Jump to content

Can We Talk About Group Queue?


377 replies to this topic

#341 EGG

    Member

  • PipPipPipPipPipPip
  • Bridesmaid
  • Bridesmaid
  • 322 posts

Posted 24 August 2015 - 01:58 AM

There was a thread on this back in March, I posted the below at that time and it's still 99% relevant. Swap Elo for PSR.

Quote

A lot of the pugs are cynical due the previous situation of 12+ months of groups farming pugs, and spewing vitriol in the forums about how "pugs shouldn't object to teams in a team game". If you don't bother to bring any concessions to the table, they will just threadnought your suggestions to Kaetetoa. I'd suggest things like an Elo increase or tonnage limit.


What you have here is a threadnought alright.

Posted Image

#342 Mawai

    Member

  • PipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 3,495 posts

Posted 24 August 2015 - 04:37 AM

View PostVlad Ward, on 22 August 2015 - 01:33 PM, said:


This is very true. Some people refuse to play as part of a team, but over time they'll just be naturally selected down to Tiers 4 and 5 where they won't see any but the absolute worst premade groups.

If people do play as part of a team, then it really doesn't matter if some number of them are on TS together or not.


Interesting thought but no.

Unfortunately, the scoring and PSR system rewards individual and not team play. Kills/assists/damage can be achieved as a solo cog in a loosely associated group. All you need is good aim and don't stand at the front so you die first (trying to die last to maximize damage is also a decent strategy.)

THIS is what the revised ranking system is based on. The folks who do this are more likely to rise to tier 1 by ignoring team work ... other than the use of other mechs as cover :) ... they are less likely to drop to tier 4 or 5 since they will still score pretty well even when they lose.

If the ranking was related to winning as opposed to match score then these players might drop in tiers since they will not earn points on losses ... so that the lack of team work would be balanced by the numbers they put up by being opportunistic. (ECM snipers would be in the same boat).

Anyway, PGI has the statistics and data to actually know what is going on ... assuming they actually bother to monitor it.

#343 Vlad Ward

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Merciless
  • The Merciless
  • 3,097 posts

Posted 24 August 2015 - 10:42 AM

View PostEGG, on 24 August 2015 - 01:58 AM, said:

There was a thread on this back in March, I posted the below at that time and it's still 99% relevant. Swap Elo for PSR.

What you have here is a threadnought alright.



There has been no shortage of possible concessions listed in this thread, from opting out to giving groups a PSR handicap based on their size. I can't help it if certain people ignore this to rail against the machine even longer.

View PostMawai, on 24 August 2015 - 04:37 AM, said:


Interesting thought but no.

Unfortunately, the scoring and PSR system rewards individual and not team play. Kills/assists/damage can be achieved as a solo cog in a loosely associated group. All you need is good aim and don't stand at the front so you die first (trying to die last to maximize damage is also a decent strategy.)

THIS is what the revised ranking system is based on. The folks who do this are more likely to rise to tier 1 by ignoring team work ... other than the use of other mechs as cover :) ... they are less likely to drop to tier 4 or 5 since they will still score pretty well even when they lose.

If the ranking was related to winning as opposed to match score then these players might drop in tiers since they will not earn points on losses ... so that the lack of team work would be balanced by the numbers they put up by being opportunistic. (ECM snipers would be in the same boat).

Anyway, PGI has the statistics and data to actually know what is going on ... assuming they actually bother to monitor it.


All players will reach a point where sitting in the back doing nothing productive will no longer allow their teams to win. Fighting 10v12 is an impossible game to win for the 10 unless they're all significantly stronger pilots, which shouldn't be happening much once PSR settles.

Any pilot who sits in the back, does "okay", but loses the match isn't going to rise very quickly. You're better off taking a win and getting a high match score.

Tiers are not absolute. People who don't rise quickly aren't going to rise at all. PSR, from what I can gather, is an absolute integer value number, while tiers are simply a percentage of the population. If you're hanging back and earning 5 PSR points a game, you'll never reach a higher tier when other players are earning 20 or 30 PSR points a game. If anything, you may drop down even if you never do poorly simply because the rest of the population is gaining points faster than you are.

