135 replies to this topic

[Fate 6]

Roadkill, on 22 September 2015 - 03:05 PM, said:

More like 80%. I didn't check my bad games, but in all of my good games the non-damage component was ~20% of match score.

Under the old system damage was more like 65% of match score, so its reliance on damage has gotten worse.

I said PSR is 70% reliant on match score, not that match score is 70% reliant on damage. PSR does not go down if you win, meaning that if you win 50% of the time statistically your PSR will go up. So yes, damage is important in how your PSR changes, but winning is actually still the most important part.

Roadkill, on 22 September 2015 - 03:02 PM, said:

Right. It's match score based with some W/L thrown in for sanity.

Match score is primarily damage based.

Ergo, PSR is basically damage mitigated by W/L.

It's not W/L though. It's Match Score with a win or loss modifier. That modifier is pretty important though (it ends up being the equivalent of something like 200 match score which is a lot of damage).

Edited by Fate 6, 22 September 2015 - 08:45 PM.

[IraqiWalker]

Ultimatum X, on 22 September 2015 - 07:07 AM, said:

First, some detailed info collected & organized by McGral18:








Then there is the original description we have of what PSR is from the August 18th patch notes news.





"As a replacement for the original Elo-based rating and matching system, the new Pilot Skill Rating system releasing in this patch will provide more accurate and diverse skill ratings to better reflect your individual level of performance match-to-match. Whereas your original Elo scores were determined solely by your win/loss rate in a given weight class, the new PSR system is primarily influenced by your Match Score. "





TL;DR:

• PSR is primarily influenced by your match score
  
• Match score is weighted very heavily by damage dealt

Therefore PSR is heavily weighted by damage dealt.





What's your opinion of this?

They also mentioned that there is a multiplier we're not seeing, that actually heavily weights other actions, such as scouting.

[Wintersdark]

Mechwarrior Buddah, on 22 September 2015 - 08:44 PM, said:

Hell yes its flawed, once you hit tier 3, youre matched with every other tier. Whats the point of a tier system then?

you CAN BE, not are. It still tries to match like tiers, and my experience so far is that matches tend to be full of a single tier or a range of 2.

It doesn't just mix everyone.

IraqiWalker, on 22 September 2015 - 08:51 PM, said:

They also mentioned that there is a multiplier we're not seeing, that actually heavily weights other actions, such as scouting.

they said that, but there is no indication its actually happening.

[nehebkau]

IMHO solo kills should be 25, points from damage should be %accuracy * damage done

[Karl Streiger]

nehebkau, on 23 September 2015 - 07:12 AM, said:

IMHO solo kills should be 25, points from damage should be %accuracy * damage done

kill efficiency and the honorable kills - good idea

[Lootee]

And yet the only thing that matters is dealing damage.

On Mordor the lemmings that march into Mount Doom for a slugfest are considered more 'skilled' than the Spider who evades the enemies and 5-caps all the resource points and wins the game for the team after all friendlies died in the volcano slugfest.

At the very least the contribution of the Spider should count as much as the brawlers. Sadly it doesn't even though the Spider did the carrying that game.

Call this broken system what it is, Player Damage Rating.

Edited by Lootee, 23 September 2015 - 07:32 AM.

[Roadkill]

Fate 6, on 22 September 2015 - 08:47 PM, said:

I said PSR is 70% reliant on match score, not that match score is 70% reliant on damage. PSR does not go down if you win, meaning that if you win 50% of the time statistically your PSR will go up. So yes, damage is important in how your PSR changes, but winning is actually still the most important part.

PSR can go down if you win. PGI stated that it can go down in a win if you suck bad enough, and that it can go up in a loss if you carry hard enough.

Therefore match score, e.g. damage, is the most important component in PSR.

Quote

It's not W/L though. It's Match Score with a win or loss modifier.

That's basically exactly what I said, except that I also pointed out that Match Score is primarily based on damage (roughly 80%) which means that PSR is damage with a W/L modifier.

Since it appears you were trying to conter-point what I said, but then re-stated what I said, I can't tell if you're agreeing with me or not?

[Donny Line Z]

so damage is a big deal, you need it to win .... thing is the longer I played the more my damage went down and my kills and wins went up. I even find times when I can charge out like a noob on crack drawing the reds into an unwinnable position at the expense of my own mech I must admit. You see I like to win, and I am big on teamwork. so is there a sweet spot where pgi can see the shift from bathing the bad guys with damage to drilling out the important bits, like hunting and killing the dual gauss jagur ,or chasin down that LRM boat , no dam for the hunt or the run but could win the game, that's a good pilot skill.

[Ultimax]

IraqiWalker, on 22 September 2015 - 08:51 PM, said:

They also mentioned that there is a multiplier we're not seeing, that actually heavily weights other actions, such as scouting.

All you need to do is watch match scores at the end for your scout mech to see that it's a pittance.

I also seriously doubt that the tiny scouting bonuses that barely affect match score are being recorded separately in a database and then re-used as a multiplier for PSR.

[r4p70r]

Hey, i dont know how to ask my question so you can understand what i mean because of my english skills.
My question is, how is it measured if you doing good or bad in a match, is there a threshold, lets say 500 ms is good and so on or is it how your scores are in relation to your teammembers?
Hope its not to bad grammar.

[DashFire61]

Mister Blastman, on 22 September 2015 - 07:33 AM, said:

Also--folks who DO farm damage will get what they deserve. They'll be thrust into tier 1 where they'll get their assess handed to them. I'm pretty darn sure the players at the top don't care about farming damage. They care about winning. And winning is the opposite of farming.



What's wrong with thinking you're awesome because you do 200 damage and get four kills? I can think of a lot of things...

