67 replies to this topic

[Darlith]

Ted Wayz, on 26 January 2016 - 06:01 PM, said:

The terms you use are relative and slanted. "You get punished in a loss if you did ok or bad". Compared to what? Right now it is compared to an arbitrary threshold that is not in relation to everyone's performance in a match. Does that remotely make sense that you should be punished if a team barely loses despite you being vital to the match being close, yet on the other side have someone rewarded where the team won despite how bad they played? Preposterous.

And these examples mostly are in relation to one mode. PSR does not adequately reward and harshly punishes assault and conquest if you play for a win via capping. Capping is legitimate in these modes yet you will have one side stay even and the other side go down if the match is played using a viable winning strategy.

PSR is plain rubbish.


"With no way to control it? You mean like by the mere act of queuing in solo queue and being at the mercy of matchmaker. Do you see the revelation in that statement? That being AFK puts you at the same mercy as playing? Are seeing how wins and losses are arbitrary?

No they are not arbitrary, the MM puts everyone on equal footing before skill, by giving you random teams each time. If you afk you cannot apply your skill to the match so you are at the mercy of the MM completely. If you play it is possible for your skill to pull out more wins than losses because you are good at making the team turn around and win more often than you lose. If you are losing more often than you win by a statistically relevant amount the trouble isn't the MM anymore it is you the player. If you are winning and losing at a fairly steady rate than you are an okay team player, not good, not great, but okay, and then if you go up or down in the long run is determined by your skill at fighting.

Actually scratch that, you might even be a good team player and win and lose at a similar rate.

Edit - lost part of my original post when I copied it over. I don't see the win/loss as a problem because if you are a good player your psr will rise, as it should. I don't think people should be catapulting to tier 1 by changing the system to "Only score matters, how your play helps your team doesn't." MWO asks a bit more of its players than the average FPS game, your rank isn't just how well you kill, it is how well you can help your team achieve victory and so wins and losses should be a key part of the PSR calculations.

Edited by Darlith, 26 January 2016 - 06:20 PM.

[Ted Wayz]

Darlith, on 26 January 2016 - 06:12 PM, said:

No they are not arbitrary, the MM puts everyone on equal footing before skill, by giving you random teams each time. If you afk you cannot apply your skill to the match so you are at the mercy of the MM completely. If you play it is possible for your skill to pull out more wins than losses because you are good at making the team turn around and win more often than you lose. If you are losing more often than you win by a statistically relevant amount the trouble isn't the MM anymore it is you the player. If you are winning and losing at a fairly steady rate than you are an okay team player, not good, not great, but okay, and then if you go up or down in the long run is determined by your skill at fighting.

Actually scratch that, you might even be a good team player and win and lose at a similar rate.

Edit - lost part of my original post when I copied it over. I don't see the win/loss as a problem because if you are a good player your psr will rise, as it should. I don't think people should be catapulting to tier 1 by changing the system to "Only score matters, how your play helps your team doesn't." MWO asks a bit more of its players than the average FPS game, your rank isn't just how well you kill, it is how well you can help your team achieve victory and so wins and losses should be a key part of the PSR calculations.

You are 'sposing.

You do not have to help your team achieve victory to advance, so scratch that. Playing good does not mean you will advance over time, that is determined by matchmaker and whether it has put you into a match there is a chance you can win in. Playing very good ensures you will stay even or advance, playing excellent determines if you will advance very little or a lot.

Matchmaker can put you on 12-2 win streaks and 12-2 loss streaks where it doesn't matter how well you play. Read the forums and you will a) understand this is true and learn how little people know about matchmaker.

But here is the kicker. Do you think players who rank on the leader board are bad players? I played well enough in an Orion IIc to be ranked but overall had a losing record. Many of my best matches were in losses. Did PSR care that I was skilled? Nope.

If you dig an old post with a system that I proposed based upon the leader board concept you will see a system based on skill.

Another note: You won't find many people who defend PSR on leaderboards. Just saying. Remember, carried in 12 man in a sub 100 damage Dire....

Edited by Ted Wayz, 26 January 2016 - 06:39 PM.

[Darlith]

Ted Wayz, on 26 January 2016 - 06:37 PM, said:

You are 'sposing.