Edited by Vlad Ward, 24 August 2015 - 11:22 AM.


#344 Aresye

    Member

  • PipPipPipPipPipPipPipPipPip
  • Heavy Lifter
  • Heavy Lifter
  • 3,462 posts

Posted 24 August 2015 - 11:21 AM

View PostMawai, on 24 August 2015 - 04:37 AM, said:

Unfortunately, the scoring and PSR system rewards individual and not team play. Kills/assists/damage can be achieved as a solo cog in a loosely associated group. All you need is good aim and don't stand at the front so you die first (trying to die last to maximize damage is also a decent strategy.)


You're half right. Individual performance does have more of an impact on a player's PSR vs the primarily Win/Loss focus of Elo. Where you're wrong is how that performance is applied (ie - the end goal).

The entire purpose for the PSR MM is to create matches with players who are similar in playing skill, because a team is more coordinated when everybody is around the same level. The purpose the Elo MM had on the other hand, was to simply balance the average of both teams, so it would pull really high players and really low players to balance out the average against the other team who were mostly players in the middle.

So while the individual performance factors behind PSR seems contradictory to a team-oriented game, having matches where every player is around the same skill level leads to more balanced games, whereas before many players found themselves unfairly being "taken up" in Elo when their team won, despite poor performance.

#345 pbiggz

    Member

  • PipPipPipPipPipPipPipPipPip
  • Overlord
  • Overlord
  • 4,827 posts
  • LocationOutreach

Posted 24 August 2015 - 03:56 PM

The only reason you see anyone further than two tiers outside your own in group queue is because the population is so minuscule that (nearly) all the release valves open. If we had one queue, then more often than not, none of the release valves would open. You would rarely see people outside your tier, and if you think that premades (even if they're the same tier as you) are still OP, once again, you have VOIP and lfg to communicate with your team or make your own premades.

The premade boogieman narrative is a construct. It's fiction.

#346 Vlad Ward

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Merciless
  • The Merciless
  • 3,097 posts

Posted 24 August 2015 - 06:46 PM

I thought he was being perfectly reasonable, and his representation of how the matchmaker works is accurate. What part of that first post was "playing the fool"?

I suppose he should have come back with worn out 90's cliches, too.

Also welcome back to the thread, again.

Edited by Vlad Ward, 24 August 2015 - 06:47 PM.


#347 Dimento Graven

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Guillotine
  • Guillotine
  • 6,208 posts

Posted 24 August 2015 - 07:18 PM

View PostVlad Ward, on 21 August 2015 - 10:16 PM, said:

I feel that the group-only queue has outlived its usefulness in 3052 Mechwarrior Online.

When the segregation between Group and Solo queues was originally implemented, MWO was running with a version 1.0 matchmaker without all the bells and whistles we have today. New players were running up against groups of 8-12 players in coordinated builds on teamspeak together and getting facerolled. No one wanted to keep playing when they knew they could drop into a match only to see 12 AS7-D-DCs rolling up over the ledge in Frozen City.

There is no longer any real danger of this happening today.

With the implementation of 3/3/3/3, it is no longer possible for large groups to strongarm their way through solo players and smaller groups by throwing 12 Assaults or 12 Lights at them. Even if one team contains a 12-man and the other a smattering of smaller groups/solos, neither team will be vastly outweighed by the other.

Likewise, the successful implementation of in-game VOIP has negated the tactical advantage of groups being able to communicate via Teamspeak. Now, anyone can call targets or spot enemies without being forced to type while dodging autocannon rounds.

Most importantly, the implementation of PSR will help dramatically with ensuring that teams consist of players of roughly equivalent skill levels. This means that brand new players won't be running up against large groups unless those groups, themselves, consist of players with approximately Tier 4 ratings. On the other side of the coin, it also means that Tier 1-2 players who drop solo can be used to fill in gaps in Tier 1-2 teams which lack a full 12.

