Tier Ranking Reset
#41
Posted 06 June 2018 - 09:51 AM
It's no fix at all, not even a bit
#42
Posted 06 June 2018 - 10:23 AM
Only possible way is to track performance for each player in each mech with each loadout. A data nightmare at first glance, but builds can be mapped to a series of values that can be used to create a unique key that can be used in the DB. It can be tracked. A "battle value" can be calculated and used by the Match Maker when building teams. However, to do it PGI will have to sacrifice a sacred cow (even teams)...which means they won't do it.
#43
Posted 06 June 2018 - 10:25 AM
But that would require PGI to actually eat a book on game design by accident and somehow still absorb some knowledge.
#45
Posted 06 June 2018 - 10:57 AM
ShaneoftheDead, on 06 June 2018 - 10:23 AM, said:
Only possible way is to track performance for each player in each mech with each loadout. A data nightmare at first glance, but builds can be mapped to a series of values that can be used to create a unique key that can be used in the DB. It can be tracked. A "battle value" can be calculated and used by the Match Maker when building teams. However, to do it PGI will have to sacrifice a sacred cow (even teams)...which means they won't do it.
I don't think you have to get that granular or complicated. Track performance by weight class and you'll get 90% of the way there. Average matchscore of a player's last 50-100 games in each weight class and then rank them. You can improve it further by not counting the rewards only lights and mediums are eligible for. This would prevent a player with a 300 matchscore in a light being treated the same as an assault with a 300 matchscore.
#46
Posted 06 June 2018 - 10:59 AM
Kargush, on 06 June 2018 - 10:54 AM, said:
Which makes those who fail to get to T1 even less competent--which means tiers DO convey overall competence.
#47
Posted 06 June 2018 - 11:13 AM
#49
Posted 06 June 2018 - 01:02 PM
Everyone with an average match score of under 250 gets dropped 2 tiers. Everyone with an average match score of 325 or less gets dropped 1 tier. If you're averaging more than that, you stay where you're at.
If you look at the leaderboards right now, that would mean the top 1000 players stay in their tier (almost all Tier 1 I'd guess). Players in the 1000-5000 ranking get dropped 1 Tier (would mostly be tier 1s doing down, although you'd have a smattering who are Tier 2 or 3 and on their way up, they'd quickly advance).
As a Tier 1 player who'd be dropped to Tier 2, I'd still be facing most of the same players I do right now. _However_, when the matchmaker looks at who to group I would no longer be considered an equivalent match for an elite player like any of the guys from EMP. This would help better match player skill within matches, even if the actual players in the match do not change.
Mind you it doesn't fix the problem of god tier players dominating (someone who averages 450 match score vs. someone who avgs 350 is no contest most of the time), but it would help improve things to some extent IMO.
Edited by SFC174, 06 June 2018 - 01:02 PM.
#50
Posted 06 June 2018 - 01:11 PM
Be clear. Be super clear. Be talking-to-a-group-of-five-year-olds clear about what exactly it is you are asking before you mention it to Russ or Paul or whoever. History shows that they kind of do their own thing anyway, even if there's a hint of a whisper of a thing someone once asked for long ago at its root. That's all it takes for a "This is what you asked for!" reply so... just be really clear. For the record, I think the PSR is broken (it accomplishes what it was designed to but it also does lots of other unhelpful things that largely negate that intention) and a reset is a great idea. It'll mean a day or two of T5's mixed with veterans but if you put that multiplier in like new players get in the beginning anyway it'll all sort out.
Happy Gaming!
-BWS2K, mistlynx4life, Z3r0 D4m4ge, Overpowered Machine Guns, and probably one or two others I can't think of, lol
#51
Posted 06 June 2018 - 01:19 PM
I wonder also, if you are THAT bad that you repeatedly score such bad numbers with heavies...maybe you should play a medium mech. If you hide or screw up, you do not rip a hole into your team. Also, it is easier to bail out with a fast medium
Edited by Bush Hopper, 06 June 2018 - 01:29 PM.
#52
Posted 06 June 2018 - 01:37 PM
SFC174, on 06 June 2018 - 01:02 PM, said:
Everyone with an average match score of under 250 gets dropped 2 tiers. Everyone with an average match score of 325 or less gets dropped 1 tier. If you're averaging more than that, you stay where you're at.
