13 replies to this topic

[AnswerIsOne]

Why is it when your in a light mech your skill rating is nerfed?
In one match piloting a Light, I killed an assault, a medium and a light.
But my skill rating still went down? WTF?!!?!?!
That is just stupid. Yes we lost the match, but my rating should have increased non-the-less.
I'm sick of dong well and getting multiple kills, doing more damage than I even have in armor and my freaking skill rating still dropping.
Something needs to be done to address this rating method and address this.
Your personal skill rating should reflect just that. YOUR personal skills, not the skills of the
PUGS you've dropped with.

[AnswerIsOne]

Sorry, I posted this in the wrong area. I do not know how to move this to a more appropriate area.

[Tarl Cabot]

Wrong subforum, you can ask a moderator to move it for you. Anyhow, the current PSR is too heavily dependent on WINNING followed by damage, not if you had 3 kills where the total damage you did accounted for a total 50pts (example only) on already severely damaged mechs.

And to add the PSR setup also has more of an upward movement, with the more you play the faster you can move up into tiers.

Edited by Tarl Cabot, 05 March 2017 - 03:21 PM.

[AnswerIsOne]

@Tarl Cabot
just for your info I solo'd the assault and medium, both were undamaged in the given example. and yet I still dropped in the ratings.
I did 990Dmg total with 3 kills in a Jenner.
Losing rating under those circumstances is just wrong.
I also ask if a MOD is available, please move this to an appropriate Sub-forum.

Edited by Answer74, 05 March 2017 - 03:38 PM.

[MaxxQ]

Answer74, on 05 March 2017 - 03:33 PM, said:

@Tarl Cabot
just for your info I solo'd the assault and medium, both were undamaged in the given example. and yet I still dropped in the ratings.
I did 990Dmg total with 3 kills in a Jenner.
Losing rating under those circumstances is just wrong.
I also ask if a MOD is available, please move this to an appropriate Sub-forum.

Yeah, that sounds odd. Earlier today in a Raven, I did 568 DMG with three kills (one solo). We lost but my PSR broke even. Someone here posted a week or so ago that 500+ DMG on a loss nets a break-even, and 600+ on a loss net a PSR gain. That fits for what happened to me, but not you.

Do you have a screenshot of the after-match report?

[Tarl Cabot]

Quote

I did 990 Dmg total with 3 kills in a Jenner.

Hai, what MaxxQ said, do you have a screenshot, cause your PSR notation at the end of the game should have been an up arrow. It is also possible it was a display bug but is something that should be submitted to PGI (with screen shots).

One of the unfortunate things is that the PSR is not setup with a zero-sum formula, and it is predicated on absolute values.

Edited by Tarl Cabot, 05 March 2017 - 03:51 PM.

[Roc Ingersol]

Tarl Cabot, on 05 March 2017 - 03:49 PM, said:

One of the unfortunate things is that the PSR is not setup with a zero-sum formula, and it is predicated on absolute values.

Huh?

[Tesunie]

This is why I've suggested that PSR not be so heavily biased on winning and more on individual match score. There should be more ways to break even (got an expected "average" match score, you stay even) and fewer ways to move up or down (did poorly on a win you go down, did well above average on a loss and you go up).

Of course, we also could use match score to get a bit more work on it, to help reward those tasks that often leads to wins but currently are not well rewarded.

PS: Even on a loss, if you did over 900 points of damage, you should have had a match score above 400, which then should have been a positive result to your PSR. Overall though, I'd recommend you just play the game and ignore PSR. It's interesting to have a goal to get it higher, but I wouldn't fret about it too much either.

[Jables McBarty]

Tarl Cabot, on 05 March 2017 - 03:49 PM, said:

One of the unfortunate things is that the PSR is not setup with a zero-sum formula, and it is predicated on absolute values.

Chuck E Finley, on 05 March 2017 - 03:58 PM, said:

Huh?

A zero-sum formula means that for every pilot experiencing an increase in PSR, another pilot in that match will experience a decrease.

The total PSR gained by "good" pilots (however that's determined) would be balanced by the total PSR lost by "bad" pilots (however that's determined) such that the net PSR change is 0. Hence zero-sum PSR formula.

For example, in Dota 2, all five winning players get +25 MMR points, and all five losing players get -25 MMR points. If the game lasts longer, the difference decreases because it theoretically indicates a more closely matched set of players.

In MWO, they could do the same, where all 12 winning players go up a standard amount and all 12 losing players go down a standard amount; or base it on match score, where the 12 highest scorers go up, the 12 lowest go down; or mixed bags where maybe the top 6 go up, the bottom 6 go down, the middle 12 stay where they are; or where say the top six on the winning team go up, the bottom six on the losing team go down, and the rest stay the same.

[Tesunie]

I feel that, if they are going to call it Player Skill Ranking (or Rating... whatever), than it should be based on Player Skill. Right now, it's based on "Team cohesion" for QP, where it's a matter of how well does your random 12 man team work together compared to your opposing random 12 man team...

I feel we should have more neutral results, which would be more what would be considered "average" match score. If you do "average", than you are probably in the correct tier for your skill level. Average would probably be match scores between 150-300 as a random example. If you scored less than 150 match score, your PSR will go down, the lower your score the more it goes down. If you can exceed 300 in a match, than your PSR would move up. (I should mention that the movement speed should be about where it is now, I'm not suggesting a single good match actually place you in a different tier.)

This way, average performance will leave you where you are, and if you wish to go up you need to do better than average more often than you under perform.

