Monday, January 21, 2008

Reds sign Jeremy Affeldt

On Saturday, the Reds signed pitcher Jeremy Affeldt to a $3 million, 1 year contract. The 28-year old left-handed has pitched as a starter and reliever in his career, though he's found most of his success in the latter role.

He has pedigree: Affeldt was a 3rd-round selection by the Royals out of Northwest Christian High School in 1997. And he's big, 6'4", 215 lbs. But let's talk performance.

Stats:
Year Age Team IP K/9 BB/9 HR/9 %GB BABIP ERA FIP OBPa SLGa OPSa R/G RAR FIPRuns
2005 26 KCR 49.7 7.1 5.3 0.5 52.8% 0.329 5.26 3.99 0.366 0.386 0.752 5.2 -1.2 3.2
2006 27 KC/COL
97.3 4.4 5.1 1.2 48.6% 0.270 6.20 5.65 0.356 0.410 0.766 5.6 -3.8 -8.4
2007 28 COL 59.0 7.0 5.0 0.5 53.0% 0.262 3.51 4.17 0.328 0.375 0.703 3.0 13.8 4.2
3years --- --- 206.0 5.8 5.1 0.8 51.1% 0.279 4.94 4.78 0.264 0.392 0.656 4.7 8.9 -1.0

Holy control problems, Batman! The first thing that jumps off the page when I see this dataset is his walk rate, which has been consistently awful over the past three years. His strikeout rates also are unimpressive over the span of his 200 innings from '05-'07...though they've been pretty variable, and were above average last year. He does seem to throw fairly hard, averaging 94 mph on his fastball last season, and had a good 16 mph difference between his heater and his breaking stuff.

The biggest asset that Affeldt seems to have is the ability to prevent the fly ball, which has resulted in below-average home run rates. He's also left-handed, of course, which is certainly beneficial--especially in the bullpen. I see him as being pretty similar to Jon Coutlangus. ... and I'm honestly not sure that he's a whole lot better, especially if Cout can drop his walk rates a tad bit next season.

Affeldt admittedly did have a nice season last year in Coors'. The R/G statistic above is park-corrected (assumes Coors' PF = 1.09), and it's responsible for the very impressive +14 RAR he posted in just 59 relief innings--and that's without adding an extra 14% for the extra leverage of the situations in which he pitched. But indications are that this was due, at least in part, to some good luck: His BABIP was a tad low, and his FIP is about 0.8 runs higher than his ERA. He was also playing in front of a good infield defense, especially that guy they have at shortstop. Furthermore, I think it's dangerous to read too much into 59 innings of work, especially when it contrasts so heavily with the other 145 innings of the past three years.

Of course, the word surrounding this signing is that he might be installed as a #5 starter if two of Belisle/Bailey/Volquez/Cueto falter or one of the Big Two get hurt. What would a move to the rotation do to his performance?

Well, here's a very dirty way to estimate it: based on numbers from Tom Tango (yes, I rely on his work too heavily, but there's not much else out there like it), a replacement-level pitcher would allow runs at 107% of league average in relief, and 128% of league average as a starter. If the league-average ERA is 4.43 (5-year MLB average), then a replacement-level pitcher in relief should have an ERA of 4.74, but an ERA of 5.67 as a starter. That puts the shift in ERA at about 0.9 runs when moving from a relief role to the rotation.

So, if we take his 2007 FIP as a good-case scenario for his skill, and add 0.9 runs to it to reflect his transition into the rotation, we're suddenly looking at a 5.00 ERA pitcher. And that's the good-case scenario. If we instead use his 3-year average FIP (admittedly includes 9 starts), we project him as a 5.70 ERA pitcher--essentially a replacement player.

One thing I will say that he has going for him as a starter: he has had relatively small lefty/righty splits in his career (0.767 OPS vs. right, 0.747 OPS vs. left), so that, at least, does suggest that the rotation is a good spot for him. If his production can best that of a replacement player in the rotation, it would be a good spot to leverage his talents.

Money Matters

$3 million for a free agent is paying for 0.7 WAR, or ~7 runs above replacement, based on Tango's salary scale.

Affeldt's FIP over the past 3 years indicates that Affeldt produces at just around replacement-level. That would put his value at 400k--league minimum.

On the other hand, his performance last year, based on FIP, was worth 4.2 RAR--and we should probably tweak that up to 4.9 RAR to reflect the leverage of the situations in which he pitched. Add to that the potential for his groundballing ways to translate well in GABP's conditions (low IF park factor--at least, it was last I checked), and you can make a reasonable argument for 7 RAR.

Update: thought about this last night while I was trying to sleep:

If Affeldt is able to pitch at a 5.00 ERA as a starter, and he gets 150 innings, he'd more than justify his price tag, as that would be the equivalent of ~12 RAR ([5.7-5.0]/9*150=11.7). I'm not sure that he can do that, but if he can keep his strikeout rate high while in the rotation, he should be able to achieve that level of "effectiveness."

So the deal is fair. I guess the big question is whether he represents an upgrade over someone like Coutlangus or Stanton--much less Bray or one of the starter candidates. And I think that's fairly questionable.

Photo by Getty Images/Doug Pensinger

5 comments:

  1. If you look at his pitch F/x data from Josh Kalk, you can immediately see why he is having control problems.

    ReplyDelete
  2. ?

    I looked at his player card (and linked to it in my post), but I don't see anything that allows much diagnosis. Pitch velocity and break does tend to vary a lot, but that happens on pitchers with really good control too. Aaron Harang doesn't look noticeably more variable than Affeldt to my eye. But I haven't tried to really measure those things either.

    Anyway, feel free to elaborate.
    -j

    ReplyDelete
  3. Taylor, can you be more specfic? I don't really have a good sense on what a good release spread looks like or if we should expect tighter plots of movement data.

    Your comment suggests you do.

    ReplyDelete
  4. I think Taylor's point was that Affeldt's control problems are due to a complete lack of control... :-)

    ReplyDelete
  5. Oh crap. Totally forgot to come back since being out of town for awhile.

    I was looking at the pitch f/x tool that can isolate the pitches, result, velocity, count, etc.

    My fault for not using the same data set to compare or make my comment.

    As I said before I am still trying to figure some of this stuff out and should probably do more homework.

    ReplyDelete