Removing the restriction on group/solo queues is the best way to ensure that the matchmaker has the largest possible player pool to draw from, resulting in the best possible matches for everyone.
I'd change this to say that the 'solo' queue has outlived its usefulness.

Just make 11 man grouping viable and voilà (spelled correctly for all you spelling facists) you can still have units pre make their own groups and put them against appropriately PSR'd mixes of solo's and smaller groups.

I'd be ok with that.

#348 Kjudoon

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Rage
  • Rage
  • 7,636 posts
  • LocationWisconsin

Posted 24 August 2015 - 07:57 PM

View Postpbiggz, on 24 August 2015 - 07:14 PM, said:


dude, you can't come into this thread and claim to be arguing a real point while apparently putting everyone who disagrees with you on your ignore list. That's psychotic.

Go back and read the thread then. I spent a long time trying to bring sense to the nonsensical. At some point you just cut the line and let it go because whatever you caught isn't worth hauling in.

I've been involved in this conversation since page 1 pretty much. I know what it's about because I just didn't read the first post and jump to the end without context.

View PostDimento Graven, on 24 August 2015 - 07:18 PM, said:

I'd change this to say that the 'solo' queue has outlived its usefulness.

Just make 11 man grouping viable and voilà (spelled correctly for all you spelling facists) you can still have units pre make their own groups and put them against appropriately PSR'd mixes of solo's and smaller groups.

I'd be ok with that.

There are 2 conditions I'd accept with that. You can only group with people of the same PSR since group averaging is the real sin of the group queue. Not the same PSR, you can't group. Sorry.

Second, lock the queues, so only the same tier can play together.

Edited by Kjudoon, 24 August 2015 - 07:58 PM.


#349 Dimento Graven

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Guillotine
  • Guillotine
  • 6,208 posts

Posted 24 August 2015 - 08:03 PM

View PostKjudoon, on 24 August 2015 - 07:57 PM, said:

...

There are 2 conditions I'd accept with that. You can only group with people of the same PSR since group averaging is the real sin of the group queue. Not the same PSR, you can't group. Sorry.

Second, lock the queues, so only the same tier can play together.
Well I don't necessarily agree with the first condition. Averaging each player's performance should, theoretically, only raise or lower group's potential drop tier by one. Sure, there could be one or more players that drop in a much lower or much higher tier than normal. The later, not so much a big deal, other than that guy will probably die first, the former... I could see that being an issue, but HOPEFULLY rare...

The second condition COULD be met with MORE population, but I think we'd need to increase the MWO population by around 10 times (I dunno maybe only 5 times) to keep wait times for each queue 'reasonable' (reasonable = no more longer than they are now).

#350 Kjudoon

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Rage
  • Rage
  • 7,636 posts
  • LocationWisconsin

Posted 24 August 2015 - 08:32 PM

View PostDimento Graven, on 24 August 2015 - 08:03 PM, said:

Well I don't necessarily agree with the first condition. Averaging each player's performance should, theoretically, only raise or lower group's potential drop tier by one. Sure, there could be one or more players that drop in a much lower or much higher tier than normal. The later, not so much a big deal, other than that guy will probably die first, the former... I could see that being an issue, but HOPEFULLY rare...

The second condition COULD be met with MORE population, but I think we'd need to increase the MWO population by around 10 times (I dunno maybe only 5 times) to keep wait times for each queue 'reasonable' (reasonable = no more longer than they are now).

And yet, we know that PSR averaging, formerly Elo averaging was a very very bad thing. Even PGI admitted it when matches were having average variations as high as 200 points (50 points was considered fairly extreme according to them). This information was gathered from screenshots of matchmaker logs by Bryan Ekman a few months back when this issue came about and it was the Solo queue that was the problem.

All problems are solved with a bigger population, but since that will not happen without steam being successful, AND have a good way to incubate the new players instead of throwing them to the wolves like the 1 queue solution will. If you don't protect them out of the gate, you WILL lose them and the game will die quickly.

