1. its not a mess. in fact theres a name for it. its called a trimmed mean. statisticians use a trimmed mean all the time in order to get a mean that more accurately reflects "the middle". however you are correct, the best game should be removed as well. doing so produces a qb rating of 80. the trimmed mean for the entire season would be ~97. in calculating these figures i didnt do anything special. i just pulled data from pro football reference.
source: [
www.pro-football-reference.com]
2. in any event, i think we agree goff played better in the early part of the season compared to the latter. the difference in opinion seems to come in who much better and why. i contend that the reason is the result of the factors that i listed, including goff bearing some of the responsibility. you seem to want to place the blame wholly or largely on goff. and if i understand you correctly, your evidence is the qb rating. in response, i will point out that the qb rating is not ALL on the qb. the line needs to block, reviever need to run routes and catch balls, and the coach needs to make smart calls. nobody on the football field works in isolation. for example, guess which game gurley had his lowest rushing total? yup, not coincidentally the same game that goff stunk it up, chicago. his second worst? the eagles...also one of goff's worst. coincidence? i guess i could run a regression to get a definitive answer , but i generally get paid for that.
3. my point is goff should be hitting is stride this year. but improvement as a player is not a straight linear line that goes up every game. it is filled with ebbs and flows. moreover goff helped them make it to the sb. they did not make it despite him.