Predictive Stats for Fantasy Baseball

[homer]

I'm curious as to what stats people use in the off season to predict value for fantasy baseball? I would think BABIP (w/ LD%) would be a good predictor on offense but what about pitching? xERA? FIP?

What stats do you use to prepare for your fantasy draft?

[TooLiveBrew]

I think BABIP & LD% are good ideas, but be sure in anything that you see more than just one year to judge anything (good/bad/avg.).

 

I don't play fantasy baseball anymore (tried for a couple years, but the constant attention required to your lineup/s really turned me off), but if I did, I'd focus upon the following for pitchers:

xFIP - be sure to factor in the defenders playing behind or projected to play behind the pitcher(s). Take a look at the differential between a guy's xFIP & FIP. If he's on a team that's relatively unchanged in INF defense, and has similar differentials, you can probably guesstimate pretty well.

BB/K ratio(s) - This one's kind of obvious, I guess. But since you get points for K's, and BBs will pretty directly impact other fantasy numbers (ER, W-L) negatively.

IP - Do you get points explicitly for this category? Even if you don't (another obvious-ish point), greater IP gives a guy a greater shot at nabbing the 'W.'

 

SLGA - check this one beyond just last year. Another one to watch similarly to BBs - pretty negative indirect impact. Ex. #1: Vargas, Claudio.

Try to find guys with good GB% on teams with good INF defense (Colorado, Arizona, etc.), and watch out for pitchers in ballparks that play toward hitters. On that note, specific to pitchers in 'hitter's parks': Watch that SLGA closely, along with HR/FB rates.

Don't know if any of this is any real help to you, homer, since you know your way around the numbers, but hope there's at least one nugget in there.

[rluzinski]

For established batters, if you use Marcel projections, they are just weighted average of the player's last 3 years, adjusted for age. It gets a lot tricker if you are trying to cut through the statistical uncertainty of a 500 AB sample. Not sure if I'm completely sold looking at BABIP compared to a batter's LD%, since defining a what's a line drive is still somewhat subjective. Of course, if a guy had a very low or high BABIP, I'd probably assume some regression, no matter what his LD% was.

[homer]

How accurate are those player projections? I was looking at Fangraphs today and they have the Bill James projections for each player and it shows Fielder regressing slightly and Braun actually getting better. I'm curious if anyone does a post mortem after the season to see how close they really are.

[TooLiveBrew]

I don't know where to find such a report or analysis, homer, but I guarantee you they're out there. I don't know a whole lot about stats & projections, but that James projection for Braun looks borderline ignorant. I've perused quite a few of his 2008 projections at fangraphs, and a lot of them look like he just guesstimated - but I doubt that's the case.

[TerraceReserved]

I've seen a couple of reports on how accurate the various projection systems are, but can't find the link anywhere. I do remember that PECOTA has been the most accurate, followed by ZIPS, then Marcels. I'm pretty sure all 3 of those systems beat Bill James.

[TooLiveBrew]

I'd be willing to bet that quite a few hometown diehards would beat Bill James on at least a player or two on their beloved team(s). It really appears that he doesn't do a ton of work to arrive at some projections imho.

[M Skjellyfetti]

Looking at the Bill James projection for Braun, I see what you mean, TLB. I can understand the slight rise in BB% and the drop in K's, but I don't think the drop in BABIP is enough. That seems really strange to me that a projection system could throw that out there. I'd love if that were correct, though.

[TooLiveBrew]

I'd love it to be true, too, C-man. I honestly feel that some of James's projections for this season are questionable at best. I can see why PECOTA & other systems have garnered more attention recently. Check out some of his other 'guesstimates' - seriously, some of them look to be just that - around the league. I don't have any ready examples, but just go check out some other guys whose names pop into your dome.

 

In particular, I feel like he glosses over ISO & BABIP. Sometimes BB rate, too. For example, he has Estrada to draw 20 BBs next year. Anyone who's seen Johnny in the past 2 seasons knows that's probably a longshot.

[homer]

Thought I'd bump this as we are closing in on various draft days.

 

What sites do you guys use/subscribe to to prepare for fantasy drafts?