11 replies to this topic

[JC Daxion]

Teir system and PSR was before we had leader boards. Now we do, Why not base the rating system off them?

go back a bunch of months, take the average score. Use it as a baseline, then go up or down if you improve/decline.. Would stop tanking to a degree if it takes you months to go down a single tier



Maybe go by averages like 0-100-200-300-400

Basically if your performing well over time you get in the level you should be. (granted those numbers could be tweaked a little but you get the idea)

[Unnatural Growth]

YES!!

Tier 20, here I come!




I like the idea of averaging the last X number (100?) match scores for tier scoring. Seems like it would even out the player skills well.

?

[El Bandito]

You mean something close to that chart ISEN came up with? That might be an issue since it will lump heavy and assaults pilots together so queue time could become long. http://leaderboard.isengrim.org/

[Tier5 Kerensky]

A much better improvement would be to separate GQ and SQ leaderboards.

[Luminis]

A new tier system is sorely needed and average match score (adjusted by class) is as good a metric as any.

But from previous discussions, it seems like that means taking away players' "progress" (because that, apparently, is what PSR is) - and that is a big no-no.

[Kiiyor]

While I think the leaderboards should have an influence, I have issues with using it for tiers.

As statiscics go, average match score is a little too streaky for my tastes. You need a lot of matches under your belt to get a decent indication of consistent performance, but we could use parts of the score to influence tier growth/reduction a little better, I think.

You'd also need to differentiate between the group queue and the solo queue (which stats should absolutely be doing anyway) as it's often harder to get decent scores in the wind blasted badlands of the pug queues than it is with a decent team behind your back.

[Pjwned]

Because the number of matches you've played has a huge effect on your leader board ranking.

PSR is somewhat similar in a way because most people will end up in tier 1 eventually with enough games played, but you can have mediocre stats with a bunch of games played and be near the top of the leader board pretty easily.

Unless you just ignore ranking entirely and focus on the stats, then that's a bit better.

[Queen of England]

Average score doesn't tell you much without tier.

A good player in tier 1 will got a good average match score.
A good player running an alt account in tier 5 will get a absurdly good average match score from seal-clubbing.

[Cy Mitchell]

Queen of England, on 13 September 2017 - 02:38 PM, said:

Average score doesn't tell you much without tier.

A good player in tier 1 will got a good average match score.
A good player running an alt account in tier 5 will get a absurdly good average match score from seal-clubbing.

This is true but if the Tier is based on the match score of the last 100-300 games then the seal clubber will find himself in Tier 1 very quickly and playing against the other "good" players and not the seals.

Basing the PSR on a players last block of sampled games (100-300) would make the Tier system dynamic and ever changing. It would result in player rising and falling based on their most recent performances. There would not be an automatic rise to the top due to the number of games played. If you rose in to Tier 1 but did not have the skill to be there and started getting your butt kicked on a regular basis then your match score would suffer and you would drop down into the levels where you could compete and maintain your ranking.

Right now their are Tier 1 players who routinely get low match scores but stay in Tier 1. That would not be the case with a dynamic system based on the match scores of your recent matches.

Tightening up the MM to a max of 1 (or 2) difference in Tiers was a good first step to get better skill matches in game. Now fixing the Tiers to make them truly a reflection of demonstrated skill would finish the task of fixing MM.

Edited by Rampage, 13 September 2017 - 03:07 PM.

[TheFallOfTheReaper]

It's sad this is the most "constructive" idea anyone has had in the last like 10 months regarding psr/tiers, +1, with he caveat group and solo has different tiers/scoring, but it's pgi and the longer we hold our breath the faster we will suffocate hoping they get their **** together and come up with something useful :3

[JC Daxion]

Pjwned, on 12 September 2017 - 11:02 PM, said:

Because the number of matches you've played has a huge effect on your leader board ranking.

PSR is somewhat similar in a way because most people will end up in tier 1 eventually with enough games played, but you can have mediocre stats with a bunch of games played and be near the top of the leader board pretty easily.

Unless you just ignore ranking entirely and focus on the stats, then that's a bit better.

If you pick average match score, if you don't hold up over 100's of matches, your average score drops. Just look at the boards, and see how average match scores are, people that are good, are good.

Yes, if you only played a few matches one month you might end up higher than normal, which is why i suggested going back a bunch of months, and then going up or down in stages.


As for tiers and lower tiers being easier to get higher, i don't think it would be that way. They would obviously have to have some other steps taken.. But after accounts mature, you can tell if they are performing or not. Push them up, if they are good, they will still perform, if not they will go back down.


I realize numbers need to be taken into account when it comes to class, but then again good players still hit high numbers in all classes. Just look at the different weight classes and the top 100. People that are bad, are far down the list..


The PSR system could be used in some way too, as the starter point. It gets good players moved up quick, and middle of the road players go the long way of playing lots of matches. This would then push those that slowly creep up back down a notch every month if they are not performing well.

[Asym]

JC Daxion, on 13 September 2017 - 04:32 PM, said:

If you pick average match score, if you don't hold up over 100's of matches, your average score drops. Just look at the boards, and see how average match scores are, people that are good, are good.

Yes, if you only played a few matches one month you might end up higher than normal, which is why i suggested going back a bunch of months, and then going up or down in stages.


As for tiers and lower tiers being easier to get higher, i don't think it would be that way. They would obviously have to have some other steps taken.. But after accounts mature, you can tell if they are performing or not. Push them up, if they are good, they will still perform, if not they will go back down.


I realize numbers need to be taken into account when it comes to class, but then again good players still hit high numbers in all classes. Just look at the different weight classes and the top 100. People that are bad, are far down the list..


The PSR system could be used in some way too, as the starter point. It gets good players moved up quick, and middle of the road players go the long way of playing lots of matches. This would then push those that slowly creep up back down a notch every month if they are not performing well.

Everyone chides me when I make suggestions and brings up my Win/Loss record all of the time....

"If you're less than 1 you're a potato....ect, etc, etc...)

OK, let's use win - loss and tier based on the actual performance to that standard ! That way, should you "get GuD" or "digress into Potato hell", you'd be matched against players of the same W/L.....????

Why not?