does create a problem where the PSR system is more of a leveling bar and not an objective placement, because most people should naturally go up. I'll be proof of this myself by ending up in T1 sooner or later, as evidenced here with when PSR launched vs where I am right now.



Matches aren't getting any harder, and there are easily over 100 games there. Unless I'm on an unusual streak, this trend will continue. Time played = bar filled.

Dude that graph is literally the one PGI published.

[El Bandito]

Nightingale27, on 27 September 2015 - 06:28 PM, said:

You need 251 points to not lose your rating. You were off by 1. Too bad, hahahahaha!
For future references, here is the scoring system.

El Bandito, on 09 October 2015 - 08:36 AM, said:

On Loss: It is <101 = large drop, 101-250 = medium drop, 251-400 = even, 400< = small rise.
On Win: It is <101 = even, 101-250 = small rise, 251-400 = medium rise, 400< = large rise.

Basically 150 points difference per level.

Edited by El Bandito, 10 October 2015 - 07:18 PM.

[Narcissistic Martyr]

I find the key to PSR happiness is to stop caring about your tier.