If you look at the leaderboards right now, that would mean the top 1000 players stay in their tier (almost all Tier 1 I'd guess). Players in the 1000-5000 ranking get dropped 1 Tier (would mostly be tier 1s doing down, although you'd have a smattering who are Tier 2 or 3 and on their way up, they'd quickly advance).
As a Tier 1 player who'd be dropped to Tier 2, I'd still be facing most of the same players I do right now. _However_, when the matchmaker looks at who to group I would no longer be considered an equivalent match for an elite player like any of the guys from EMP. This would help better match player skill within matches, even if the actual players in the match do not change.
Mind you it doesn't fix the problem of god tier players dominating (someone who averages 450 match score vs. someone who avgs 350 is no contest most of the time), but it would help improve things to some extent IMO.
Firstly, in my plan everyone would stay I thier current tier until thier first match, no tier 5s meeting tier 1s ever. Secondly there would be no average to seed from as it would be a history wipe, if you got under a 250 your first match, you would drop a tier for your second. Thirdly, there would not be the huge momentum folks currently have where you cannot change tiers because of being 10,000 games in. Best players would rocket upwards, 240 players like myself would coast slowly downwards. You could even start folks at tier 5 with say 1000 points left till 4. Abritraily make the tiers 2000 points deep, no idea what this number should be, the larger it is the slower folks change tier. Sorry, this does little to protect the very good from the best in game, but I have a feeling those folks can better handle it and would look forward to the challenge. The 250 matchscore was chosen by me based on the PGI target, but so are all the numbers that make up matchscore, and I do believe those need a second look anyway.
#54
Posted 06 June 2018 - 01:40 PM
Most people evolve a chassis to a relatively stable setup that changes only incrementally or infrequently. This would allow them to stabilize an overall Elo for that chassis.
Then assign a tier for that player in that chassis based on their position within a rank ordering of the population that has been active within the past 30 days.
There's no real need to publish tiers if a system is working to put similarly capable players into matches together. Similarly, in Solaris, there's no need to divide mechs into divisions if the Elo is calculated across the full population.
ShaneoftheDead, on 06 June 2018 - 10:23 AM, said:
Only possible way is to track performance for each player in each mech with each loadout. A data nightmare at first glance, but builds can be mapped to a series of values that can be used to create a unique key that can be used in the DB. It can be tracked. A "battle value" can be calculated and used by the Match Maker when building teams. However, to do it PGI will have to sacrifice a sacred cow (even teams)...which means they won't do it.
#55
Posted 06 June 2018 - 01:47 PM
Edited by kapusta11, 06 June 2018 - 01:48 PM.
#56
Posted 06 June 2018 - 02:01 PM
#57
Posted 06 June 2018 - 02:06 PM
#58
Posted 06 June 2018 - 02:15 PM
Tahawus, on 06 June 2018 - 01:40 PM, said:
Most people evolve a chassis to a relatively stable setup that changes only incrementally or infrequently. This would allow them to stabilize an overall Elo for that chassis.
Then assign a tier for that player in that chassis based on their position within a rank ordering of the population that has been active within the past 30 days.
There's no real need to publish tiers if a system is working to put similarly capable players into matches together. Similarly, in Solaris, there's no need to divide mechs into divisions if the Elo is calculated across the full population.
What if they seeded your initial match with your average performance on my tier 2.0 system, then drifted based on performance localized to that chassis? Say you are in tier 2 because you sometimes play assaults and average 350 per match but that only represents 25% of your play, while in mediums you averaged 200? They could seed you based on:
1. Your average across all play
2. Your average across that weight class
Or 3. Tier 5 because you just bought that chassis?
I don’t think option 3 is very realistic, they have to start with some data to accurately place you. You could even place a modifier on that initial seed based on global performance of that variant. I also don’t see ditching historical data because you moved armor around or tweaked your heatsinks.
#59
Posted 06 June 2018 - 03:09 PM
Edited by LordNothing, 06 June 2018 - 03:15 PM.
#60
Posted 06 June 2018 - 03:12 PM
Nightbird, on 06 June 2018 - 09:51 AM, said:
It's no fix at all, not even a bit
That is where you miss then point.
A bunch of people in Tier 1, should be in Tier 4. With them no longer in games for Tier 1 (or 2) players and doing their sub 100 damage constantly... Such things are removed. Also their game experience would also improve as they are now back with people of their own skill level and will never have Tier 1 in their game again.
That is a win-win.
27 user(s) are reading this topic
0 members, 27 guests, 0 anonymous users