Zergling, on 09 August 2017 - 01:10 AM, said:
Just worked out averages for some prior Leaderboards.
So yeah, once again Cougar and Uziel are demonstrating that they really are godawful bad.
And Annihilator is performing worse than the Supernova, so it isn't OP or in need of nerfs.
I don't think it's quite that simple. The fact that this leaderboard event divided players up and dissuaded people from farming really high scores (vs. trying to place on multiple leaderboards) means the scores can't be accurately compared against prior leaderboard events.
Averages were brought down by the number of variants that people simply didn't play in volume. The Uziel is bad and needs buffs, but the majority of people on the Uziel -5P leaderboard didn't even play 10 games- depressing its average below what you would get with a real sample size. More generally this is seen in reinforcement variants often having worse leaderboard scores- SNV-C vs. SNV-1 for instance. If we think of all the participants as existing on a bell curve, then we can imagine the effect number of number of participants has on average score on the leaderboard. If there are only like 150 people that play a given variant (like the UZL-5P), then that leaderboard is a cross-section of the top 50% players. If there are 5000+ participants (like the KDK-3), then the top 75 players are going to be the absolute cream of the crop. It only ends up being a very tiny data point and doesn't really tell you anything about how much better one variant actually is from another.
Fewer people playing = worse average scores. Good rewards and more leaderboards = less incentive to push leaderboard score on specific variants.
Dividing the ANH-2A and the ANH-2A(S) and the MCII and MCII(S) is a direct and practical example of how dividing up the players between more leaderboards drops the average. Take the top 75 from both leaderboards (like how they were combined on earlier leaderboards), and the ANH-2A average jumps up to 2790, the MCII average to 2846. They are both real powerhouses compared to anything else released in the past year- comparatively the MAD-IIC averaged only 2671. However, we can't really make an accurate judgement whether this means they're both better than the MAD-IIC (or Supernova) was on release, because of how influenced it is by the volume of players to pump the average up.
We don't really know how many were purchased and played in the leaderboard event in earnest, so we can't really say if the top 75 players are the top 1% or top 5% or top 50% of the bell curve except in cases like the Uziel where it's clear that you didn't even have to play 10 games to place in the top 75.
While I think we all hope that other 100t mechs get much-needed buffs, recent nerfs suggest that Chris is taking an organic approach, and nerfing variants just because they're overperforming in relation to other variants of the same chassis. Even if the Annihilator isn't as good as the MCII, it wouldn't be surprising if it got nerfed literally just because it's the best 100t mech.
On top of that, there's always the small but nonetheless latent potential for a general nerf to Assaults (probably Heavies as well). Assaults are easily the best-performing class in terms of match score, and with the release of two immensely strong chassis, coexisting in different niches, it would be foolish to think Assaults are in anything other than a golden age right now.
4