You obviously arent an efficient player then, because getting 6 kills and only doing 200-300 dmg means im better at cockpitting and coring rear than you are.



What they should do, is add a bonus for having a higher hull integrity at the end of the match, if you have 80% or greater armor you get an effectiveness modifier.

[Deathlike]

To answer the topic question: Yes
https://www.reddit.c..._to_settle_out/

[CtrlAltWheee]

Doesn't seem to reward things like narcing or spotting.

Agree there is room for improvement.

[Mister Blastman]

Lt James Steiner, on 23 September 2015 - 02:08 PM, said:

What they should do, is add a bonus for having a higher hull integrity at the end of the match, if you have 80% or greater armor you get an effectiveness modifier.

Good players--no, the best ones, soak damage for their team and often end a drop with 25 - 40% of their 'mech integrity left.

[fat4eyes]

Grisbane, on 23 September 2015 - 02:51 PM, said:

to be honest, am glad i am not in tier 1 or 2.. matches that have only tier 1 or 2 players are easy to spot.. the entire team is in a stormcrow, arctic cheetah, timberwolf, hellbringer, ebon jaguar, or Dire Wolf.. always running some version of gauss vomit or laser vomit.. basically very boring, repetitive play with no deviation from the meta whatsoever. if i were tier 1 or 2.. i would probably get bored of constantly playing the same thing over and over and quit outright. thus why i purposely lower my psr to the lower tiers.. you never know what builds you may fight next, or what mechs.. it's simply more interesting.

You'd be surprised how different games are at higher tiers. I can only speak for tier 2, but depending on the map, there's a lot more movement and actual flanking (i.e. part of the team pins down the enemy while the rest go around and flank). Envelopments are more likely to lead to victory than sheer attrition. And games are fast, usually over at 5-7 minutes. There's a lot more sitzkreig at tier 4, unless you yourself do something to goad the team forward.

The Meta is more diverse than you think (though it does have a LOT of pulse lasers). A lot of IS mechs still get played. Dakka Maulers and Crabs are legitimately scary. Some players are really, REALLY good at SRM brawling, even in lights.

Don't give up getting to the higher tiers, the gameplay and tactic changes are worth seeing.

[Bloody]

the rating is completely ********, i got caught being semi afk, got 60 score and WENT UP cos my team won,. i am like wut?

the rating is completely ********, i got caught being semi afk, got 60 score and WENT UP cos my team won,. i am like wut?

[Karl Streiger]

...ok after Pauls last post made it official - PSR rating doesn't differ much from ELO - victory / loss is most important

It also shows flaws in direct comparison with ELO.

Look we had a minimal range in ELO difference in the group queue of 150? At a time where it was not possible to choose the mission type.
That was a huge difference - a team of 12 stomping over several smaller groups - but while this was not important for ELO - hey the MM bet that you would not win - this stomp is archived and it did have a huge impact on your current PSR rating.
Same for the winners - while stomping would not have riseen there ELO - their PSR rating got a boost.

so using archived data was the first mistake

Ok an while i don't know exactly how the MM works - i guess - Tier4 player see if i get 24 Tier4 players - no take some T3 - well 2 slots remaining - ah a waiting T1 and a T2.
But it is unlikely that the Tier of a player in comparison with the others have any impact. A T1 player in T4 pool (as a extreme should have the highest ms - the rise and fall ranking should be adjusted)

[IraqiWalker]

Karl Streiger, on 24 September 2015 - 11:48 PM, said:

...ok after Pauls last post made it official - PSR rating doesn't differ much from ELO - victory / loss is most important

It also shows flaws in direct comparison with ELO.

Look we had a minimal range in ELO difference in the group queue of 150? At a time where it was not possible to choose the mission type.
That was a huge difference - a team of 12 stomping over several smaller groups - but while this was not important for ELO - hey the MM bet that you would not win - this stomp is archived and it did have a huge impact on your current PSR rating.
Same for the winners - while stomping would not have riseen there ELO - their PSR rating got a boost.

so using archived data was the first mistake

Ok an while i don't know exactly how the MM works - i guess - Tier4 player see if i get 24 Tier4 players - no take some T3 - well 2 slots remaining - ah a waiting T1 and a T2.
But it is unlikely that the Tier of a player in comparison with the others have any impact. A T1 player in T4 pool (as a extreme should have the highest ms - the rise and fall ranking should be adjusted)

You can only see 2 tiers above, and below yours.

Meaning if you are a tier 4, you can only be matched against players from tiers 5-2

If you are tier 5, you only see tiers 5-3, if you are tier 2 you can go from 4-1, tier 1 players can only see 3-1. Only tier 3 players have access to all 5 tiers.

[EgoSlayer]

IraqiWalker, on 25 September 2015 - 07:10 PM, said:

You can only see 2 tiers above, and below yours.

Meaning if you are a tier 4, you can only be matched against players from tiers 5-2

If you are tier 5, you only see tiers 5-3, if you are tier 2 you can go from 4-1, tier 1 players can only see 3-1. Only tier 3 players have access to all 5 tiers.

They have never said it's only +/-2, only that Tier 1 will never face tier 4 or 5. Tiers 2 and 3 can play anyone, Tier 1 can only play 1, 2 and 3s. And that only applies to solo queues, because in group queue players can create groups that include members from any tier and the groups are averaged.

[Karl Streiger]

And of course this works only when there are enough players - otherwise I'm pretty sure the typical Mm behaviour will remove those borders.
Result : the known ugly face - long wait times and bad games.
Think it would be simpler to analyse the current numbers of players - and of course PGI need an average number of players the MM need for a good game.

If the numbers don't match the player got a message - so he could check his game mode / and server

Edited by Karl Streiger, 25 September 2015 - 10:31 PM.