You do not have to help your team achieve victory to advance, so scratch that. Playing good does not mean you will advance over time, that is determined by matchmaker and whether it has put you into a match there is a chance you can win in. Playing very good ensures you will stay even or advance, playing excellent determines if you will advance very little or a lot.

Matchmaker can put you on 12-2 win streaks and 12-2 loss streaks where it doesn't matter how well you play. Read the forums and you will a) understand this is true and learn how little people know about matchmaker.

But here is the kicker. Do you think players who rank on the leader board are bad players? I played well enough in an Orion IIc to be ranked but overall had a losing record. Many of my best matches were in losses. Did PSR care that I was skilled? Nope.

If you dig an old post with a system that I proposed based upon the leader board concept you will see a system based on skill.

Another note: You won't find many people who defend PSR on leaderboards. Just saying. Remember, carried in 12 man in a sub 100 damage Dire....

If you don't help your team win you will advance slower than the person who does, which is good. If anything is wrong it isn't win/loss affecting things it is the amount of gain and loss of psr, and the thresholds. Win and loss streaks do not matter, in the long run all things being equal those will even out, which is why PSR is meant to rise and fall slow, why it should be taking thousands of matches to move up or down a tier. That way statistical variances should be ironed out in favor of actual skill.

12 man carrying a sub 100 damage dire is likely to end with him not moving his psr rating, and if that is his consistant performance instead of a fluke match he isn't going anywhere fast regardless. If anything I would support the notion of a seperate PSR for group queue and solo queue, that way if you are carried in group queue you stay with your group and they suffer for carrying you, but if you go solo queue you rise and fall on your skill.

Edit - note I consider myself a mediocre pilot at best and I have ranked on the leaderboards in past events, I don't think you need to be good to have some good matches

Edited by Darlith, 26 January 2016 - 06:47 PM.

[Ted Wayz]

Darlith, on 26 January 2016 - 06:46 PM, said:

If you don't help your team win you will advance slower than the person who does, which is good. If anything is wrong it isn't win/loss affecting things it is the amount of gain and loss of psr, and the thresholds. Win and loss streaks do not matter, in the long run all things being equal those will even out, which is why PSR is meant to rise and fall slow, why it should be taking thousands of matches to move up or down a tier. That way statistical variances should be ironed out in favor of actual skill.

12 man carrying a sub 100 damage dire is likely to end with him not moving his psr rating, and if that is his consistant performance instead of a fluke match he isn't going anywhere fast regardless. If anything I would support the notion of a seperate PSR for group queue and solo queue, that way if you are carried in group queue you stay with your group and they suffer for carrying you, but if you go solo queue you rise and fall on your skill.

Again you don't get it.

I do not have to help my team at all to advance. I can shoot the wrong targets and pretty much spin in circles and still advance based upon either what my 12 man is doing to carry me or whether the matchmaker has me on a roll in solo queue.

You say "how fast" to try and change your argument to try and make it valid. But how fast is just as much determined by the same factors above, matchmaker or how good the players you are grouped with. Show me, with numbers, how statistical variances are ironed out when there is hard evidence that this is not so. Many people are successful despite PSR only because they are several sigmas better than the mean.

And you are admitting that a sub 100 Dire can continue to advance, at least that is progress. The fact that they do, no matter how fast or slow, is evidence that PSR has nothing to do with skill.

As said before, by using the leader board system you could rank everyone based upon chassis and skill, period. This way I would have the freedom of playing a mech I am bad at or new at or unskilled at without the threat that my "skill" in a chassis I am excellent at would be affected. What a concept! A system that promotes experimentation with mechs in a game whose primary source of income is the sale of mechs!

PSR is garbage and promotes bad behaviors for the sustainability of the current financial model. What is your opinion on that!

[Darlith]

Ted Wayz, on 26 January 2016 - 06:56 PM, said:

Again you don't get it.

I do not have to help my team at all to advance. I can shoot the wrong targets and pretty much spin in circles and still advance based upon either what my 12 man is doing to carry me or whether the matchmaker has me on a roll in solo queue.

