[NoOne]

"so I don't need a huge sample size to tell if a product feels "right" or not."

that pretty much explains it all. if you rely on feelings you will be fooled by randomness more often if you did not mix feelings and objective facts. that is the entire point of a larger sample size... you cannot tell with confidence in the absence of a suitable sample size. believing that you can is like believing in magic relative to the results you reported. i.e. they are not extreme outliers consistently occurring one after another. a little common sense of course, but that's a very small portion of situations when it's actually obvious.

in addition to the game environment outlined by others (league totals, modifiers, team strategies, talent distributions of eras etc etc ), it looks like you are pairing really good teams to play each other. that's certainly not representative of any year's average. it's quite difficult to have a good idea of a baseline since there is none. you can run X games of simulations between 2 teams, but i don't know what kind of data you can get out of it and if it's easily fed to a 3rd party spreadsheet/db.

i think this frustration is mostly due to the complexities of players from different eras playing concurrently. if i was more familiar, i'd make some suggestions.

if you do like these types of games, i'd put it down for a week or two and relax. then, at some point in future make a basic mlb 2016 and turn on the auto-calc league totals modifiers (make sure it runs each year)... think it's default? either way use all defualt otherwise.... just play and don't get detail orientated with the settings.

i bet you'll enjoy it a bit more... then you can get into customizing with the knowledge that when properly balanced and set up in a logical way, you can do all sorts of things with a game world in ootp.