#351 Dimento Graven

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Guillotine
  • Guillotine
  • 6,208 posts

Posted 24 August 2015 - 08:43 PM

View PostKjudoon, on 24 August 2015 - 08:32 PM, said:

And yet, we know that PSR averaging, formerly Elo averaging was a very very bad thing. Even PGI admitted it when matches were having average variations as high as 200 points (50 points was considered fairly extreme according to them). This information was gathered from screenshots of matchmaker logs by Bryan Ekman a few months back when this issue came about and it was the Solo queue that was the problem.
I think that was specific to the solo queue as you say, and considering that elo was measuring a statistic that ultimately, in this game, has little to do with actual pilot skill and contribution to the team's efforts, kind of seems like what happened while it was being employed is irrelevant and has no bearing on PSR's ultimate affect.

I totally admit it's a warning and something to watch out for, but I don't think that in the pre-made group queue, it'll be that significant an issue since it's 'skill' not 'wins verses losses verses elo's prediction'.

My recent drops makes me believe that either solo or grouped I am, for the most part, being grouped with individuals of not to dissimilar skills.

Is my experience the norm, or not, I can't tell you. The forums represent only the loudest minority of the player base, so we may never know the 'real' answer to this particular question.

Quote

All problems are solved with a bigger population, but since that will not happen without steam being successful, AND have a good way to incubate the new players instead of throwing them to the wolves like the 1 queue solution will. If you don't protect them out of the gate, you WILL lose them and the game will die quickly.
True, which is why I hope that keeping Tier 5 out of the reach of the top 2 tiers, and Tier 4 out of the reach of the top tier will provide as much of that protection that is as reasonably possible.

#352 Vlad Ward

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Merciless
  • The Merciless
  • 3,097 posts

Posted 24 August 2015 - 09:12 PM

The matchmaker is already locked to same tier only for entities (solo players or group averages) unless you've waited for more than 60 seconds (solo queue) or 30 seconds (group queue).

It used to be a 60 second lock across the board, but PGI decided to scrap the wait time for group queue since it wasn't making same-tier games anyways and 30 seconds off the wait is 30 seconds off the wait.

Locking groups to same-tier only would kill units and pretty much get rid of referrals. A straight average may not be the ideal case, but it could be worse. There are a million different small modifications that can be made to the matchmaker to help even those things out - like weighted averages for groups and PSR handicaps.

#353 Kjudoon

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Rage
  • Rage
  • 7,636 posts
  • LocationWisconsin

Posted 24 August 2015 - 09:27 PM

@ Dimento

For now, there are 2 critical flaws to the ranking system. I like PSR and what it's going to score on even though it is still too heavily weighted to damage for my tastes.

1. It is too easy to go up in PSR. There is like only one condition in which your psr goes down and that's doing very badly on a loss. There is no way for it to go down on a victory. The Loss scale should be the inverse of a Win scale. That way it does not have a natural buoyancy that will ultimately create the same Elo hells we are used to of people bouncing off the ceiling like flies trapped in a jar and ultimately quitting the game in frustration because their losses do not take them back down fast enough to where they are balanced. This is also known in business as "The Peter Principle". Promoting someone beyond their capacity to do well and leave them stuck at a job they are incompetent at.

2. With low population, you cannot lock the tiers. I acknowledge that fact. The problem is you are not going to grow the population when you have seal clubbing going on. New players who have a bad experience will not play the game. The same way a kid who joins a baseball team and sucks will quickly quit and usually hate the sport from then on out. This is not new. Sure, you're not seeing Tier 1 when you start, because you are seeded at Tier 4. Okay, we get this. The problem will be in how fast you rise (which looks to be quite fast really) till you're tier 3 and facing those tier 1 thugs AND being teamed with the same people derping up the joint. Just not all at the same match. This will lead to an unsatisfactory gaming experience and drive the player away as quickly as a kid who sucks at baseball is teased and ridiculed for his bad playing and inability to hit the ball. Kids are cruel, and it seems the internet doesn't help em grow up.

