Jump to content

- - - - -

The Player Skill Rating (Psr) System Explained... (As Best I Can)

Guide

149 replies to this topic

#141 Bilbo

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The Nimble
  • The Nimble
  • 7,864 posts
  • LocationSaline, Michigan

Posted 09 March 2016 - 06:13 PM

View Postbar10jim, on 09 March 2016 - 06:02 PM, said:

I've always felt that the PSR needed to be "adjusted" for matchmaking purposes based on the mech that you are piloting. Now, I don't mean having a separate PSR for each variant, chassis, or weight class. Merely and adjustment considering if the mech you are piloting has completed Basic or Elite levels of training.

For example: as a player you have a PSR of 30% of tier 3. If you have not finished Basic in your current mech, you enter the queue a full tier lower (30% tier 4) for matchmaking purposes ONLY (your base PSR stays the same, and is adjusted as normal following a match. If you have completed Basic, but not Elite, you enter the queue 1/2 tier lower (80% tier 4 in the above example).

PGI could vary the "spread" or percentage of tier reduction as result of performance metrics to better tune the matchmaker results and match balance.

So what happens with a tier 1 player that takes a trial mech (which are fairly well built these days) they don't own and now automagically get to play against the tier 4 players they are currently barred from playing.

Personally, I don't like the removal of weight class specific ratings, but downgrading tiers based on mech efficiencies seems a little to easily exploitable. Taking trial mechs out of the equation, why would someone spend XP on their mech until they could just max everything out when they will generally have an easier time of it in a brand new mech.

#142 Wintersdark

    Member

  • PipPipPipPipPipPipPipPipPipPipPip
  • 13,375 posts
  • Google+: Link
  • Twitter: Link
  • LocationCalgary, AB

Posted 09 March 2016 - 06:17 PM

View PostBilbo, on 09 March 2016 - 06:13 PM, said:

So what happens with a tier 1 player that takes a trial mech (which are fairly well built these days) they don't own and now automagically get to play against the tier 4 players they are currently barred from playing.

Personally, I don't like the removal of weight class specific ratings, but downgrading tiers based on mech efficiencies seems a little to easily exploitable. Taking trial mechs out of the equation, why would someone spend XP on their mech until they could just max everything out when they will generally have an easier time of it in a brand new mech.

I liked class specific weightings too, but we don't have enough tiers to allow downgrading.

The reality is that the tiers really exist solely to protect newer players from experienced players. Put me in a brand new Awesome, and I'll still hold my own in T2. Put me with brand new T5 players, and I'll mop the floor with them.

Yes, you're less effective in some mechs, particularly brand new ones without pilot skills etc, but it's not THAT big a gap.

If you're in T2, T1, it doesn't matter what mech you're in, under no circumstances should you be put into a match with new players.

#143 Tesunie

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The Seeker
  • The Seeker
  • 8,586 posts
  • LocationSeraphim HQ: Asuncion

Posted 09 March 2016 - 07:18 PM

I appologise now to the OP for my upcoming long post. However, consider it a good topic bump for you.




View PostRoadkill, on 09 March 2016 - 02:28 PM, said:

Elo ratings are derived from the results of playing the game, using win or loss as the only input. That's not the same as saying it is "based on W/L."


Ever heard of the term "splitting hairs"? Because that is what you just did. Let me explain:
Derived from = Based on
Same thing. Different words.

It rates you based on your wins and losses.
Which means it derives a number from your wins and losses.

Shall I pull out a dictionary?
Derive: base a concept on a logical extension or modification of (another concept).
Base: a conceptual structure or entity on which something draws or depends.

Thus: Elo is derived on the concept of win and loss ranking = Elo is based on the concept of win and loss ranking.

View PostRoadkill, on 09 March 2016 - 02:28 PM, said:

Sort of. It isn't based on the player's W/L rate, it's based on their results against other players. That sort of sounds like W/L, but it actually isn't.

For example, you can't look at a player's stats and say "this player has a W/L rate of .500, so his Elo rating is 1600." His Elo rating will depend on the ratings of the players he beat and lost to, not simply on his W/L rate.


Read my statement farther on...
Yes. It is based on W/L, and adjusts a value based upon (I mean... Derived upon) the value of the opponents you fought.

View PostRoadkill, on 09 March 2016 - 02:28 PM, said:

False. Elo is a rating system only, it doesn't have anything to do with matchmaking.

A matchmaker can use Elo ratings as input into its matchmaking process, but that's not at all the same thing.


Where we NOT talking about Elo as a Match Making tool?

We where talking about Elo as being used in MW:O to produce matches. So, TRUE. Not False.

View PostRoadkill, on 09 March 2016 - 02:28 PM, said:

An expected outcome can be derived from relative Elo ratings, yes. But that's not really "predicting" the outcome of the match like people think. It simply gives you a percent chance that side A will beat side B.

PGI's matchmaker attempted to keep the aggregate Elo ratings of the two teams equal. If the two sides' Elo ratings are equal, then the expected outcome is 50/50. I.e. both sides have an equal chance to win.


