76 replies to this topic

[Tristan Winter]

In the latest Town Hall, Russ explained that PSR will be adjusted to put more emphasis on personal accomplishments as opposed to winning matches, because this will reward the more skilled players.

Prediction: When you adjust any rewards (C-bills, XP, PSR) with more emphasis on personal accomplishments (e.g. damage or kills) and less emphasis on victory, you increase incentive for selfish behavior and reduce incentive for teamwork.

Have you been in matches where 2-3 harassing light mechs manage to get around and behind the enemy, causing a panic and then distracting half the enemy team, so their teammates are basically fighting a 10 vs 6 battle on the front line? Or matches where someone rushing back to stop a base cap bought your team enough time to either counter-cap or defend base?

In the solo queue, those kinds of selfless acts aren't exactly standard, but they'll be even more rare if you put less emphasis on winning the match.

Altruistic suicide doesn't pay C-bills, it doesn't give you XP, and your precious WLR stats are invisible to everyone but yourself. So right now, the only incentive is a slight PSR increase if you can get a match score above 100. However, if PGI puts more emphasis on damage and kills, then it's probably going to be easier to just accept defeat and focus on shooting bad guys. In the long run, it's going to give you more C-bills, more XP and higher PSR.

So yeah. I know everyone doesn't like PSR, but even if we accept that it's here to stay, it seems unwise to put so much emphasis on personal damage and kills.

TL;DR - On a match-by-match basis, it probably seems more fair to reward the guy who did 500 dmg on a losing team. He may have carried like a boss. But in the long run, rewarding those players more than the guys who consistently win matches will increase selfish behavior in the solo queue.

[Sandpit]

I don't get things like this. Russ specifically stated they're looking at making a minor adjustment to give a little less weight to W/L and a little more to personal match score.

It's not like they said they're tossing out the formula in favor of individual match scores. It's not going to increase "selfish" behavior, it's going to reward the players who get stuck on team derp, do really well individually, but have zero chance to win the match because 3 guys on their team decided to rambo off and die 30 seconds into the match.

THAT'S selfish behavior.

[Metus regem]

I'd like to see stronger rewards for team work, like spotting for support mechs, scouting for light mechs, capping bases and control points. I'd love to see rewards for bringing AMS to a match... It's funny when I pull out a light mech, that does all of those things, I get punished on my match score and rewards, because I'm doing what a light mech should be doing, and rather than rewarding the mech for doing what it should be doing, they are only fostering the current problem where all that matter is damage, as damage accounts for most of your match score.

[Hit the Deck]

Do people actually actively try to increase their PSR with regularity? How many of them do it? This relates to your prediction whether PSR actually has any influence on players behavior. Events seem to have some effect on it but I'm not sure about PSR mechanics.

Edited by Hit the Deck, 02 March 2016 - 10:21 AM.

[Mystere]

Metus regem, on 02 March 2016 - 10:13 AM, said:

I'd like to see stronger rewards for team work, like spotting for support mechs, scouting for light mechs, capping bases and control points. I'd love to see rewards for bringing AMS to a match... It's funny when I pull out a light mech, that does all of those things, I get punished on my match score and rewards, because I'm doing what a light mech should be doing, and rather than rewarding the mech for doing what it should be doing, they are only fostering the current problem where all that matter is damage, as damage accounts for most of your match score.

This is why I think PSR should take wins and losses more into consideration, not less as is being demanded by people.

[STEF_]

PSR total should be zero: in this way the skilled goes up, the potato goes down.

Then, not only less emphasis on victory, but in defeat too.

It is absurd I tried to carry the team, I did dmg and assists.... and psr stays equal.

Mystere, on 02 March 2016 - 10:18 AM, said:

This is why I think PSR should take wins and losses more into consideration, not less as is being demanded by people.

ELO was taking in consideration w/l only. And it was a really bad system

[L A V A]

Stefka Kerensky, on 02 March 2016 - 10:21 AM, said:

ELO was taking in consideration w/l only. And it was a really bad system

One that really only rewards damage isn't much better.

[Sandpit]

Metus regem, on 02 March 2016 - 10:13 AM, said:

I'd like to see stronger rewards for team work, like spotting for support mechs, scouting for light mechs, capping bases and control points. I'd love to see rewards for bringing AMS to a match... It's funny when I pull out a light mech, that does all of those things, I get punished on my match score and rewards, because I'm doing what a light mech should be doing, and rather than rewarding the mech for doing what it should be doing, they are only fostering the current problem where all that matter is damage, as damage accounts for most of your match score.

Same here, although Russ did say they're adding rewards for things like AMS shooting down missiles and such, that's a good step in that direction. Rewarding players for taking support systems like that.

L A V A, on 02 March 2016 - 10:26 AM, said:

One that really only rewards damage isn't much better.

