GAC Skill and banners.
Maybe I'm misunderstanding something, so please correct me if I'm wrong--But if matchmaking is based off of Skill Rating, and Skill Rating is gained or lost depending on whether you win or lose (respe...
Forum Discussion
"PeachyPeachSWGOH;c-2324285" wrote:
Banner disparity can reflect the "closeness" at one level more granular than win/loss. It doesn't reflect how close at individual battle level of course, but still better than a simple binary outcome.
That said, coming from a chess background, I'm perfectly fine with a system where a win is a win, regardless of it was because your opponent blundered or you genuinely pulled off some Fischer style brilliance. :smiley: A big part of the rationale behind these ELO-like rating systems is that over the long run noises (like opponent blunders or disagreeable lunches) are going to get filtered out.
Weird scenario, but what about a couple runs of the binary win-is-win, loss-is-loss rating adjustments, then once a nice bit of separation and seeding has taken place switching to banner v banner adjustments (where a win is always going to be good, but the gain is mitigated by the closeness, for instance) so that initial noise doesn’t overly-affect the data?