Definition:
Predicting: say or estimate that (a specified thing) will happen in the future or will be a consequence of something.
Predicting = An expected outcome.

Second part, true. I believe I said that. But, if one side was not equal in Elo, than the system attempted to calculate if the lower side had a reasonable chance to win. If they did not, the system either did not drop their ratings, or if it did it only dropped it a small margin. If the lower Elo team won however when it was expected to lose, they would see a larger increase in Elo. As used as part of a match maker system.

View PostRoadkill, on 09 March 2016 - 02:28 PM, said:

This is where PGI deviated from a proper Elo system, and their deviation made the system less responsive. Their change still works, it just works even more slowly than normal.

In a proper Elo system, your rating will change after every match (with some minor corner-case exceptions) in line with the expected outcome.

What that means is that if you and I play, and I have a 90% chance to win, then if I win there will be very little change in either of our ratings. But if I lose there will be a very large change in both of our ratings. If I only have a 50% chance to win, then win or lose both of our ratings will change by a moderate amount. The winner's rating always increases while the loser's rating always decreases. (Again, except for some corner cases that aren't worth worrying about.)


I believe this is what I had said... was it not?

"- If the player who is predicted to win wins, then there is little change in their rating. If the player predicted to win losses though, then changes happen accordingly."

Oh yeah. It was. I just didn't go into as much detail in a quick bullet point.

View PostRoadkill, on 09 March 2016 - 02:28 PM, said:

It was first used for chess, but it isn't correct to say that it was "designed for Solo PvP" in any way. Elo ratings are simply a mathematical formula that works in any 2 sided system. Each side can have as many players on it as you want.


It was created and designed for Chess. Chess is a 1v1 game.

Elo, as far as an individual rating system, works BEST in 1v1.

The problem with Elo in a team based game, especially one with randomly generated teams (unlike if the teams were set and remained unchanged), is one match I may be the highest Elo player there (expected to carry the team) and the next I may be the lowest Elo player in the match (and expected to die a very quick death). Elo doesn't account for this, so if I managed to carry a team by myself (not saying I can) and win, everyone is better off. If I get carried in the reverse situation, and my team wins even though I (in theory) died to better Elo/skilled players, my Elo is still better served.

So, as far as randomized teams, Elo really isn't a great metric to work off of. For set teams, it would probably work rather well.

This is why PSR is suppose to be a better system, accounting for individual player skill in the, otherwise normally, randomized team play environment. If PSR worked as it should (as a match making tool), in theory if I pair off against someone of the same tier/ranking/exact PSR, we should have a 50% chance of beating each other in a test of skill. With Elo as it was working before hand, I could have a higher Elo than another player, yet that other player might have more actual skill and beat me easily in a 1v1. Why? Because my Elo was derived from my randomized (PUG) team match scores, where I alone am not in control of if the team will win or not. (Reminder, there are 23 other players in the game with me. A single person rarely can determine a win all on their own.)

View PostRoadkill, on 09 March 2016 - 02:28 PM, said:

False. Elo works just as well for multi-player as it does for 2-player, it just takes longer to reach equilibrium. PGI's Elo ratings took even longer because of the change they made. However, players with thousands of games easily had accurate Elo ratings even with PGI's modified Elo ratings. (The math says that a couple hundred games at most would be plenty, and possibly as few as 50.)


No. True. In randomized teams, often with teams not being balanced in Elo scores, it was not working in MW:O, which I might remind is a game with multiple players on each team, and each team (for the most often played game mode) is completely random in it's build (rarely having the same players on the same teams, or even in the same match very often). 1 out of 24 (an influence rate of 4.1667% or 1/24=0.041667) isn't exactly game influencing enough to make strictly a system based on W/L very effective.

For group queue, it probably would work well, especially for 12 man teams. The teams are set, and thus changed the win influence rate for the set elements up to 50%. (Teamwork wins games, not a single solo player often times.) Then, calculating for Elo and if there is a difference...

Elo is not accurately effective (AKA: Doesn't work so well) in a multi-player environment with randomly assigned teams per match. Especially when it often times wasn't creating balanced teams within the match maker. (Recall having to wait a long time? The longer you waited, the more unbalanced your Elo was in the match between teams.)

View PostRoadkill, on 09 March 2016 - 02:28 PM, said:

You fell into the typical gamer trap, that's all.

Elo ratings are a very simple, very accurate mathematical formula. All of the rest of that baggage really belongs to the PGI matchmaker and it's use (or misuse) of Elo ratings as input.


You fell into the "splitting hairs" trap.

Elo, as being discussed as it was implemented in MW:O (Match Maker included) was very inaccurate on individual player rankings. I might also want to mention, if Elo was so completely accurate, why is it many games seem to start out using Elo, and then change over to another (though often related or based on Elo) system? If it was so accurate, wouldn't they remain with Elo, and not change anything? Extra information found here.
Spoiler


Any farther questions on my knowledge on the topic? Anything else I missed? Any more hairs/words you wish to split, cut, ream, shread, or mince?

