Can Someone Explain the Bill James Projection of Scott Baker?

by Jeff Zimmerman on Nov 22, 2010 3:00 PM EST in Player Profiles

I am doing some fantasy previews and got to Scott Baker and looked at his stats at Fangraphs which currently includes Bill James's projections. Bill's projections have always been optimistic compared to other projection systems, but Scott's was low and I just can't follow it. Take a look:

Season	Team	W	L	SV	G	GS	IP	K/9	BB/9	HR/9	BABIP	ERA	FIP
2005	Twins	3	3	0	10	9	53.2	5.37	2.35	0.84	0.265	3.35	3.82
2006	Twins	5	8	0	16	16	83.1	6.7	1.73	1.84	0.359	6.37	4.99
2007	Twins	9	9	0	24	23	143.2	6.39	1.82	0.94	0.327	4.26	3.89
2008	Twins	11	4	0	28	28	172.1	7.36	2.19	1.04	0.290	3.45	3.79
2009	Twins	15	9	0	33	33	200	7.29	2.16	1.26	0.287	4.37	4.08
2010	Twins	12	9	0	29	29	170.1	7.82	2.27	1.22	0.329	4.49	3.96
2011	Bill James	9	8	0	26	26	152	7.16	2.19	1.18	0.311	4.03	4.01

Problem 1: Why does he have him starting 29 games in 2011, at least 2 less than any of the past 3 years. This tough does match with the drop in IP of 18 from 2010 to 2011. Maybe he is looking at other Twin pitchers getting some starts. Definitely no kind of standard weighting used in other systems is being used here.

Problem 2. The K/9 drop is quite rough. It is lower than any of the previous 3 seasons. He must be factor in a pretty heavy aging factor.

Problem 3. ERA of 4.03. Some his ERA is going to be lower than anytime in the past 2 years, but his K/9 is going down. Also with a 4.03 ERA, you would think the Twins will play him more.

I just can't make any sense of this at all. He is is supposed to allow a half a run less than 2010, by having his K/9 drop by 0.76. With his ERA dropping, the Twins also decide to play him less.

I just don't get it.

Tweet Comment 11 comments | 0 recs |

Do you like this story?

Comments

Thought he was unlucky last year?

That 4.49 ERA came in the same year as a 3.96 FIP. He expects quite a BABIP drop, and a slightly lower HR and BB rate. With regards to the start/innings drop I guess he just believes he’s injury prone?

Roto Hardball | @CajoleJuice

by CajoleJuice on Nov 22, 2010 3:30 PM EST reply actions

Maybe James’ algorithm thinks his 2010 ERA was unlucky and just regresses it towards the FIP. I think that one is ok. The IP algorithm is probably flawed and could be just looking at the trend down in IP from 2009 to 2010 and is extrapolating it down again further in 2011. The K/9 drop is likely some combination of regressing and aging.

I usually only use these projection systems BJ, Marcel, Zips, Chone (take your pick) for their rate stats and change the raw totals based on my own PA projections or IP projections etc… I think only the IP number is suspect here.

by Xeifrank on Nov 22, 2010 3:31 PM EST reply actions

Any particular reason you want to rely on Bill James projections?

Or are you just curious?

Blogger and Editor, Rational Pastime Blog. Twitter: @RationalPastime.

by J-Doug on Nov 22, 2010 3:32 PM EST reply actions

I don't rely on it at all, I just can't figure it out where his is getting his numbers from

- .-. ..- … – / – …. . / .—. .-. - .. . … …

by Jeff Zimmerman on Nov 22, 2010 3:39 PM EST up reply actions

I wouldn't be surprised if IP were regressed to league average starter

or league average # whatever starter Baker is.

by Dan Turkenkopf on Nov 22, 2010 8:38 PM EST reply actions

1) He’s only had one season out of 6 where he made a full ML season of starts. You can say that he’s just as likely to miss significant time with injury as he is to pitch every 5th day for 6 months. It makes sense for a projection system to kind of split the difference

But that’s just an undetailed, educated guess on my part.

2) Age, and his 2010 K/9 does seem like a bit of an outlier – however, his 2007 also sticks out (6.39) and, according to fangraphs, he’s added 1.2 mph to his fastball since then.

3) His BABIP was high (compared to average and the two years before that), and his 2010 FIP was a lot closer to that 4.03 number. It is fair to expect his ERA to fall more a little more in line with his FIP and his peripheral stats in the future. ERA is very inconsistent year to year due to all the factors outside a pitcher’s control (defense, park, luck).

Hope this helps a little, even though I don’t have much insight into all of Bill’s methodology.

Eugenio Velez: Really fast... at getting picked off. SICK BURN!!

by jasomack on Nov 22, 2010 8:54 PM EST reply actions

Also: like CajoleJuice said, they have him projected to have slightly lower HR and BB rates, and grammar fail a lot on my part from me typing run-on sentence. It’s the vicodins.

Eugenio Velez: Really fast... at getting picked off. SICK BURN!!

by jasomack on Nov 22, 2010 8:57 PM EST up reply actions

An Answer Can Be Found...

RE: Innings Pitched
I would gather James’ projection suggestions that 2009 was an anomaly as that is the only season Baker started more then 29 games. Possibly this data suggests to James that a pitcher who consistently misses a handful of starts a year (for whatever reason) is likely to see his starts further decrease – specifically during his age 30 season.

RE: Strikeouts
His K/9 figure is actually higher then his career rate. While the projection calls for a figure much less then 2010, this again can be coined as an aberration as his 2009 and 2008 seasons are quite a bit lower. Thus, the 2011 season can be those two seasons PLUS attrition.

RE: ERA
James calls for a K/BB essentially on par with Baker’s career rate and as such, one could expect his ERA to fall close to his career FIP. While this is .3 runs lower then his career ERA, it has to be difficult to adjust for misfortune, which, according to FIP, Baker has consistently fell victim to.

The innings pitched certainly sticks out, but everything else seems to line up with his career numbers. Baker saw a spike in his K/9 in 2010 as well as a spike in his BABIP. Maybe both of those are legit, one of those, or none of those. James seems to think that the previous data speaks higher for the “type” of pitcher Baker is as he pretty much expects a “career average” season.

To me, this isn’t overly strange and doesn’t really require an explanation.

The internet's latest attempt at understanding the insides of baseball.

by bbdbrandon on Nov 23, 2010 12:00 AM EST reply actions

Johan Santana's projection makes much less sense

http://www.fangraphs.com/statss.aspx?playerid=755&position=P

How is his K/9 going to jump back up that much? And does he not realize he went under major shoulder surgery? 195 IP??

Roto Hardball | @CajoleJuice

by CajoleJuice on Nov 23, 2010 1:25 AM EST up reply actions

He looks like he takes the career stats with pitchers and adds an aging factor.

- .-. ..- … – / – …. . / .—. .-. - .. . … …

by Jeff Zimmerman on Nov 23, 2010 8:59 AM EST up reply actions

Ah, alright

So there’s not exactly individual analysis going on here, just a model. I admit to being fairly ignorant as to how most of the projection systems work.

Roto Hardball | @CajoleJuice

by CajoleJuice on Nov 23, 2010 12:08 PM EST up reply actions