As far as winning and losing having an impact on match score, getting that condition could be a booster to your score, say adding 20 points to your match score total if you win. This way, it would be that much easier to "do better" if you won, but not so crippling "you had better been the best performer in that match or better if your team lost" like it currently is.

(Of course, I also believe before something like this could be added in, the actions that apply to match score could use a little revamp, as potentially game winning actions should provide more rewards, such as capturing points in Conquest or Assault should provide some more bonus. Instead, I've had people win matches because they caped the points, but their match score and rewards was abysmal despite doing what the team needed for them to do for the win. Another example, an AMS unit protecting allies... Can save teammates but has no current rewards...)

[Rogue Jedi]

Tesunie, on 07 March 2017 - 11:07 AM, said:

I feel that, if they are going to call it Player Skill Ranking (or Rating... whatever), than it should be based on Player Skill. Right now, it's based on "Team cohesion" for QP, where it's a matter of how well does your random 12 man team work together compared to your opposing random 12 man team...

not anyway near as much as it used to be.

Player Skill Rating was introduced, I think about a year to 18 months ago, before then we had an Elo system, in which the entire thing was based on if your team does better or worse than expected, there were no published figures but you would have a ranking out of 1500, the matchmaker would assemble 2 teams then make an estimate of how each side would do, (e.g. reds will win but only have 2 Mechs left) if your team did exactly what was predicted ratings do not move, do worse and you loose ranking, do better and gain ranking, so your ranking was completely dependant on the teams sucess (or lack thereof).

with PSR at least persional contribution is a significant factor, on a win 250 score or better will give a gain in PSR, on a loss I think you require 450 or better match score for a gain, and 250 or greater for a maintain, those standards are the same regardless of which Mech you are piloting.

Edited by Rogue Jedi, 07 March 2017 - 11:49 AM.

[Tesunie]

Rogue Jedi, on 07 March 2017 - 11:45 AM, said:

with PSR at least persional contribution is a significant factor, on a win 250 score or better will give a gain in PSR, on a loss I think you require 450 or better match score for a gain, and 250 or greater for a maintain, those standards are the same regardless of which Mech you are piloting.

I thought it was 300ish match score on a loss to break even. Actually, I'm sure it's near 300 if not closer to 350 just for an even PSR on a loss...


I've always kinda felt that PSR was just Elo reskinned with a few "exception" clauses placed in. AKA: It's still about winning or losing, but instead of being 100% based on W/L, it is not more like 80-90% based on W/L. (Percentage is randomly generated to how I feel it seems.)

I just feel that, as far as rankings go, if you are getting "average" results in your matches, you are probably at the correct tier for your game play. If that is the case, than winning or losing with "average" match scores shouldn't shift your position around. As you go higher in tiers, it theoretically becomes harder to get "above-average" with the same level of skill and harder to get the same performance. (As in theory you are fighting tougher opponents which should keep you from getting the same match scores.)

I agree that winning and losing should have some minor play in PSR, but your ranking shouldn't be so dependent upon it either. It should be ranking the individual, not the other 23 players in the given match. A single player can only do so much after all.

(Of course, this is just my opinion on the system.)

[Rogue Jedi]

Tesunie, on 07 March 2017 - 11:53 AM, said:

I thought it was 300ish match score on a loss to break even. Actually, I'm sure it's near 300 if not closer to 350 just for an even PSR on a loss...


I've always kinda felt that PSR was just Elo reskinned with a few "exception" clauses placed in. AKA: It's still about winning or losing, but instead of being 100% based on W/L, it is not more like 80-90% based on W/L. (Percentage is randomly generated to how I feel it seems.)

I just feel that, as far as rankings go, if you are getting "average" results in your matches, you are probably at the correct tier for your game play. If that is the case, than winning or losing with "average" match scores shouldn't shift your position around. As you go higher in tiers, it theoretically becomes harder to get "above-average" with the same level of skill and harder to get the same performance. (As in theory you are fighting tougher opponents which should keep you from getting the same match scores.)

I agree that winning and losing should have some minor play in PSR, but your ranking shouldn't be so dependent upon it either. It should be ranking the individual, not the other 23 players in the given match. A single player can only do so much after all.

(Of course, this is just my opinion on the system.)

I have screenshots showing defeat with an = and 262 match score, as well as a \/ with 248, so I am pretty sure 250 is the cutoff for maintain on a loss, and have simular for a win, 240s is an = 260s is ^.

winning plays a big part in PSR but is no longer the whole story as it was with Elo which is my point, a win means an average game is an ^ on a loose the average is an =, however your persional contribution does have an impact, get a low score on a win and you maintain, a low score on a loss and you go down, get a high score and you go up regardless how the rest of the team does.

[Tesunie]

Rogue Jedi, on 07 March 2017 - 01:48 PM, said:

I have screenshots showing defeat with an = and 262 match score, as well as a \/ with 248, so I am pretty sure 250 is the cutoff for maintain on a loss, and have simular for a win, 240s is an = 260s is ^.

winning plays a big part in PSR but is no longer the whole story as it was with Elo which is my point, a win means an average game is an ^ on a loose the average is an =, however your persional contribution does have an impact, get a low score on a win and you maintain, a low score on a loss and you go down, get a high score and you go up regardless how the rest of the team does.

For the cut offs, I'll take your word for it. I'll be honest, as I don't typically give it that much consideration. I tend to look and go "oh, it went up" without really observing what my match score may have been. They also may have adjusted the values needed over time...

I just feel it should be less based on W/L and a bit more on individual performance/match score. I just feel it's still to heavily biased on the win, similar to how Elo worked.