Tiers ultimately need to be locked once the population grows enough. But without allowing for fast matches, nobody will play so we're in a vicious circle. The compromise is not piling everyone together. You can loosen the tiers to maybe the neighboring tier only, but any more than that, you're better off saying "Match not found, Search again?" than to inflict a horrifying game experience. It's the lesser of two evils. Yes the "neighbors" level of Tier matching tells Tier 5 and 1 to suck it, but really, I don't have a problem with ignoring these edge cases. The tier 1 players are probably in organized groups playing in CW and tournaments anyway and getting their match on that way. They don't have many peers left because they beat them all. Tier 5s are either declining to the point of being out the door (since your psr only declines under doing extremely bad with losses) and they make for easy victims for the Tier 4 new guys and just blah casuals. Nobody likes being the omega food stockand Tier 5s are it through their own effort or lack their of. Otherwise, they'd improve as it's currently weighted. Right now it takes effort to be bad. Lots of it. Even when new.


To be honest, I think the bigger problem is going to be the Tier 2 tryhards. They'll get upset their not Tier 1, and take it out on everyone else... except tier 1s who will beat them pretty consistently when matched. When matched with Tier 3's they're going to beat the tar out of them pretty bad and be the worst sports about it since their power has been thwarted. It's a big reason why locking the tiers is best. Right now, Tier 2s can also face anyone. Just not Tier 4-5 and Tier 1s mixed. They will usually crow their easy victories and kvetch endlessly when beaten by the betters, and have an apoplectic fit if they lose to the lower tiers and fire off endless accusations of cheaters. Knowing my luck I'm probably a Tier 2. Gawd, that'd be horrifying!

So you can see why I am the originator of the "Lock the Tiers" movement, and all for protecting the mechanisms, that if we get the population will incubate and grow this game. It's horrid while we're so low, but PGI only has to look in the mirror for much of that, and look at the forums to see the results of their well... less than stellar choices. Games get the gaming populace they deserve.

And yes, the forums represent a small, vocal minority of the game population.

Edited by Kjudoon, 24 August 2015 - 09:28 PM.


#354 Vlad Ward

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Merciless
  • The Merciless
  • 3,097 posts

Posted 24 August 2015 - 09:58 PM

I don't even know if he can read this, but there are several issues with the assumptions presented above.

1) PSR values are integer values with no known limit. Tiers are simply delineations defined by certain percentages of the population. For example, let's say Tier 1 is the 90th percentile. This could be 500 people or 500,000, depending on the total population of MWO. Their Pilot Skill Scores could also be anywhere from 600 points to 573,000,000 points, as long as they're the top 10 percent of the population.

The important takeaway from this is that your PSR doesn't exist in a vacuum. It's very possible to go up or down a tier without ever playing a match, so long as enough other people either surpass you or fall below you.

This means 2 things:

1) Probably the single most important factor in determining your overall ranking is number of games played. People who turn the game off for months at a time will come back to find their Tier 3 account has slipped into Tier 4 as other players have surpassed their Pilot Skill Score. This is good, because no one's going to retain all of their mecha fighting ability after a 6 month break. Likewise, it prevents all but the most tryhard of tryhards from taking a Tier 1 spot and holding on to it indefinitely without continually expending effort. There is no gold star for being good at the game at some arbitrary point in the past. Trust me, I'd know.

2) Consistently winning games is not enough to actually cause you to go up a tier. In order to move up the rankings, you have to not only win, but win more often and more thoroughly than the rest of the population of your tier.

Tiers are big places. They're huge swathes of a huge population. We're talking tens of thousands of players. Just because someone wins a game doesn't mean there aren't hundreds of other players also winning games at the exact same time. This will prevent all but the strongest of players in a tier from making the move into a higher bracket - exactly the way it should be.

Edited by Vlad Ward, 24 August 2015 - 09:58 PM.


#355 Dimento Graven

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Guillotine
  • Guillotine
  • 6,208 posts

Posted 24 August 2015 - 10:05 PM

View PostKjudoon, on 24 August 2015 - 09:27 PM, said:

...

