45 replies to this topic

[Lily from animove]

Thorqemada, on 23 September 2015 - 12:56 AM, said:

Winning seems very urgent - like i wrote in a different Thread - Sean Lang (Tier 1) got a Green Arrow for Matchscore of 106 in a won Group-Q-Match i saw on Twitch.

Imho that is ridiculous low!

PS - Proof at the ~25:30 mark for that matchscore and PSR increase:
http://www.twitch.tv/ngngtv/v/17255295

wow, really? that already kinda sucks, rofl, so doing 450 damage and losing is a down, because you just lost?

That looks quite flawed given that you suddenly say a 200 damage dealing winner is said to be upskilled while a 450 damage loser is having to get "downskilled"

That brings me exactly back to the initial question after PGI posted the patchnotes introducing tiers:

what are PGI's ideas of good, bad and average

Quote

• If a players' team wins, and the player did well during the match (achieving a high Match Score), the player will rise in skill rating.
  
• If a players' team wins, but the player did not perform well (achieving a low Match Score), the player will not move in terms of skill rating.
  
• If a players' team loses, but the player does exceptionally well (achieving a very high Match Score), the player will go up slightly in skill rating.
  
• If a players' team loses, but the player performed well (achieving a high Match Score), the player will not move in skill rating.
  
• If the players' team loses, and the player performed poorly (achieving a low Match Score), they will drop in skill rating.

so looks like "doing well" means 200 damage - in case of winning, but in terms of losing "doing well" is not even achieved with 450 damage. In fact it means you could efficiently have killed 2 or mechs. LOL Instead those damage is said to be "performed poorly"

edit: also winning nets 20 score. so you only need to win and do 160 damage to get "up" since I guess 100 score is probably what starts to "put you up"


OUCH that screams FLAWED to my ears.

Edited by Lily from animove, 23 September 2015 - 01:42 AM.

[Kotzi]

Sounds great. If you are stomped you will most certainly loose your tier. If you stomp you will gain a tier.Those 2 will mostly be the case. Anything else is neglectible. Not very accurate imho.

[TWIAFU]

riverslq, on 22 September 2015 - 01:32 PM, said:

what is more of the determining factor then? win or a loss?
I think we need to see some low scores on wins (aka, spiders ) and see how it changes things

Lily started doing something that I wanted to know about myself, so, started getting info.

Win/Loss(W/L) | Match Score | Up/Down/Even (U/D/E)

W | 50 | E
L | 257 | E
L | 300 | U

I have many more, just got to find notes.

Basically, if you win and do a minimal score, you will go up.

Loose, and you have to preform very well, about 300 match score to go up, about 250ish to break even.

[PyckenZot]

• If a players' team wins, and the player did well during the match (achieving a high Match Score), the player will rise in skill rating.
  
• If a players' team wins, but the player did not perform well (achieving a low Match Score), the player will not move in terms of skill rating.
  
• If a players' team loses, but the player does exceptionally well (achieving a very high Match Score), the player will go up slightly in skill rating.
  
• If a players' team loses, but the player performed well (achieving a high Match Score), the player will not move in skill rating.
  
• If the players' team loses, and the player performed poorly (achieving a low Match Score), they will drop in skill rating.

I read this as personal score in the drop not being of that much effect, but rather your relative scoring against the others in your team.

[Lily from animove]

PyckenZot, on 23 September 2015 - 03:41 AM, said:

• If a players' team wins, and the player did well during the match (achieving a high Match Score), the player will rise in skill rating.
  
• If a players' team wins, but the player did not perform well (achieving a low Match Score), the player will not move in terms of skill rating.
  
• If a players' team loses, but the player does exceptionally well (achieving a very high Match Score), the player will go up slightly in skill rating.
  
• If a players' team loses, but the player performed well (achieving a high Match Score), the player will not move in skill rating.
  
• If the players' team loses, and the player performed poorly (achieving a low Match Score), they will drop in skill rating.

I read this as personal score in the drop not being of that much effect, but rather your relative scoring against the others in your team.

that could be as well, yet havign 106 score in the winning team is nothignI ever would consider as a "up" because thats literally impossible since evryone would have to be quite "bad" and then its questionable if anyone even should have been "upped".

But if it is that much in comparison to others, we would need the entire Scoretables to see a pattern.

[SirMad]

https://www.reddit.c..._to_settle_out/

[Speedy Plysitkos]

Ach 200 damage = green
mal-R 400 damage = red.

WHY ?

[Thorqemada]

Sean Langs 106 matchscore was based on 5 Assists and 123 damage...

[hero2zero]

I just got a score of 307 (461 damage) in a loss, PSR stayed the same ('=') in tier 3 (2/3 full bar).

[Ghogiel]

Just do 300 match score on a loss and it'll stay the same.

[El Bandito]

As I have more than 1:1 WLR, win based tier climb is fine with me!

As for me, this is what I had to do to not fall in rankings in a complete loss.

[Koshirou]

Does CW figure into the tier ranking?

[Yellonet]

Titannium, on 23 September 2015 - 04:03 AM, said:

Ach 200 damage = green
mal-R 400 damage = red.

WHY ?

Because doing damage in ACH is very difficult.






LOL.

[hero2zero]

I would think I had done enough to raise my PSR with this next loss...

