www.omnipotents.com/forums

For anyone interested I updated the UTstatsDB analysis Google Sheet.
The second sheet has August 1st scores. I also added a column for the low end estimate of win/loss percent.
There is a 95% chance that the player's win percent is higher than this number. Basically, the error bars shrink with the number of matches the player has completed.

I know we've been through this before, but is there any way we can use win/loss for the balancer?

DrScrotey wrote: ↑Mon Aug 01, 2022 11:50 pm ...
I know we've been through this before, but is there any way we can use win/loss for the balancer?

Yes, but it requires a type of math that isn't represented on this tree yet. LOL

I know we've been through this before, but is there any way we can use win/loss for the balancer?

Not directly. However, we have our PPH stats, and snarf made an addition that gives Known Players a multiplier in an attempt to account for better players even though they may have same scores. Think of the scenario where an average player gets an inflated score from Mino or Levi kills...and then the next time the team ends up facing a Leon/Xexx/Enyo combination.

We just turned that off last week, as it was thought that since there's no damage points and some other changes, that scores would be more "accurate". After a week I don't think that's the case, IMO it was better before with the KnownPlayers Multipliers. The idea there was that some players play "better" than score.. for instance, and I'll pick on Leon, he's a very smart player--knows what nodes to go to, gets Double Damage often, uses vehicles to good effect... Enyo same, Xexx same etc. We could take the win percentage and create a multiplier list based off that-- that wouldn't take long to do as its all there... and IMO would be better than what we have now. The downside of KnownPlayer is the multiplers previously were subjective and has to be manually maintained -- for example some of our "better" players recently Sanka, 4Doors come to mind and did not have a multiplier--if we took it from statsDB we'd probably get most/if not all.

Back on Topic, the idea on fixing the scoring was that node building should generate more points, as the game is driven by nodes not kills. Maybe 8 pts for full node build vs. 5 right now, same with destroying, and I don't think you get points for destroying a building node.

I attempted to combine these here: https://www.omnipotents.com/utstats/onsrankings.php Its a little sparse right now since the statsDB reset, as you need 40 hours of play time to count...but we have many 30+ hours plus that will likely get there in the next couple of weeks.

Thanks for the detailed reply. I have a love/hate relationship with the rankings and such. I've gone months forgetting they are there and obliviously played happily, but when I remember they exist it becomes something that I compulsively try to optimize, as opposed to optimizing for enjoyment. That being said they're useful for balancing; I just need to lay off the compulsion.

My vote would've been No for myself, but yes for others lol.