10 replies to this topic

[LordNothing]

so right now your psr score is relative to the other players in the current match. so if you happen to be the king potato, you can still move up if everyone else is starchier than thou even if you yourself did terrible. or likewise when faced with a match full of pros, its a good chance you are going to lose psr just because of present company regardless of how well you do in match. you will simply tend to do not so well when paired with higher ranked players and get downranked. incidentally ive had both those matches within 2 hours of each other.

now this is not so much a complaint as wanting to discuss the alternative. what if psr is based off of the global community average instead of the local match average? would than not place players in a better position than the former? seems now people on the edge of a tier will ping-pong back and fourth (ive done this many times since the psr redux). the match maker can easily put you in one of the edge slots and the selection of other players in the match has more influence than your actual performance.

perhaps it so pgi can give you a red or green arrow to show progress or its inverse rather than a meh yellow = sign. perhaps humans have some kind of psychological need to be approved of or disapproved of and that being neutral is somehow worse than being inferior, and this somehow affects the bottom line. thus a system that puts you where you belong and keeps you there is worse than a system that moves you around a lot. or perhaps the alternative would lead to a better match quality. discuss.

Edited by LordNothing, 07 June 2021 - 01:23 AM.

[Gagis]

Global stats depends heavily on your opposition.

A tier 3 player may have equal or superior statistics compared to a tier 1 player simply because their opposition (temporarily) puts up less of a fight. This gets a tad complicated during times players are moving up or down.

[LordNothing]

so a third option of 'maybe it doesnt matter' exists and that however its going to be configured the complexity of mm is just going to eat it alive, so to speak.

[crazytimes]

No. Worse players in matches of peers get much higher match scores. Some of the tier 5 match cards that get posted can have median scores in the 3-400 range as players sit around for 15 minutes farming spread damage and AMS off each other.

Tier 1 matches at extreme can have median scores as low as the <200 range as people will tend towards PFLD builds and accurate fire, with less lurms and effective AMS boating. Takes 200 accurate damage to core the biggest assaults.

Compare apples with apples.

[il1il]

Winning and losing would be clear indicators of performance.

[Davegt27]

LordNothing, on 07 June 2021 - 01:20 AM, said:

so right now your psr score is relative to the other players in the current match. so if you happen to be the king potato, you can still move up if everyone else is starchier than thou even if you yourself did terrible. or likewise when faced with a match full of pros, its a good chance you are going to lose psr just because of present company regardless of how well you do in match. you will simply tend to do not so well when paired with higher ranked players and get downranked. incidentally ive had both those matches within 2 hours of each other.

now this is not so much a complaint as wanting to discuss the alternative. what if psr is based off of the global community average instead of the local match average? would than not place players in a better position than the former? seems now people on the edge of a tier will ping-pong back and fourth (ive done this many times since the psr redux). the match maker can easily put you in one of the edge slots and the selection of other players in the match has more influence than your actual performance.

perhaps it so pgi can give you a red or green arrow to show progress or its inverse rather than a meh yellow = sign. perhaps humans have some kind of psychological need to be approved of or disapproved of and that being neutral is somehow worse than being inferior, and this somehow affects the bottom line. thus a system that puts you where you belong and keeps you there is worse than a system that moves you around a lot. or perhaps the alternative would lead to a better match quality. discuss.

discussion would be good

but any honest discussion will either be moderated or shouted down by the mob

Edited by Davegt27, 07 June 2021 - 10:36 AM.

[Khobai]

all i know is its pretty depressing when you do 500 damage in a light mech and still get a =

i definitely think it should be based on the global average instead of the local average

[TheCaptainJZ]

Over enough time, the local average should come out to being similar to the global average. At least, it would even out to the global average out of all the games you have experienced, which seems like a reasonable way to do it.
There is tweaking that can be done to change how it's calculated and what actions are rewarded and by how much. AMS is over-represented in MS for instance and that needs toned down. Once we've rebalanced Match Score, then hopefully it would feel more accurate for more pilots and not be as biased around a mech class. This part is the complicated part but also where the true meat and potatoes are. Local or Global average isn't meaningful if the calculation of those averages is still skewed towards certain actions.

[Khobai]

TheCaptainJZ, on 07 June 2021 - 01:00 PM, said:

Over enough time, the local average should come out to being similar to the global average. At least, it would even out to the global average out of all the games you have experienced, which seems like a reasonable way to do it.
There is tweaking that can be done to change how it's calculated and what actions are rewarded and by how much. AMS is over-represented in MS for instance and that needs toned down. Once we've rebalanced Match Score, then hopefully it would feel more accurate for more pilots and not be as biased around a mech class. This part is the complicated part but also where the true meat and potatoes are. Local or Global average isn't meaningful if the calculation of those averages is still skewed towards certain actions.

thats not the case because sometimes you get games with tons of assault mechs

[Leone]

Okay. First we gotta make sure we're all on the same page.

As it stands a hefty portion of matchscore is based offa damage. Which incentivizes avoiding lights and picking on heavies and assaults. Which in turn incentivize nascaring to avoid flanking lights and get free shots on distracted tailing assualts.

Now, when switching to this, 'global matchscore' You'll find more fast movers like mediums and lights being piloted by better players in order to incentivize others to go elsewhere for the damage and leading the nascar, leaving higher tier matches with less damage to spread around, thus lowering matchscores. Then in low tier matches you'll have newbs dropping heavy and picking eachother apart with higher damage to go around and so much spread damage with missiles and poor aim and huge matchscores.

Then they'll all meet in tier 3 and the salt will flow.

~Leone.

Edited by Leone, 07 June 2021 - 09:42 PM.

[il1il]

So a big problem with the idea of a standard global average is that your not adjusting the psr globally in a game you are taking psr from 24 players and redistributing it among them. that means you need to localize the standards by which you judge the players. in a game with less tonnage there will be less damage or more tonnage more damage. that means the damage that happens in a game has no objective standard.
Now if you follow this logic this also means that damage to a light is not appropriately weighed locally when compared to other local damage (on higher tonnage).
IF you were to say damage was a worthy measure, by the logic of the match score system it should be assessed as a percentage of the components max health in the weight in match score.
Also taking higher tonnage your self should perhaps reduce the weight of damage as it is easier to score high damage on heavier mechs.

In other words I argue the opposite of the proposal is true. The problem is not that you are judged situationally, its that you are not judged situationally enough.. only win loss average sorting will remove that bias from the sorting system.