14 replies to this topic

[Katotonic]

Hi All,

I have seen a ton of threads talking about PSR, They range from "its Bull*#$^" to "how to improve your score" to "Help me understand how it works", etc.

One of the biggest things I have seen is that some people have seen drastic movement in just a few games and some have seen no movement. I have two thoughts to explain this:

1) If you do not have enough games played to give the system an adequate idea of your skill level (and we are probably talking 1000+), 1 game is going to create a big movement, good or bad as your average score will change more drastically.

and

2) (This is the big one that I have not seen mentioned anywhere and (I believe) is true): All PSR tiers/scores are relative to the rest of the MWO playerbase. They have effectively taken all players (which by current guesstimate is about 6000 to 7000) and divided us into 5 lvls. Top 20% is tier 1, next 20% tier 2, etc.

This means that you tier ranking could move without you even doing anything!!!! Say you are lower Tier 3 (like me). If I don't play for 2 weeks and a bunch of people in upper tier 4 get better (like a bunch of new players) while not too many new people join and the rest of the player population stays roughly the same in skills. I might be bumped to tier 4 as now I would not make the cut off number to be in the middle 20%.

Now, don't freak out as this would probably not happen too often as people general playing skill level won't change too much that quickly. But do consider that your first 20 games compared to your game 100-120 you probably improved a lot so it could happen. Especially in the T4 to T3 area as new players are auto placed into T4.

Just something to keep in mind.

[Gnume]

Katotonic, on 23 September 2015 - 05:54 AM, said:

2) (This is the big one that I have not seen mentioned anywhere and (I believe) is true): All PSR tiers/scores are relative to the rest of the MWO playerbase. They have effectively taken all players (which by current guesstimate is about 6000 to 7000) and divided us into 5 lvls. Top 20% is tier 1, next 20% tier 2, etc.

This means that you tier ranking could move without you even doing anything!!!! Say you are lower Tier 3 (like me). If I don't play for 2 weeks and a bunch of people in upper tier 4 get better (like a bunch of new players) while not too many new people join and the rest of the player population stays roughly the same in skills. I might be bumped to tier 4 as now I would not make the cut off number to be in the middle 20%.

We haven't heard them say that Tier's are relative to population and that they are broken up by percentages. It could just be a Hardcoded Number (a la experience points) that once you reach that number, you magically move up a tier. Lose some Exp, and you could drop down a Tier...but there are no limits to the number of players that can be WITHIN a given Tier.

[Raggedyman]

CaptRosha, on 23 September 2015 - 07:21 AM, said:

We haven't heard them say that Tier's are relative to population and that they are broken up by percentages. It could just be a Hardcoded Number (a la experience points) that once you reach that number, you magically move up a tier. Lose some Exp, and you could drop down a Tier...but there are no limits to the number of players that can be WITHIN a given Tier.

Thing is it's supposed to be a banding to give you a decent game experience, so if it isn't relative then it's going to get terrible pretty fast as people get better over time which will result in Tier 1 becoming an insane slaughter zone of death/you having a skill cliff at the Tier boundaries. You would also see some serious issues at the edges of each banding, where people try and stay inside bands to max out returns on time.

Not saying if it's hardcode or banding, as I don't know. Just suggesting why I hope to gosh it's not hardcoded.

[Rhaythe]

Not to mention - if you get good, enjoyable and competitive games, who cares what tier you're on.

[Lily from animove]

there is no drastic improvement. the bar only goes up very slowly, so you gonna need a laod of games, thats all.

[Lorian Sunrider]

I was mildly surprised I was Tier 4 with only a small smidgen to go to get Tier 3. Then I read the information about the amount of games being played playing a large part and realized even though my skill level may not be representative of my PSR, the amount of time I put into the game compared to others may be.

But I'm content playing Tier 4 right now. I carry hard plenty of matches, which is fun. And when I do drop with my friends who are Tier 1 and 2, I'm right there with them in match score and damage.

Its easy to get caught up with a system telling you that you are worse than you think you are. But its an imperfect system and as long as you are having fun who cares.

And on a side note I think this allows for some fun with streamers/content creators. Some of my favourite streams to watch on Twitch are League players starting Alt accounts and grinding up as high as they can. Now as long as PGI doesn't get any idea's of Tier 1's being able to "prestige".

[LordSkippy]

Katotonic, on 23 September 2015 - 05:54 AM, said:

All PSR tiers/scores are relative to the rest of the MWO playerbase. They have effectively taken all players (which by current guesstimate is about 6000 to 7000) and divided us into 5 lvls. Top 20% is tier 1, next 20% tier 2, etc.

BTW, people with alt accounts have found that a bad one game match score ranked their alt around 25,500th place in the challenge last weekend. So, the number of accounts has to be at least close to 26,000.

Considering how long the game has had open membership, 26,000 is a very low number for total accounts. Especially when one weekend saw around 25,000 play. So, your 6-7k guesstimate is orders of magnitude off.