For now, there are 2 critical flaws to the ranking system. I like PSR and what it's going to score on even though it is still too heavily weighted to damage for my tastes.

1. It is too easy to go up in PSR. There is like only one condition in which your psr goes down and that's doing very badly on a loss. There is no way for it to go down on a victory. The Loss scale should be the inverse of a Win scale. That way it does not have a natural buoyancy that will ultimately create the same Elo hells we are used to of people bouncing off the ceiling like flies trapped in a jar and ultimately quitting the game in frustration because their losses do not take them back down fast enough to where they are balanced. This is also known in business as "The Peter Principle". Promoting someone beyond their capacity to do well and leave them stuck at a job they are incompetent at.
Well I disagree that there should be any reason your PSR goes down when your side wins. There's just too many circumstances in a WELL performing team where 'doing well' may not be possible because all the enemies are dead by the time you get catch up to everyone else.

My understanding is that it won't go up if you perform badly yet still win. This SHOULD prevent most people from rising to tiers they don't actually qualify for, OR AT THE LEAST, not staying in that Tier for very long.

With PSR you have to have some pretty strong downward inertia before you move downward, that's a lot of consistent losses doing very badly score-wise. This keeps you from being overly punished for getting matched with a lot of bad/not serious players on your side.

Quote

2. With low population, you cannot lock the tiers. I acknowledge that fact. The problem is you are not going to grow the population when you have seal clubbing going on. New players who have a bad experience will not play the game. The same way a kid who joins a baseball team and sucks will quickly quit and usually hate the sport from then on out. This is not new. Sure, you're not seeing Tier 1 when you start, because you are seeded at Tier 4. Okay, we get this. The problem will be in how fast you rise (which looks to be quite fast really) till you're tier 3 and facing those tier 1 thugs AND being teamed with the same people derping up the joint. Just not all at the same match. This will lead to an unsatisfactory gaming experience and drive the player away as quickly as a kid who sucks at baseball is teased and ridiculed for his bad playing and inability to hit the ball. Kids are cruel, and it seems the internet doesn't help em grow up.

Tiers ultimately need to be locked once the population grows enough. But without allowing for fast matches, nobody will play so we're in a vicious circle. The compromise is not piling everyone together. You can loosen the tiers to maybe the neighboring tier only, but any more than that, you're better off saying "Match not found, Search again?" than to inflict a horrifying game experience. It's the lesser of two evils. Yes the "neighbors" level of Tier matching tells Tier 5 and 1 to suck it, but really, I don't have a problem with ignoring these edge cases. The tier 1 players are probably in organized groups playing in CW and tournaments anyway and getting their match on that way. They don't have many peers left because they beat them all. Tier 5s are either declining to the point of being out the door (since your psr only declines under doing extremely bad with losses) and they make for easy victims for the Tier 4 new guys and just blah casuals. Nobody likes being the omega food stockand Tier 5s are it through their own effort or lack their of. Otherwise, they'd improve as it's currently weighted. Right now it takes effort to be bad. Lots of it. Even when new.


To be honest, I think the bigger problem is going to be the Tier 2 tryhards. They'll get upset their not Tier 1, and take it out on everyone else... except tier 1s who will beat them pretty consistently when matched. When matched with Tier 3's they're going to beat the tar out of them pretty bad and be the worst sports about it since their power has been thwarted. It's a big reason why locking the tiers is best. Right now, Tier 2s can also face anyone. Just not Tier 4-5 and Tier 1s mixed. They will usually crow their easy victories and kvetch endlessly when beaten by the betters, and have an apoplectic fit if they lose to the lower tiers and fire off endless accusations of cheaters. Knowing my luck I'm probably a Tier 2. Gawd, that'd be horrifying!

So you can see why I am the originator of the "Lock the Tiers" movement, and all for protecting the mechanisms, that if we get the population will incubate and grow this game. It's horrid while we're so low, but PGI only has to look in the mirror for much of that, and look at the forums to see the results of their well... less than stellar choices. Games get the gaming populace they deserve.