[MechWarrior3671771]

"OUCH that screams FLAWED to my ears."

I am seeing players rant at their team now, more often than before. That's not good.

PGI may want to reconsider tying PSR to Wins, as chat is starting to get as toxic as it was during that kill/assist/win challenge.

Although I am seeing more careful play and less Leroy - peeps aren't throwing away matches as much.

[Sarlic]

Fenrisulvyn, on 23 September 2015 - 05:47 AM, said:

"OUCH that screams FLAWED to my ears."

I am seeing players rant at their team now, more often than before. That's not good.

PGI may want to reconsider tying PSR to Wins, as chat is starting to get as toxic as it was during that kill/assist/win challenge.

Although I am seeing more careful play and less Leroy - peeps aren't throwing away matches as much.

I haven't seen any rant yet.

Pretty sad if somoene rants over PSR.. There will always be a few rotten apples.

[Mawai]

Thorqemada, on 23 September 2015 - 12:56 AM, said:

Winning seems very urgent - like i wrote in a different Thread - Sean Lang (Tier 1) got a Green Arrow for Matchscore of 106 in a won Group-Q-Match i saw on Twitch.

Imho that is ridiculous low!

PS - Proof at the ~25:30 mark for that matchscore and PSR increase:
http://www.twitch.tv/ngngtv/v/17255295

If you read what PGI said about the new system ...

http://mwomercs.com/...30-18-aug-2015/

"First off, the win/loss condition plays a much smaller role compared to the current Elo system."


With the new PSR system:

• If a players' team wins, and the player did well during the match (achieving a high Match Score), the player will rise in skill rating.
  
• If a players' team wins, but the player did not perform well (achieving a low Match Score), the player will not move in terms of skill rating.
  
• If a players' team loses, but the player does exceptionally well (achieving a very high Match Score), the player will go up slightly in skill rating.
  
• If a players' team loses, but the player performed well (achieving a high Match Score), the player will not move in skill rating.
  
• If the players' team loses, and the player performed poorly (achieving a low Match Score), they will drop in skill rating.

If you win, your PSR goes up or stays even. If you lose, your PSR can go up (with an exceptional match), stay even or drop.

Win/loss is the primary determinant of direction of change in PSR. The magnitude of the match score is the determinant of how much of a change occurs in your PSR.

Finally, if PGI implemented the system correctly then it must also be CONSERVATIVE. This means that the total change in PSR summed over all players in any match MUST be zero. Otherwise you will get PSR inflation over time and eventually it will be pretty meaningless. This means that if the good players went up by 100 points then the rest have to go down by 100. What this means is that a match score of 100 for a win could result in a different change in your PSR in different matches depending on how everyone else did in the match.

There are several ways that PGI could have done the change in PSR. They could say that the total change in PSR for a match is +100 on one side and -100 on the other. They then distribute this fixed number among all the players depending on their match scores and whether they won or lost. This is probably the easiest to do and by changing the fixed value of PSR change they can increase or decrease the net change in PSR and how many matches it might take to converge a player's value. The other approach is to calculate some sort of absolute change in PSR based on match score for either the winning or losing team and then distribute that number of points to the other team based on their match scores. This would result in a certain match score in a win always getting the same number of points for example ... but the numbers allocated to those who lost points would vary.

The last (and broken way) to do PSR change is to calculate an absolute change in PSR based on match score and win/loss. At first glance this would sound like a good approach but the problem is that the system isn't conservative. Over time the total PSR will either go up or down ... it would be almost impossible to balance such a system to avoid this problem and defining tiers then becomes an issue unless you constantly change the tier boundaries (which is possible if they define tiers to have roughly equal numbers of players rather than a particular range of PSR.)

[Fate 6]

Thorqemada, on 23 September 2015 - 12:56 AM, said:

Winning seems very urgent - like i wrote in a different Thread - Sean Lang (Tier 1) got a Green Arrow for Matchscore of 106 in a won Group-Q-Match i saw on Twitch.

Imho that is ridiculous low!

PS - Proof at the ~25:30 mark for that matchscore and PSR increase:
http://www.twitch.tv/ngngtv/v/17255295

The system definitely needs some fine tuning. Heck, their algorithm for initial placement was off from the start (I'm easily a better pilot than Sean Lang and I am a few pixels shy of Tier 1), so we'll have to wait a good month or so before players get to their true Tier. The only problem is that, like you said, a crap match score on a win still brings you up.

The good news is that it'll be easy for me to raise my Tier to where it should be

[4rcs1ne]

FWIW, I did 851 damage and got 2 kills in a Cicada last night on a loss. My PSR went up.

[Nothing Whatsoever]

Another element I curious about is the population distribution, and how that might look over time.

Assuming PSR hits the nail on the head, then I'd guess we'd see a bell curve.

IIRC, the initial setup was even distribution, and I wonder how long that lasts if that was the case (Not sure need to try more googling to see if anything has been stated).

What I figure is highly likely to happen is the upper Tiers are gonna balloon over time with how the game is currently balanced and depending on how much actually changes with the re-balancing.

First is bell curve, narrow at the top and bottom, 2nd is even distribution, and then with population growth (Steam release) filling of the upper tiers, knowing that brand new accounts are being reported to be upper Tier 4.