One of the problems with making player's PSR rankings relative to the rest of the playerbase in a game with frequently shifting populations is what do you define as the playerbase and what to do with inactive accounts. If a lot of players stuck in tier 3 quit, their scores don't go away. And you develop a blockage for active players to progress out of tier 4.

It safe to say that a large number of players quit before the beginning of the year. Which means their initial placement is in tier 4, near the top. That creates a huge swelling of players clustered around a point, which breaks relative rankings.

Plus, it creates situations where a player's rank can change without any sense of reason to the player. "I played every night for three weeks to break into tier 1, to a weekend break, and I'm tier 2 again when I logged in! Forget this!"

I doubt it's relative based.

[Weeny Machine]

Rhaythe, on 23 September 2015 - 08:27 AM, said:

Not to mention - if you get good, enjoyable and competitive games, who cares what tier you're on.

I totally second this. I am tier 3 and get away with playing non-meta mechs which I enjoy. Why would I want to move upwards to the meta wh...meta oriented people? I have fun here.

[Death Proof]

Pretty sure as your average match score increases you will progress through the tiers. If it starts to decrease you will start sliding toward the next lesser skilled tier...until you settle into the tier you belong.


The thing that is not clear is what match score PGI deems to be the perfect mean that neither raises or lowers your PSR. 100? 200? 300?

Basically, if you're regularly putting up scores of 400+ you can safely assume that you will progress through your current tier pretty quickly until you reach the next tier, and your scores start getting lower on average. If you are still putting up scores of 400+ at T1, then you definitely belong there. Just the same, if you have a hard time breaking 100+ scores, then you can likely expect your progress to start slow down and possibly drop down.

[Ghogiel]

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

Considering how long the game has had open membership, 26,000 is a very low number for total accounts. Especially when one weekend saw around 25,000 play. So, your 6-7k guesstimate is orders of magnitude off.

The total number of regisitered accounts is over a million since forever. http://mwomercs.com/...gistered-users/

Yep, that 26000 is just who logged in and played at least one match over the weekend.

[sycocys]

I think a lot of what people miss is that a bunch of your score is dependent on the skill level of the other 11 MM throws on your team.

Been seeing it a lot while leveling these Stormcrows (all with identical loadouts, from 0 skills through basics) - I put up the same numbers in nearly every match: 400-500 damage, 1-2 kills, 9 assists and a bunch of the other rewards bonuses. The same scores can gain me PSR one match where I end up on a mostly competent team, lose points when I end up an an incompetent team, or stay the same if its 50/50. W/L doesn't really seem to matter as much, have stayed the same and gone up in wins, not down yet.

The thing is my numbers aren't changing much at all, but the numbers of the rest of my team fluctuate and really feel more like a determining factor than my own performance.

Just seems weird, I'll give it a bunch more games now that it's possible to have at least some insight on whats going on.

[LordSkippy]

sycocys, on 23 September 2015 - 09:55 AM, said:

400-500 damage, 1-2 kills, 9 assists and a bunch of the other rewards bonuses. The same scores can gain me PSR one match where I end up on a mostly competent team, lose points when I end up an an incompetent team, or stay the same if its 50/50. W/L doesn't really seem to matter as much, have stayed the same and gone up in wins, not down yet.

That would roughly be a match score around 300. That seems to be a magical match score on losses. On wins, it's a straight up boost. On losses, it could be anything.

There's a thread with people posting match scores and if they went up or down. Going up, down, or staying the same appears to rely most on victory or defeat, then by match score in relation to others in the match.

[sycocys]

The thing is what I'm seeing is that this average, at least with this particular mech because its been very consistent, doesn't seem to take my score into account as much as how well the other 11 players perform.

Its far more apparent on losses when you can see it go either way with the same performances, the different factor is your teams performance.

[LordSkippy]

sycocys, on 23 September 2015 - 10:30 AM, said:

The thing is what I'm seeing is that this average, at least with this particular mech because its been very consistent, doesn't seem to take my score into account as much as how well the other 11 players perform.

Its far more apparent on losses when you can see it go either way with the same performances, the different factor is your teams performance.

The primary factor is win or loss.

Win, you either go up or stay the same. From what I've seen and read, you have to do very poorly in a win in order not to go up.

On a loss, you will most likely go down. Depending on your match score in relation to the other players, you have a chance to stay the same or even go up in PSR.

Whether it takes just your team or both teams in consideration on a loss is anyone's guess.

But the primary factor is whether your team won or lost.

[sycocys]

Well usually on the losses I've been seeing I'm in the top 3 on my team and generally in the upper half for both sides.

Usually end up equal, sometimes lose. It's hard to count w/l since in solo drops it is essentially just a roll of the dice and my rng in that regard has been horrible since shortly after the last event started. Went from 15-20 games with a positive win count to 30 games under - I literally could feel the quality of matches just tank the more I played the event.