And yes, the forums represent a small, vocal minority of the game population.
Your concerns, while noted, I don't think are all that critical.

First we really won't know how 'quickly' an average ACTUAL new player will move up or down tiers, we don't know and PGI apparently won't tell us, how they're actually generating the PSR scores. We know the components of the scores, and have a general idea on how those components are weighted, BUT, we don't know what's going on in the background. We really don't know for sure where the "+points to PSR score" and "-points to PSR score" triggers are. All we know is that it's based off performance, a winning match with 'good' score means a bump. A winning match with a non-good score means little if any change. A losing match with a 'great' score means a bump up, though again, the criteria for a 'great score' we don't know, and a losing match with a bad score means a bump down.

The scoring seems to me to indicate that the rise between tiers is designed to be very gradual, even if you are winning 100% of your matches. So seeing new players rocket to tier 3 or 2 seems, unrealistic, though again, I admit I could be wrong, as pointed out earlier, we really don't know the algorithms behind the scene and PGI won't tell us (to prevent gaming the system obviously). Though if you think about it, it's probably why all NEW players are put in the upper portion of Tier 4 by default instead of Tier 5. We're going to get new players that once they learn the basics are going to be THAT good. There's a lot of MMO FPS's out there to have built up the ability to aim, understand tactics and strategy and how to cooperate with other people on your team.

'Quickly' moving to the next tier up may be a matter of perspective, but it WILL give them more opportunity to lose bad, BUT learn the game, before they subjected to being stuck in Tier 5 and having to work their way all the way back up again.

If we HAD the population necessary for locking like tier to like tier or even the single tier up/down spread as you mentioned, hey, that'd be awesome, and I agree we'd probably want to do THAT first, but EVEN THEN, we'd still want to include CW as part of the PSR growth or shrinkage.

The problem of generating too many 'Match not found messages' is that, that too is a bad player experience, wait 5 minutes to try and get a match only to be told to 'try again' isn't the sort of thing you want to be doing to existing or new players. At least by locking the elite vets to the first three tiers, THEORETICALLY, they're committed to playing the game anyway (and have long since gotten used to long MM searches anyway, even the bald old elo hell days, you could wait 5 minutes to get a match that ended in 2 minutes).

I'll agree PGI and (more originally IGP) made some pretty bad gaming decisions when it came to the development and progression of this game. It has gotten better and certainly the faux pas of late are not some the more cataclysmic variety that they've had in the past.

Again on this one I think we're going to end up seeing a big win for us and PGI in the long run.


View PostVlad Ward, on 24 August 2015 - 09:58 PM, said:

I don't even know if he can read this, but there are several issues with the assumptions presented above.

1) PSR values are integer values with no known limit. Tiers are simply delineations defined by certain percentages of the population. For example, let's say Tier 1 is the 90th percentile. This could be 500 people or 500,000, depending on the total population of MWO. Their Pilot Skill Scores could also be anywhere from 600 points to 573,000,000 points, as long as they're the top 10 percent of the population.

The important takeaway from this is that your PSR doesn't exist in a vacuum. It's very possible to go up or down a tier without ever playing a match, so long as enough other people either surpass you or fall below you.

This means 2 things:

1) Probably the single most important factor in determining your overall ranking is number of games played. People who turn the game off for months at a time will come back to find their Tier 3 account has slipped into Tier 4 as other players have surpassed their Pilot Skill Score. This is good, because no one's going to retain all of their mecha fighting ability after a 6 month break. Likewise, it prevents all but the most tryhard of tryhards from taking a Tier 1 spot and holding on to it indefinitely without continually expending effort. There is no gold star for being good at the game at some arbitrary point in the past. Trust me, I'd know.

2) Consistently winning games is not enough to actually cause you to go up a tier. In order to move up the rankings, you have to not only win, but win more often and more thoroughly than the rest of the population of your tier.

Tiers are big places. They're huge swathes of a huge population. We're talking tens of thousands of players. Just because someone wins a game doesn't mean there aren't hundreds of other players also winning games at the exact same time. This will prevent all but the strongest of players in a tier from making the move into a higher bracket - exactly the way it should be.
I didn't interpret what PGI had disclosed on this as meaning what you've posted, HOWEVER, I am willing to acquiesce due to an EXTREME ignorance of how they're doing it.

