71 replies to this topic

[Duke Nedo]

Paigan, on 28 January 2016 - 01:59 AM, said:

Highly unfair:
Say you have 12 super guys stomping 12 noobs:

The 6 super guys with score 500-550 get uprated, while the 6 super guys with score 450-500 get downrated.
Did the 6 "lower" guys really play so bad that they deserve a downrate?

Same on the other side:
The 6 noobs with score 20-30 get rewarded for having performed SO well.
The 6 noobs with score 0-20 get downrated because they suck so much.


My observation:
People are REALLY quick to come up with extremely short-sighted, half-baked solutions for complex problems.
For a proper solution, see the link provided in my post and sig.

I agree that relevant damage > total damage, but even so, that is also simplified... even with perfect scoring for doing objectives etc, a good score in a very homogenous game may be a bad score in a heterogenous game... your team mates and your opponents have a lot to say about whether your score is good or bad... at the end of the day, the only thing that matters is whether your team won or lost. But then we're back at Elo and that is horribly insensitive in a 12v12 game with random team compositions from one game to the other.

I think we just have to settle for a solution that is as little bad as possible. Imo winning must be the most important objective, as in really really must be, in order to not create foul game play. Some degree of PSR is also nice to speed up the process of finding out whom out of these 12 players was actually driving the win, but that PSR component will never be perfect.

Edited by Duke Nedo, 28 January 2016 - 02:15 AM.

[Kilo 40]

Paigan, on 28 January 2016 - 01:59 AM, said:

Highly unfair:
Say you have 12 super guys stomping 12 noobs:

The 6 super guys with score 500-550 get uprated, while the 6 super guys with score 450-500 get downrated.
Did the 6 "lower" guys really play so bad that they deserve a downrate?

Same on the other side:
The 6 noobs with score 20-30 get rewarded for having performed SO well.
The 6 noobs with score 0-20 get downrated because they suck so much.


My observation:
People are REALLY quick to come up with extremely short-sighted, half-baked solutions for complex problems.
For a proper solution, see the link provided in my post and sig.

no no no....

I mean across both teams. the top six players, from both teams increase, while the lowest 6 across both teams decrease. so using your example, The 6 super guys with score 500-550 get uprated, while the 6 super guys with score 450-500 and The 6 noobs with score 20-30 stay the same and the 6 noobs with score 0-20 get downrated because they suck so much.

also your example sucks and isn't very realistic.

[Onimusha shin]

TexAce, on 28 January 2016 - 01:41 AM, said:

(to be fair with myself I tried an unskilled Hunchy 4H with bad loadout and got distracted by boobs early).

To be fair, that's the only thing that matters. Nothing matters more than bewbs, especially not over some online game. =PPP

[Haakon Magnusson]

Lord0fHats, on 27 January 2016 - 11:33 PM, said:

I think the change that's really needed is that people should lose PSR even when their team wins if they do poorly enough. That's the real issue. The outcomes of a match are;

Win, Gain points
Win, = Outcome
Lose, = Outcome
Lose, Lose points

Nominally, I think people make too much of the 'exp bar' bit. If someone did well enough to not lose points even in a loss, then they did pretty well. But really if someone did really really badly, I think they should still lose points even if their team won. Likewise, maybe people should gain points even if they lose if they exceptionally well.

I think the bigger issue is the way match score is calculated, namely the massive weight derived from damage done. Want an easy way to climb the PSR ladder? Just load up LRMs and SSRMs (don't we have a dozen threads on why there's so many LRM boats now?). These weapons are bad (very bad), but they can put up high damage numbers while requiring little if any skill to get the damage. Because of the weight for damage in Matchscore, putting up 600 dmg in missiles pretty much assures you won't lose PSR. It's 'fluff' damage. It looks good on paper, and really brings up your match score while not really helping the team much.


The tabulation of Matchscore really needs adjustment for a number of things I think (torso twisting should be rewarded) and that would help solve the current PSR issues.