Which PSR does not do

[STEF_]

L A V A, on 02 March 2016 - 10:26 AM, said:

One that really only rewards damage isn't much better.

it can be twicked.
Dmg, and kills, is what makes you win.
But not only: I would put good matchscore to tag/ams/ecm covering, shooting down UAV. etc.
Every contribution should be rewarded in adequate way.

W or L is the result of the team efforts, but can be very penalizing for a single or too much easy for the ones who like to be carried. So, imo it's more fair to reward efforts than the results

Edited by Stefka Kerensky, 02 March 2016 - 10:36 AM.

[Kodyn]

So wait, they're making it even easier?

Now everyone really is going to end up T1 eventually, it's already quite simple, it's just putting in the time. I look at hitting T1 now as my warning that I've been playing too much, make it any easier, and I won't have that bar.

[Tristan Winter]

Hit the Deck, on 02 March 2016 - 10:13 AM, said:

Do people actually actively try to increase their PSR with regularity? How many of them do it? This relates to your prediction whether PSR actually has any influence on players behavior. Events seem to have some effect on it but I'm not sure about PSR mechanics.

Some do, some don't. Hard to say anything about this without guessing, but let's use game theory and look at it from a different perspective.

You're a fast mech.
Your base is being capped.
If you go back to defend, your team may win but you will probably get focus fired and killed really fast. You won't have time to do much damage. At best, you get a 100+ match score, at worst you get a 0-100 match score.
If you ignore base and focus on killing, your team will probably lose, but you will get to do more damage.

So what are the rewards for defending? No extra c-bills, no extra XP, but maybe you get the win. What is a win good for? Nothing, except PSR.
So what are the rewards for attacking? You get extra c-bills, extra XP, but unless you get a 400+ match score, you won't get a higher PSR rating.

Now, if you make it easier to gain PSR rating in defeat, you basically eliminate the one single reason anyone in a fast mech would ever want to defend. Because really, who cares about the honour of a win or a loss when each match lasts 5 minutes? It's meaningless, in the long run. You have to look at the rewards.

[Mystere]

Stefka Kerensky, on 02 March 2016 - 10:21 AM, said:

PSR total should be zero: in this way the skilled goes up, the potato goes down.

Then, not only less emphasis on victory, but in defeat too.

It is absurd I tried to carry the team, I did dmg and assists.... and psr stays equal.

How are people who play vanguard rated?
Are spotters richly rewarded?
How do you reward people who consistently divide, disrupt, confuse, and delay the enemy enough for the rest to take advantage and win, but ends up having poor damage scores and/or gets killed for the effort?

Stefka Kerensky, on 02 March 2016 - 10:21 AM, said:

ELO was taking in consideration w/l only. And it was a really bad system

I say at best that is debatable, especially because there are those who think PSR is also a really bad system because it encourages stat-whoring even more.

Edited by Mystere, 02 March 2016 - 10:46 AM.

[Metus regem]

Mystere, on 02 March 2016 - 10:43 AM, said:

How are people who play vanguard rated?
Are spotters richly rewarded?
How do you reward people who consistently divide, disrupt, confuse, and delay the enemy enough for the rest to take advantage and win, but ends up having poor damage scores and/or gets killed for the effort?

Poorly, I can attest to that....

[STEF_]

Mystere, on 02 March 2016 - 10:43 AM, said:

I say at best that is debatable, especially because there are those who think PSR is also a really bad system because it encourages stat-whoring even more.

But also psr (the one we have now) is pointless because sooner or later everybody can be tier 3 and matched up with everyone.
PSR numbers are going up.
That's why previously I first posted the more important thing in my opinion: psr total of all 24 players must be zero.


So even if a player "wins", his psr must go down if he did nothing for the win.

Beside, I agree with L A V A: dmg cannot be so damn important in determining matchscore.

Edited by Stefka Kerensky, 02 March 2016 - 10:59 AM.

[Raso]

Look I'm gonna level with you all.... I don't want to play in higher tiers. I have zero desire to do so. The amount of mental and social efforts needed to enjoy my time in those tiers is beyond my current capabilities as an almost functioning adult with a career and family. I dig losing because it makes my bar go down. I count on the poor performance of my fellow PUGers to keep my PSR fairly neutral. In return I fight hard and play hard and help those PUGers get thenodd win. Everyone has a generally great time.
Don't pin my PSR on me. I don't want to be leet. I don't want to have to boat meta trash builds. I like my LRMs and AC2s and LBX10s. Don't ruin that for me.

Edited by Raso, 02 March 2016 - 10:58 AM.

[Mystere]

Stefka Kerensky, on 02 March 2016 - 10:55 AM, said:

So even if a player "wins", his psr must go down if he did nothing for the win.

