Combined Queues - Final Discoveries
#21
Posted 25 May 2020 - 04:52 PM
Honestly, the Tier should be calculated on a real+time basis. Breakpoints between tiers should be set as quintiles of the PSR of players that have been active in the recent past. Don't be shy about demoting people of their past performance doesn't keep them in a current tier.
#22
Posted 25 May 2020 - 04:53 PM
I do not need to be constantly matched up with pre-made tryhards who plow through mine or the red team in roughly 2 minutes after contact is made, leaving bugger all to do for everybody else. There's no fun or challenge in that.
I also am not that new to this, that I need to dropped into a match with the poor bugger who stopped their light mech right in front of my lazervomit Timby, to plink away at my nosecone with their mgs for a bit. Until I alpha'd their CT to smithereens, still mildly confused. There's not much fun (well, a little), but certainly no challenge in that.
If I can get reasonable amounts of fun and challenge out of this game, I'd stick around. The above examples do not provide that, so if a tier reset can actualy assist in that, all for it.
Spawns and weight distribution should be high on your list too though.
Edited by TheFourthAlly, 25 May 2020 - 04:54 PM.
#23
Posted 25 May 2020 - 04:56 PM
Kurlon, on 25 May 2020 - 04:44 PM, said:
As someone who's career is based on data, shoddy communication like Paul's on this does not inspire confidence in his analysis. If you can't communicate analysis effectively and precisely, you probably also can't do analysis effectively.
#24
Posted 25 May 2020 - 04:56 PM
#25
Posted 25 May 2020 - 04:57 PM
Edited by Scout Derek, 25 May 2020 - 04:57 PM.
#26
Posted 25 May 2020 - 04:57 PM
#27
Posted 25 May 2020 - 04:58 PM
#28
Posted 25 May 2020 - 04:58 PM
#29
Posted 25 May 2020 - 05:01 PM
#30
Posted 25 May 2020 - 05:05 PM
#31
Posted 25 May 2020 - 05:05 PM
Yes. 1 Vote for resetting the PSR and back to Zero-sum.
#32
Posted 25 May 2020 - 05:06 PM
#33
Posted 25 May 2020 - 05:06 PM
#34
Posted 25 May 2020 - 05:06 PM
Paul Inouye, on 25 May 2020 - 03:41 PM, said:
The last time I communicated I set the Match Maker configs (release valves) in this manner and left it for a while
We need your feedback on whether this is acceptable or not (losing current Tier standings).
Reset the tiers, but until you actually share real data I don't trust a word you write. You and Russ clearly wanted this to begin with and you've been very cagey about actually telling us real results, but Russ gave up the game in his twitter posts. Since you don't have a clue about how the game plays these days, I simply don't trust your conclusions. You guys have simply been too deceptive too much in the past to be trustworthy. Share the real data - let us draw conclusions - I guarantee there are a lot of people that understand stats better than you do playing the game. It's the only way you'll earn any trust back.
Edited by Anomalocaris, 25 May 2020 - 05:47 PM.
#35
Posted 25 May 2020 - 05:09 PM
~Leone.
#36
Posted 25 May 2020 - 05:09 PM
#37
Posted 25 May 2020 - 05:10 PM
#38
Posted 25 May 2020 - 05:11 PM
#39
Posted 25 May 2020 - 05:12 PM
First, I definitely have seen stomps up recently, not down. Worse, it looks to be directly related to the groups in the match. If one team's group plays well and the other team's group poorly, your team crushes and vice versa. Worse, if this disparity happens, it doesn't appear to matter how well or poorly the other players on each team play - the 4 man groups control the tide. I'm still for mixed queues, but doesn't this problem need to be addressed for it to work out?
Second, why does win/loss affect PSR at all? Isn't it pilot skill, not teamwork skill? In fact, if anything, your team getting stomped should count as a mitigating circumstance to your performance if you're in the, random number for example, second half of your team killed since you can't have as good personal statistics when your team gets stomped as you would have otherwise as the opponent can gang up on you so quickly. Even if you do well, you probably would have done better if you hadn't suddenly had a 4 on 1 situation. To a lesser extent, the opposite is true as well, i.e., its easier to do well personally when your team stomps, so why does the PSR system reward you even further for being on the winning team?
Edited by JoeCold, 25 May 2020 - 05:14 PM.
#40
Posted 25 May 2020 - 05:13 PM
We know you want to minimize dev investment at this point. Why not crowd source a new PSR measure? Pull together a sample data set (match, map, mode, match result, mech, damage out, damage taken, etc) and publish it. I would guess there would be multiple proposals for you to consider.
8 user(s) are reading this topic
0 members, 8 guests, 0 anonymous users