You say "how fast" to try and change your argument to try and make it valid. But how fast is just as much determined by the same factors above, matchmaker or how good the players you are grouped with. Show me, with numbers, how statistical variances are ironed out when there is hard evidence that this is not so. Many people are successful despite PSR only because they are several sigmas better than the mean.

And you are admitting that a sub 100 Dire can continue to advance, at least that is progress. The fact that they do, no matter how fast or slow, is evidence that PSR has nothing to do with skill.

As said before, by using the leader board system you could rank everyone based upon chassis and skill, period. This way I would have the freedom of playing a mech I am bad at or new at or unskilled at without the threat that my "skill" in a chassis I am excellent at would be affected. What a concept! A system that promotes experimentation with mechs in a game whose primary source of income is the sale of mechs!

PSR is garbage and promotes bad behaviors for the sustainability of the current financial model. What is your opinion on that!

My opinion is that yours is wrong. I think the PSR system has flaws but taking wins and losses into account is not one of them.

If that mythical dire pilot is somehow being carried to tier 1 with his consistant sub-100 damage scores than it is evidence that group queue needs its own ranking where he can continue to drag his own team down or not without it affecting his psr in solo queue. If he is progressing with that kind of damage in the solo-queue than it is evidence that threshold for skill raises is off, not that wins are the problem, because if he sucks in the solo-queue consistantly than the skill drops on his losses should be dragging him down more than his gains from those sub-100damage performances in the wins.

If you shoot the wrong targets and spin in circles in solo queue, then yes a win streak might raise your psr, and then the next loss streak will drag it back down. If it isn't the thresholds and amounts of skill changes are the problem, not the consideration of wins and losses in the equation.

The leaderboards would be just as crap a way to judge skill, because they take the 10 best rounds out of however many rounds you choose to play and only check number of matches on a tie. Which means I can have 10 good matches out of 1000 and might rank higher than someone who has 10 good matches but only played 20 and is a more skilled player than I am, because I played more. Not really any different than the way everyone is claiming PSR currently works, the people who play the most will likely rise more.

As a small sample because I don't feel like coming up with tons of numbers we will rank these two mythical players here based on their best match.

Player 1 has scores of - 450, 240, 200, 130, 60, 255, 200, 180, 70, 145
Player 2 has scores of - 400, 380, 365, 390
The leaderboard would rank player 1 higher than player 2 because of 1 good match out of a bunch of mediocre to bad matches even though player 2 is arguably the better player. Just imagine instead of that small sample that player 1 had 10 good matches out of hundreds of mediocre to bad, and player 2 had 10 good scores out of say 20 good matches.

Now if you are suggesting we put a PSR ranking on each chassis/variant and take each and every match into account using some sort of score averaging system, I would be curious about it. But only if it takes the full average and not just a handful of best matches like the current leaderboards do. But I would also like the see what criteria you take into account for it, because if it goes purely off match score there are many flaws to that as well with the current scoring system. Even with that potential system I would like to see weighting to it for wins and losses.

[Mawai]

Lykaon, on 25 January 2016 - 06:02 PM, said:

ANSWER: PGI has tailored nearly all the support mechanics in the game to casual solos. The "average" MWO player apparently does not belong to a unit or play in groups or play a game mode other than skirmish for that matter.

Almost the entire game is being designed to spoon feed "casuals".

"Casuals" often refered to as pugs or puggies in MWo generally play in a rather narrow margin of variables.

They pilot generally the same chassis.ussually heavies.Having more mechs means more mech bays (mechbays cost money and why would they want to spend any on MWo)

They rarely paint their mechs (paint cost moneyand why would they want to spend any on MWo)

They will almost always select skirmish mode over anything else,Apparently secondary objectives are to hard or confusing or distract them from playing skirmish.

They are the ones who flood the forums with complaints about "premades" or "pugstomps" or any number of other topics that are actually about their lack of ability to cooperate and perform team work. Because they choose to play this game in a less efficent manner the players who do play effectivley have to be punished.

Essentially "casual solos" need to have the PSR system in place as a means of funneling them into generally similar pools of players. Since the PSR system doesn't really work well in the group queue it's clearly for them and not the units.

Why does it not work in groups you may ask?

PSR should keep players in a match with other players within 2 steps of their tier. Yet id a player group contains a tier 1 player and a tier 4 player the group will be placed in the same match obviously this supercedes the 2 step rule so PSR is not working as intended.