While I think you are somewhat right having to adjust Pauls bars a bit



I believe the biggest thing we should be adjusting is the threshold where these moves happen.
We know that over 400 match score on a loss, and you psr goes up... on all tiers

This threshold (And all others also) should be modified per tier to be higher, something akin to this

t5: >400
t4: >440
t3: >480
t2: >520
t1: >560

For simplicity just the psr rise on loss, but changing the thresholds of rising on win is just as important, if not more

Of course the actual number would need to be based on something what PGI has (numbers) I just threw a guesstimate of flat increase to previous level. It might even be that points needed at t1 and t2 may need to be less, since you are competing over a finite resource of points in a tier with people who know how to kill efficiently.

Also maybe tweaked multiple times until population distribution resembles more a Bell curve, some at tiers 1 and 2 should start downward movement untill they are at level where they can stay.
Naturally the way match points are calculated is a deciding factor in this as well, but the threshold for improving per level really should rise some.

[Paigan]

Kilo 40, on 28 January 2016 - 02:18 AM, said:

no no no....

I mean across both teams. the top six players, from both teams increase, while the lowest 6 across both teams decrease. so using your example, The 6 super guys with score 500-550 get uprated, while the 6 super guys with score 450-500 and The 6 noobs with score 20-30 stay the same and the 6 noobs with score 0-20 get downrated because they suck so much.

also your example sucks and isn't very realistic.

It's an intentional exaggeration to make a point.
I could explain for pages and pages that performance has to be measured objectively (as objectively as possible) and not in comparison to the rest of the players. To shorten this, I used an exaggeration to show a problem in principle.

You should try to understand such rhetorical figures instead of instantly shouting "your example sucks".

However, your additional explanation was quite interesting. Still not very objective, but probably more suitable.

Edited by Paigan, 28 January 2016 - 04:49 AM.

[pyrocomp]

XX Sulla XX, on 27 January 2016 - 11:45 PM, said:

You are exactly right. Being able to go down on a win is a must. For example having some one playing in a group and they make like 5 damage each game and the rest of the team is making 400-900 each game and they are winning. 5 damage person should be going down.

Squirrels usually do little damage but give their team an advantage. The first mech in a push through a bottleneck is usually just a damage absorber dealing low damage and going down allowing his team to rush in (on therma in a dakka build you have a chance to do something before going down, but that's an exception). The squirel can try to gain some 'scouting' rewards directly (if there are not many ECM mechs nearby) or by popping and UAV, but what for pushers?
If you can mathematically reliably detect the 'squirrel' and 'pusher' cases in general low-damage-unwise-moving population using movement/shooting/postitions, then I'll agree with you. In current state - no. And there will be nascar-abandons that absorb damage without shooting back.
I can't come up with a solution that will be able to recognize divertions (yes, wasted shots and looking the other way) and suppressing fire (low damage but area denial) and reward it. Such things are and invesment into the teams general position and rewarding for a win even with a low score is a safe way to ensure that such things will still be used.

[Duke Nedo]

pyrocomp, on 28 January 2016 - 05:47 AM, said:

Squirrels usually do little damage but give their team an advantage. The first mech in a push through a bottleneck is usually just a damage absorber dealing low damage and going down allowing his team to rush in (on therma in a dakka build you have a chance to do something before going down, but that's an exception). The squirel can try to gain some 'scouting' rewards directly (if there are not many ECM mechs nearby) or by popping and UAV, but what for pushers?
If you can mathematically reliably detect the 'squirrel' and 'pusher' cases in general low-damage-unwise-moving population using movement/shooting/postitions, then I'll agree with you. In current state - no. And there will be nascar-abandons that absorb damage without shooting back.
I can't come up with a solution that will be able to recognize divertions (yes, wasted shots and looking the other way) and suppressing fire (low damage but area denial) and reward it. Such things are and invesment into the teams general position and rewarding for a win even with a low score is a safe way to ensure that such things will still be used.

Exactly, in a team based game "skill" = "ability to drive a win". It's very difficult to identify and quantify that... that said, Rewards 2.0, PSR and onwards is a huge improvement over what we used to have. I just want to stress that we don't move away from the over all aim, which is to win the match. If you can "farm" PSR without winning we risk encouraging really poor team play.

Edit: A parallell (a bit extreme, but just to prove a point) could be to compare to EVE where the FC may sit cloaked on the battlegrid just observing and calling targets and coordinating with other FCs etc, not firing a single shot but clearly being the one largely deciding the outcome of the battle. The skill needed for the "grunts" in Eve is mainly to click on the target and press F1... it's not that bad in MWO, but there is a hint of it, isn't there...? There are many flavors of important contributions that doesn't end up on the score board.