#144 L Y N X

    Member

  • PipPipPipPipPipPipPip
  • The Nimble
  • The Nimble
  • 629 posts
  • LocationStrana Mechty

Posted 16 April 2016 - 09:32 PM

For new players, the first few pages of this thread explains how the PSR rating system works. The rest is for entertainment only. Some players want to be T1, while others could care less. PSR matchmaker helps to separate the two groups, not perfect, but far better than what used to be.

#145 El Nino

    Rookie

  • Bad Company
  • 1 posts

Posted 24 April 2016 - 02:25 PM

I really don't know about this system. I started being pretty bad and consider myself to be decent these days. I played a a few hours today and am nor barely broke into T3. My Win/Loss ratio is 1.08 and my KDR is 1.40. I think I am better than I am rated here. Moving up seems slow. I would just like to understand this better.

#146 Tesunie

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The Seeker
  • The Seeker
  • 8,586 posts
  • LocationSeraphim HQ: Asuncion

Posted 24 April 2016 - 03:19 PM

View PostEl Nino, on 24 April 2016 - 02:25 PM, said:

I really don't know about this system. I started being pretty bad and consider myself to be decent these days. I played a a few hours today and am nor barely broke into T3. My Win/Loss ratio is 1.08 and my KDR is 1.40. I think I am better than I am rated here. Moving up seems slow. I would just like to understand this better.



PSR is not designed to make massive movements in it's ranking of your skill. It is suppose to be gradual changes as they are earned. That is, unless you are massively performing different than your PSR currently indicates you should be.

For example, if you constantly got match scores in the thousands (I don't even think it's possible, but for example sake), you would find you would be quickly moving through the ranks till your performance slowed down to a more reasonable level, or you hit T1 (highest rank).

If, on the flip side, you constantly got a match score of 0-20 every match, you would probably find your match score dropping more visibly.

If you are seeing match scores between 200-300 (what is more or less average), you will find your PSR will increase slowly or even remain the same.


PSR is designed for "over the long haul" game play. It doesn't shift quickly, but is suppose to nudge you into your PSR rank, as you improve. It's suppose to reward longer term play more than short term play. Sure, you may be doing great now (for whatever reason), but tomorrow you may have a harder time. You wouldn't want to jump from PSR T5 to PSR T1 after a couple battles and then right back down to T5 a few matches later.

Keep playing. It will play you into the proper tier over time. T3 is considered "average" by most players. I myself am mid T3, and I've had people comment (in match) that I appear to be an above average player (with a lot of bad luck Posted Image ). Personally, I'd recommend that you just ignore your PSR tier level. It really doesn't mean all that much and is only used to better place you in matches with more evenly skilled opponents. Posted Image

#147 King LaLa

    Rookie

  • Philanthropist
  • 8 posts

Posted 13 June 2021 - 04:47 AM

Why my PSR went down when I have 290 match score in a winning game. 1 Kill and 1 KMDD, not once but 3 times?

I am in a light mech

#148 martian

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 8,429 posts

Posted 13 June 2021 - 05:11 AM

Greetings!

View PostKing LaLa, on 13 June 2021 - 04:47 AM, said:

Why my PSR went down when I have 290 match score in a winning game. 1 Kill and 1 KMDD, not once but 3 times?

The first things first: Ignore everything what is in this thread. The PSR system has been changed since July 2020, so posts from 2015, 2016 or 2017 are no longer relevant.

And now to the point ... Posted Image

Read this topic, please: Psr Community Version 1.0


Short summary:

Your Match Score is compared with the Match Score of the remaining 23 players (the members of the victorious team get a small bonus). In theory, the best 12 players will get the green arrow (PSR rise) and the rest gets the red arrow (PSR slump). Sometimes you get the yellow "=" sign (your PSR is unchanged).

Note that in one game your MS 290 can put you among the best players and thus earn you the green arrow, while in some other game with the same MS 290 you might end up with the red arrow (if there are 12+ players with the higher MS).

The number of kills and KMDDs is just one factor that adds to your Match Score and thus to your PSR. There are other activies, that you can do, to compensate for the lack of kills.

View PostKing LaLa, on 13 June 2021 - 04:47 AM, said:

I am in a light mech

The weight class of your 'Mech has no direct influence on your PSR. You can get the green arrow in a light 'Mech and the red arrow in a badly piloted Assault 'Mech.

#149 Void Angel

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • The Marauder
  • The Marauder
  • 7,062 posts
  • LocationParanoiaville

Posted 13 June 2021 - 07:40 AM

Necro, necro, necro,
How do you find these posts?!
Necro, necro, necro,
This thread should be a ghost!

You had a game experience,
and wonder what it means -
And somehow you found this thread,
from Two Thousand Sixteen!

Necro, necro, necro,
You're old enough to know.
Necro, necro, necro,
this thread is way too old!

#150 Ekson Valdez

    Volunteer Moderator

  • PipPipPipPipPipPipPip
  • 716 posts

Posted 15 June 2021 - 04:43 AM



Please refer to the above linked official post for PSR change. Thread closed.







14 user(s) are reading this topic

0 members, 14 guests, 0 anonymous users