This is a good set of questions on a fairly complicated subject. Some things I'll note:
- ZiPS calculates a players underlying numbers (Ks, BBs, HRs, etc.) and then calculates FIP and estimates ERA from there. The fWAR you see on various player pages is based entirely on that projected FIP. HOWEVER, when you read the big hefty ZiPS projection article it uses a variation of WAR that attempts to include contact quality. Ben Brown is a great example, he's projected at 1.0 WAR above and 1.6 WAR on his player page
- This is a bit confusing and annoying, but the flip side would be every projection system on a player's page following different rules which would be even more confusing
- The Cubs' defense is stellar, but we have to take care to not double count that. If the Cubs defense saves 50 runs in a season that's 0.30 runs of ERA. So you can't give PCA/Swanson/Hoerner their defensive runs and then also give Boyd full credit for beating his FIP by 0.40 runs
- I do not use bWAR much. I believe it takes a pitcher's ERA, strips out the impact of defense, and then gives the pitcher the rest of the credit. I philosophically do not agree with this. IMO there's far too much luck even in a full season for that to be a reasonable approach. Over larger samples I'm less opposed (e.g. I'd use bWAR for Kyle Hendricks' full career but not for an individual season)
- The magnitude of the contact management stuff isn't huge on a team level. The Cubs last year had a 3.81 ERA, 4.04 xERA, and matching 4.16 FIP/xFIP marks. That xERA/xFIP gap is what you'd potentially consider the soft contact skill to be. 0.12 runs of ERA even over a full season is 20ish runs total. Add in roster turnover and regression and the effect is small enough that it's not going to significantly move the needle. I also don't blame someone for just throwing it out with the bathwater entirely
- I do not have the citation for this, it was in a chat years ago, but Dan Szymborski has said you need hundreds of innings to confidently say someone is a FIP beater. It's basically a situation where it takes so long to establish that someone is a FIP beater, by the time you can say it definitively looking backwards you have to start worrying about whether the talent level will hold up looking forward
And then I know this is way too long already, but a few tangentially related thoughts on evaluating pitching:
- I tend to look at xFIP first when evaluating a pitcher and also give it the most weight. It is focused on only the things a pitcher can control, and isn't an overengineered black box like BaseballProspectus' stats.And IMO
- xERA is the ideal compliment to xFIP. There is a good bit of luck to contact quality, so I do think xFIP carries a good bit more weight, but all of the reasonable counterarguments to xFIP are handled by xERA. If you use xFIP and xERA as brackets you are almost certainly capturing the range of how a pitcher actually performed IMO
- This is a little reductive, but for the sake of nice round numbers half a run seems to be the extent to which someone can under or overperform their peripherals at a true talent level. Anything more than that and you should credit/blame the defense or assume it'll smooth out with more time