Edited by Duke Nedo, 28 January 2016 - 06:12 AM.

[M T]

Imho the closest statistics that (we players) can observe that remotely reflects skill is the accuracy stats.

And should be one of the main deal breakers for tier placement.

- Gauss, PPC's, one shot weapons. (Perhaps autocannons too)
- Lasors
- LRM/Streak SRM (Auto aim weapons)

Put them in three different classes to compare against different players.

Sadly PGI thinks that a PSR system with 10 variables (kills, assists, damage, match score, win, lose etc is an accurate reflection of skill, but its not. Having intentionally played a few matches solely aiming at L/R torso to pad my damage numbers (1350+) and having 0 kills is just one example of that. Can think of many more scenarios.

Ultimately im afraid this PSR will eventually have zero value as all the dedicated LURM per second damage farm boats will reach Tier 1 with plenty of fulfilled xp bar farm time.

Edited by x MT x, 28 January 2016 - 06:58 AM.

[JernauM]

Whatever the solution, PSR must be fixed. Otherwise, we're just going to experience progressively worse and worse matchmaking as everyone generally trends upwards in tier.

[MrMadguy]

Two simple changes needed:
1) PGI should decide, what behavior should be rewarded, how it should be rewarded, set this "rules" in stone and adjust match score accordingly. For example, it feels like, they are delaying MM improvements, just because they are unsure, how they should reward scouting.
2) Make MM purely performance-centric. And as rewards are proportional to match score and match score is proportional to performance - MM should depend on match score only. Performance -> Match Score -> Rewards.

W/L now is may be 90% of PSR. With new MM system Win will affect rating the same way, it affects your rewards. How much Match Score/CB/XP PGI gives us for Win? That's how much Win really "costs". Win isn't 90% of my rewards, sorry. It causes terrible imbalance between players: some are in PSR Paradise, having K/D = 3, dealing 600+ dmg per match and having CB/XP raining on them and some are in PSR Hell, so they barely can equip DHS on their 'Mechs. Both have W/L = ~1. But W/L doesn't mean anything - it's nothing, but number at your stat page. Only performance and rewards mean.

[Duke Nedo]

MrMadguy, on 28 January 2016 - 11:49 PM, said:

. But W/L doesn't mean anything - it's nothing, but number at your stat page. Only performance and rewards mean.

It's a team vs. team game, w/l should be heavily weighted otherwise you'll get lots of players only playing for their personal rewards and not for their team. There really needs to be incentives for "taking one for the team". Already now we have enough mechs trying to stand behind each other to have your teammate tank for you... we don't want more. For the team to win, everyone has to tank their share. To get as high personal score as possible, you want to tank as little as possible while dealing damage, and damage is king.

[Chagatay]

I am ok with the way it is. Is it perfect? No. But is it good enough? Yep.

[MrMadguy]

Duke Nedo, on 28 January 2016 - 11:58 PM, said:

It's a team vs. team game, w/l should be heavily weighted otherwise you'll get lots of players only playing for their personal rewards and not for their team. There really needs to be incentives for "taking one for the team". Already now we have enough mechs trying to stand behind each other to have your teammate tank for you... we don't want more. For the team to win, everyone has to tank their share. To get as high personal score as possible, you want to tank as little as possible while dealing damage, and damage is king.

Then Winning isn't being rewarded enough. Give 90% CB/XP solely for Win. But what will happen then? Some players will be gaining free rewards for doing nothing. PGI don't want to do it. But, as I said, it causes imbalance - rating is biased towards winning, but reward isn't. See "LRMs are OP" threads. Players are saying, that they are gaining ton of rewards and rating via playing LRM boats. I could do it too. But I don't want to play Meta 'Mechs only - I want to play various builds. So my rating should be adjusted according to my performance and rewards only. Meta players should play against Meta players only.

Edited by MrMadguy, 29 January 2016 - 12:20 AM.