#356 Vlad Ward

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Merciless
  • The Merciless
  • 3,097 posts

Posted 24 August 2015 - 10:18 PM

View PostDimento Graven, on 24 August 2015 - 10:05 PM, said:

I didn't interpret what PGI had disclosed on this as meaning what you've posted, HOWEVER, I am willing to acquiesce due to an EXTREME ignorance of how they're doing it.


I'll readily admit I'm neither a PGI employee nor a secret squirrel so most of this is conjecture on my end as well, but it fits with all of the official statements from Russ and Paul thus far and seems to me to make the most sense as a replacement to ELO.

I'll gladly take an official correction if I'm wrong, but there would be serious issues if everyone just kept rising in rank until everyone was Tier 1, no?

#357 Dimento Graven

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Guillotine
  • Guillotine
  • 6,208 posts

Posted 24 August 2015 - 10:38 PM

View PostVlad Ward, on 24 August 2015 - 10:18 PM, said:

I'll readily admit I'm neither a PGI employee nor a secret squirrel so most of this is conjecture on my end as well, but it fits with all of the official statements from Russ and Paul thus far and seems to me to make the most sense as a replacement to ELO.

I'll gladly take an official correction if I'm wrong, but there would be serious issues if everyone just kept rising in rank until everyone was Tier 1, no?
Mmm... not in the way that I think you're inferring, BUT, we're back to having individual 'mech PSR's if we take it what in my currently insomniac mind is a logical conclusion, namely if there's only ONE PSR per PLAYER, what happens to my game play when I move from my favored dual gauss KGC, to a brand new "neato-whiz-bang" 'mech that they've just released and I haven't optimized yet?

Well, it's obvious: My play goes to crap (in comparison) and I do badly in a lot of matches until I get used to the 'mech, modify it to my play style, skill it up to a minimum of elited.

But again, I totally agree that each 'mech should have its own PSR tracked, just like each has its own KDR and WLR tracked.

#358 Kjudoon

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Rage
  • Rage
  • 7,636 posts
  • LocationWisconsin

Posted 24 August 2015 - 11:04 PM

Balance in matches should run through 2 numbers.
PSR multiplied by your mech's Battle Value as described in original TT, a score based on loadout.

Good pilots in bad mechs balance with bad pilots with good mechs better. Screw tonnage. A firestarter is not equal to a Panther, nor is an Atlas equal to a Dire Whale.

#359 Vlad Ward

    Member

  • PipPipPipPipPipPipPipPipPip
  • The Merciless
  • The Merciless
  • 3,097 posts

Posted 24 August 2015 - 11:06 PM

View PostDimento Graven, on 24 August 2015 - 10:38 PM, said:

Mmm... not in the way that I think you're inferring, BUT, we're back to having individual 'mech PSR's if we take it what in my currently insomniac mind is a logical conclusion, namely if there's only ONE PSR per PLAYER, what happens to my game play when I move from my favored dual gauss KGC, to a brand new "neato-whiz-bang" 'mech that they've just released and I haven't optimized yet?

Well, it's obvious: My play goes to crap (in comparison) and I do badly in a lot of matches until I get used to the 'mech, modify it to my play style, skill it up to a minimum of elited.

But again, I totally agree that each 'mech should have its own PSR tracked, just like each has its own KDR and WLR tracked.


Yeah. I meant that it makes the most sense as a glimpse of the whole system given the information we have so far, lol.

Russ has mentioned that there's only one PSR per player, unfortunately.

#360 Dimento Graven

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Guillotine
  • Guillotine
  • 6,208 posts

Posted 24 August 2015 - 11:10 PM

View PostVlad Ward, on 24 August 2015 - 11:06 PM, said:

...

Russ has mentioned that there's only one PSR per player, unfortunately.
Only proper response to that is 'Meh.'





1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users