And which is why I asked these very questions:

Mystere, on 02 March 2016 - 10:43 AM, said:

How are people who play vanguard rated?
Are spotters richly rewarded?
How do you reward people who consistently divide, disrupt, confuse, and delay the enemy enough for the rest to take advantage and win, but ends up having poor damage scores and/or gets killed for the effort?

[PEEFsmash]

Actually, the -only- thing that should count is wins and losses. You are the only constant factor in your games. The way you win doesn't have to be through damage. The damage role is not for every mech. If you do your part, whatever your part is and no matter how intangible thru stats (like spotting a push on seismic and calling it out, directing friendlies to cap or counter-cap, holding locks, etc), if you are pulling a high winrate, the factor in the long term must be you. The only factor that should matter is your impact on winning the game, which can only be measured by whether you win or not.

[STEF_]

Mystere, on 02 March 2016 - 10:43 AM, said:

How are people who play vanguard rated?
Are spotters richly rewarded?
How do you reward people who consistently divide, disrupt, confuse, and delay the enemy enough for the rest to take advantage and win, but ends up having poor damage scores and/or gets killed for the effort?

Stefka Kerensky, on 02 March 2016 - 10:35 AM, said:

it can be twicked.
Dmg, and kills, is what makes you win.
But not only: I would put good matchscore to tag/ams/ecm covering, shooting down UAV. etc.
Every contribution should be rewarded in adequate way.

Imo, MWO already has the means/instruments to try to track game style.
It's already possible to reward spotters/scouts
About your example: "How do you reward people who consistently divide, disrupt, confuse, and delay the enemy enough for the rest to take advantage and win, but ends up having poor damage scores and/or gets killed for the effort?" it is already possible to track mech positions (for instance, there already are lance rewards).
Those teamates, cutting off 2-3 mechs from the rest of the team, should be rewarded, indeed.
It is possible to track those 2-3 enemy mechs cutted off from the rest and also friendly units engaging them.

[Sjorpha]

In my opinion the problem with measuring performance through results other than winning and losing is not that it would increase or decrease any given behavior.

I really doubt it would, unlike rewards which are a tangible and direct carrot&stick type incentive, the calculation of PSR changes aren't tangible or direct. Some players might reflect on them from time to time, but I don't think it would change behavior, especially since the advantages of increasing or decreasing are pretty fuzzy as well.

Instead, the problem is that it is just a bad type of measurement:

It doesn't measure your contribution to the teams coordination, via leadership or otherwise, which can be one of the most important to a win and may even mean more than your scores.

It doesn't measure your timing and accuracy, so we don't know if the damage and kills were strong contributions to the win, only that they happened.

Even worse: It removes the only way to capture those aspects, because if we are no longer measuring the players trends towards winning and losing, we can't know if his scores and unmeasured behavior actually contributed or not.

Consider for a moment a player that is a great leader, and therefore has a very high win rate because he force multiplies whatever skill pool he leads with good coordination. Let's assume this player is also bad at shooting, he almost never gets good scores in terms of damage and kills.

Under a system that primarily measures scores, this player would stay low tier where he will unbalance every match he drops in, they will mostly be winning stomps because in reality he an amazing player that brings his contribution in the form of leadership.

Under a system that measures overall tendency to win, which is to say that it measure how much more likely a team with player X is to win than a team with player Y is, this great leader will propel correctly up the tiers to where great leaders should be.

Leadership is of course not the only attribute that suffers from this kind of invisibility under reductive measuring systems, they don't just reduce the result to less than the sum of it's parts, it even leaves parts out!

Your average impact on your teams chance to win is the only data point that captures the sum of your contribution to a match, it is therefore the only data point you CAN use if you want accurate skill measurements, with the actual PSR changes weighed by the predicted chance to win the match in the first place. That is how ELO works, and there are other formulas that work this way as well.

Now any choice to use other data points are necessarily inaccurate, but the benefit may be a quicker arrival at the desired measurement. In other words ELO was to slow for PGI, so they reduced inaccuracy in favor of speed. They also deliberately added more inaccuracy by giving the system a bias towards PSR increases, which is truly mindbogglingly stupid.

If you are suggesting further removal of the win datapoint you are in effect suggesting a further decrease of accuracy and a further removal of any chance to measure meta elements like leadership, strategy and timing.

How large part of being good at this game comes down to leadership, strategy and timing? 50%?, 70%? More?

Well that is the part you are suggesting we should no longer factor into matchmaking if you want to remove winning from he equation.

Edited by Sjorpha, 02 March 2016 - 12:07 PM.

[Sandpit]

Sjorpha, on 02 March 2016 - 12:05 PM, said:

Well that is the part you are suggesting we should no longer factor into matchmaking if you want to remove winning from he equation.

maybe I missed it, can you show me where anyone is suggesting W/L not be considered in the formula for PSR?