I try to go with the stats that give me the finest granularity for the effects I am looking for... for example, for adjusting plate discipline, etc., I am looking at "per-pitch" stats like Swing %, Miss %, O-Swing % (how often a ball is chased), etc. For play results effects, I'm looking at per-plate-appearance stats, like BB%, SO%.
I am actually not going by "per-game" stats because extra innings, lack of bot 9th being played can skew results. Theoretically you can get perfectly realistic 4-run per game results, even when every game is 18-innings marathon, with each team hitting exactly 4 HRs but striking out the rest of plate appearances... you know what I mean?
I prefer BABIP because it removes plate discipline factors from evaluating hit results... The premise of BABIP is that once the pitch leaves the pitcher's hand, he doesn't have a whole a lot of control over what happens on the (batted) ball, and therefore a whole a lot of pitchers tend to have their BABIP close to .300 over a long run (there are exceptions). A stable stat like that is a better target than some stats that can change wildly depending on situations and abilities (like SB%).
If all hits are fly balls, the vast majority of them (except HRs) would be caught so BABIP tends to be low... if all hits are grounders, a majority would be caught by infielders but some go through the holes, resulting in a slightly higher BABIP... if all hits are line drives, the majority would be solid hits somewhere, and BABIP tends to be much higher. In that sense, from BABIP we can infer if we are getting realistic hit type fractions (as well as how realistic fielders are playing the batted balls) as well.
All in all, if the game is doing a decent job on per-pitch statistics (which The Show mostly does a fantastic job), the rest of the stats tend to come out very balanced and that's why I am tracking them.
Not sure what you mean by SIM (as in simming games?), but the only elements that maxing out Pitch Speed sliders affect is the base stealing in CPU vs. CPU games. I prefer to max them out because that settings brings visual speed of pitches close to real-life pitch speed, i.e., 90 MPH on display travels like a real 90 MPH pitch. If you leave that at default, I think 90 MPH pitch actually feels like a high school fastball.
It's purely a preference.
I'm definitely on the side seeing more low-offense games, but there seems people who are not seeing this.
It could be the effect of "small sample" (likely) or perhaps some settings might be affecting how the game plays (like default roster vs. carryover)? Will see once people play more games.
And a slider set from one year tends not to work that well on another year. The game is tuned quite differently each year, so we always kinda have to start from a clean slate.