Lol .. wasn't sure whether to laugh or cry at all the misinformation and ludicrous opinion in this one.

1) Most MWO players are "casuals" because they have a life.

2) For a lot of the "hardcore" MWO fans, who started with Battletech, this includes jobs, wives, children and in some cases likely grand children.

3) The time available to a devote to a unit is minimal ... at best a lot of these "casuals" have a chance to hop on for a couple of games a night or maybe a couple a week. So they don't join a unit or drop with a group.

4) The particular group of "casuals" I have described have (a) jobs ( disposable income (c) the desire and ability to spend money on MWO. In fact, these "casuals" may well be a primary revenue source for PGI.

5) Heavies are popular because they give a good mix of firepower, armor and maneuverability ... they are fun to drive ... though I find lights to be a better adrenaline rush but I tend to play a bit risky with my lights

6) I bought about a dozen colours when they were on sale ... combined with the free colours it gives more than enough customization for me since I don't even look at my mechs during a match and could care less what others think of the appearance. It certainly has nothing to do with spending money.

7) Teamwork is OP ... as the original poster mentioned ... always has been. Luckily PUG stomps generally became a thing of the past everywhere except CW when they introduced the solo/group queue split. Having played sufficient matches where an 8 man pre-made rolled my PUG side ... coordination and team work are king. It was worth complaining about.

8) Huh. In the solo queue, everyone drops by themselves, there is no efficiency unless you speak up and organize your team ... and honestly, most teams respond well to a little organization. It doesn't punish folks who want to play more efficiently since they are in the group queue in the first place.

9) PSR works as well as it can in groups. The matchmaker averages the PSR in a group. If a Tier 1 CHOOSES to group with a Tier 4 ... then that is their choice ... they decided to group with someone from a lower tier whom they would not usually see in a match. It sounds like you'd prefer that the matchmaker boot those lower tier folks out of groups ... which I can guarantee would not go over well with group players. Anyway, so PSR gets averaged for each group and that value is used for matchmaking which is as good as it will get unless the matchmaker starts kicking folks out. Other than that, PSR should work fine in groups ... while Elo had problems. Why? The biggest component to match score is damage done and that contributes to PSR change. If a player contributes to a win without directly contributing that much then the PSR rise is modest. However, under Elo, all they needed was a win ... the change wasn't scaled by their performance. This resulted in folks that played well on a team but didn't necessarily do much getting inflated Elo values when playing in the solo queue. PSR does get around some of this issue in the group queue.

[MrMadguy]

Worst thing about bias towards winning - is that one good match compensate 20 stomps. So game just turns into non-stop stomp fest: 20 stomps - 1 good match - 20 stomps - 1 good match - 20 stomps... Just because those tiny bits of PSR, you lose while stomp, are being compensated by just one good match.

P.S. Game is completely unplayable at any time, except prime. I don't know, which Tiers lack players - lower or higher (I suspect higher), but mixing players from different tiers causes game to became unplayable for lower ones. PGI have to do something with this situation, otherwise they won't survive. Tier 3 is unplayable and I can't drop to Tier 4 to avoid being matched against Tier 1 - it's simply impossible.

Edited by MrMadguy, 27 January 2016 - 12:59 AM.

[Ghogiel]

Ted Wayz, on 26 January 2016 - 05:46 PM, said:

Your statement is based on what? Fact?

Given matchmaker is still highly influential on whether you have a win streak or loss streak when not playing group then you could easily stay even despite frequent AFKs. In most cases a single pilot has very little effect on the outcome. So why does PSR reward people for just showing up when the matchmaker has them on a roll?

Sorry, but many of you really do not understand PSR and matchmaker.

Yes. 1 mech down every game (only way to ensure this is to afk/suicide) causes an imbalance over time. Statistically this shows up as a lot more losses. I expect the effect to be exacerbated for T1 players.

Win streaks and loss streaks are statistically irrelevent. Being carried or sand bagged has no bias either way. Any player has no way of increasing the odds of either of those, except by afk/suiciding. Over a large data set they will even themselves out.

Edited by Ghogiel, 27 January 2016 - 10:43 AM.