[Ratpoison]

MrMadguy, on 29 January 2016 - 12:16 AM, said:

Then Winning isn't being rewarded enough. Give 90% CB/XP solely for Win. But what will happen then? Some players will be gaining free rewards for doing nothing. PGI don't want to do it. But, as I said, it causes imbalance - rating is biased towards winning, but reward isn't. See "LRMs are OP" threads. Players are saying, that they are gaining ton of rewards and rating via playing LRM boats. I could do it too. But I don't want to play Meta 'Mechs only - I want to play various builds. So my rating should be adjusted according to my performance and rewards only. Meta players should play against Meta players only.

Your ideas are always so entertainingly crazy.

[Duke Nedo]

MrMadguy, on 29 January 2016 - 12:16 AM, said:

Then Winning isn't being rewarded enough. Give 90% CB/XP solely for Win. But what will happen then? Some players will be gaining free rewards for doing nothing. PGI don't want to do it. But, as I said, it causes imbalance - rating is biased towards winning, but reward isn't. See "LRMs are OP" threads. Players are saying, that they are gaining ton of rewards and rating via playing LRM boats. I could do it too. But I don't want to play Meta 'Mechs only - I want to play various builds. So my rating should be adjusted according to my performance and rewards only. Meta players should play against Meta players only.

I think the rewards balance between winning and losing is rather OK now. My main complaint is that damage and killing blows are rewarded a bit too much relative to all the other entries. Just a bit though. I am not too worried by free-riders, if you have too many you won't win anyways. I don't think it's worth it to try to exclude them because you stand a big chance of instead hurting team oriented players that currently are too little rewarded for their contributions, like for example the unusual brave pilot leading the charge losing all his armor in the process and dying before dealing half as much damage as his team mates. I'd like to encourage that but it's a bit tricky to reward tanking (because also discos will "tank"). Perhaps reward taking damage near team mates or something like that.

[MrMadguy]

I have changed my EBJ-C into LRM boat, just because it's impossible to level it in SRM+UAC5 build - it's simply not viable. And I hate myself for doing it. I don't want to return to those terrible times, when every 'Mech should have been carrying LRMs in order to be viable. I have two Highlanders for example. One has LRMs, other - SRMs. First is viable, second - is not. And this problem is caused by MM problems. Players, who play more viable 'Mechs should have higher rating, so if I prefer to play less viable 'Mech - I should be playing against players, who have made this decision too - not against Meta players.

Edited by MrMadguy, 29 January 2016 - 12:38 AM.

[MrMadguy]

Lol. This is crap. Complete crap. Recently I suspected one thing, about how current PSR really works. This match confirmed this suspicion.

Edited by MrMadguy, 29 January 2016 - 02:07 AM.

[Ratpoison]

MrMadguy, on 29 January 2016 - 02:06 AM, said:

Lol. This is crap. Complete crap. Recently I suspected one thing, about how current PSR really works. This match confirmed this suspicion.
-snip-

I've done less damage than that while making plays that turned the tide of the game before, so you don't really have a point, as usual.

[MrMadguy]

tortuousGoddess, on 29 January 2016 - 02:19 AM, said:

I've done less damage than that while making plays that turned the tide of the game before, so you don't really have a point, as usual.

The question is - how low should your match score be in order for your PSR to drop in case of Win? How much PSR is biased towards winning?

Question to PGI: Have you licensed your MM from other company or it's details are hidden from us, just because you haven't done it and now you are trying to avoid suing?

[Ratpoison]

MrMadguy, on 29 January 2016 - 02:25 AM, said:

The question is - how low should your match score be in order for your PSR to drop in case of Win? How much PSR is biased towards winning?

Question to PGI: Have you licensed your MM from other company or it's details are hidden from us, just because you haven't done it and now you are trying to avoid suing?

No, that isn't the question. The question is how can the scoring for players be altered and improved to better reflect a player's contributions to the team, so that the system can better differentiate who is contributing to the team effort. Right now the system is designed to be more lax on taking away PSR, and that's because it's too inaccurate to be effective as a hard skill measurement. The best it can do is measure experience, and to a lesser extent, a player's ability to regularly contribute something to the team. Doing just this has already improved the matchmaking, but further improvements depend on improvements to what is scored, and only after that happens can we honestly consider drawing harsher lines for gaining and losing PSR.

And please don't break out into conspiracy theories like that, it's rude.