Who Should Start Game Four for the Phillies?

Philadelphia Phillies skipper Charlie Manuel is expected to decide soon that Roy Oswalt will be the fourth starter for the club heading into the postseason. The decision obviously would come at the expense of 23-year-old standout Vance Worley, whose rookie season has made some folks in the City of Brotherly Love wonder if he’s the real fourth ace in this star-laden rotation.

Certainly, Manuel’s decision would have been expected in April when the Phillies entered the season with Roy Halladay, Cliff Lee, Cole Hamels and Oswalt on staff; but a flurry of ineffective starts — and two trips to the disabled list — have marred Oswalt’s season. Worley, meanwhile, now looks like the guy who should have been manning that fourth spot all along. His ERA is 3.00 — compared with Oswalt’s 3.66 — and Worley’s defense-neutral estimators all are about a third of a run better than Oswalt’s, as well.

So should Oswalt really be such a lock for that start? Well, here’s the way to tell: Tell both pitchers to warm up, and then put the radar gun on the 34-year-old’s fastball. If it’s above 91.5 mph, put in Oswalt. If it’s below — well, you know the answer.

Below is FanGraphs’ plot of velocity by start for Oswalt. Note that these charts don’t include his Sept. 12 start (91.87mph) and his Sept. 17 (91.85mph), which are a small decrease from the last start pictured, on Sept. 7 (92.18mph).

Notice that I’ve drawn a thick red line down the center at 91.5 mph, dividing his starts into games when he had his fastball velocity working and games when he didn’t. This red line represents the difference in between the fourth ace that the Phillies thought that they had all along and the below-average pitcher who has been bested by a bespectacled rookie. The ace Oswalt gives up half as many runs and strikes out 50% more hitters — while the other Oswalt has a strikeout rate that would rest somewhere between Livan Hernandez and John Lannan on the National League’s K% leaderboards.

Velocity

GS

IP

ERA

SIERA

91.59-93.04

8

52

2.25

3.32

89.85-91.44

13

73.1

4.66

4.29

While ERA can be a useful tool to evaluate pitchers, it struggles when dealing with small sample sizes. Because of that, I dug deeper to confirm my suspicions. I broke down Oswalt’s performance into buckets — shown above by thinner black lines on opposite sides of the thick red line, representing at 92 mph and 91 mph — to see how his peripherals and his SIERA moved with his velocity. In fact, Oswalt’s SIERA is more than run lower when he’s at his best, thanks to a strikeout rate almost double that of his inferior alter ego.

Velocity

GS

IP

ERA

SIERA

K%

BB%

GB%

Over 92.0mph

4

27

1.67

3.26

22.0%

7.3%

52.7%

91.5-92.0mph

4

25

2.88

3.40

17.3%

3.9%

49.3%

91.0-91.5mph

6

29.1

5.22

4.31

12.7%

5.2%

39.6%

Under 91.0mph

7

44

4.30

4.29

13.4%

5.7%

41.3%

Deciding between Oswalt and Worley isn’t really that hard because Manuel doesn’t have to choose between two similarly competent pitchers. Rather, he’d be choosing among two of three possible pitchers who show up in the bullpen: Worley, Good Oswalt or Bad Oswalt. Fortunately for the Phillies, the difference between Good Oswalt and Bad Oswalt is easy to root out. Charlie Manuel just needs to bring his radar gun.




Print This Post



Matt writes for FanGraphs and The Hardball Times, and models arbitration salaries for MLB Trade Rumors. Follow him on Twitter @Matt_Swa.


27 Responses to “Who Should Start Game Four for the Phillies?”

You can follow any responses to this entry through the RSS 2.0 feed.
  1. BDF says:

    Fantastic. I’ve been imagining that “What Do Advanced Stats Have To Tell Us About Charlie Manuel’s Game 4 Starter Decision?” would be my first article if I were fortunate enough to be selected as Fangraphs’ new writer, with the thesis that there would be anything. This isn’t exactly advanced stats, but it is advanced analysis, and it’s great.

    Although, the bullpen performance of a pitcher, especially a veteran like Oswalt, I think is not strongly correlated to game performance. Bouton has a great section on this lack of correlation in Ball Four. So maybe Manuel should go Koufax/Drysdale and start Oswalt and have Worley on the verge of warming up if it looks like Roy-O doesn’t have it.

    Thanks, Matt.

    Vote -1 Vote +1

    • Yirmiyahu says:

      So Oswalt pitches better on days when he has his good stuff. I’m guessing, if you did this same study with any power pitcher, you’d get similar results.

      Vote -1 Vote +1

      • jorgath says:

        While true, Oswalt is one of a “type” of pitcher wherein there is a much greater difference between “good Oswalt” and “bad Oswalt.” I’m guessing that if you did this same study with another power pitcher (if someone wants to try, I suggest Verlander) that yes, you’d see a dropoff on days when his pitches aren’t as good, but the quality difference wouldn’t be as severe as it is for Oswalt.

        Vote -1 Vote +1

  2. Jon Bongiovi says:

    Unfortunately for the Phillies, Oswalt could hit 70 on the gun pre-game and Charlie would still go with Oswalt because he’s “his guy”.

    Vote -1 Vote +1

    • Dekker says:

      The odds of Charlie Manuel sitting Oswalt down in his office to say “We’re gonna go with someone who’s performed better than you” is only as likely as Oswalt’s right arm falling off.

      Vote -1 Vote +1

  3. Telo says:

    Halladay.

    +17 Vote -1 Vote +1

    • joe says:

      Seems like the obvious answer, especially with the game 2 starter being able to come back for game 5 (if needed) on normal rest.

      And with the Phillies rotation the impact on the next series is not as significant – if you go game 4 with your #1), that means either game 2 of the following series on short rest again or game 3 on a bit extra rest… having Lee and Hammels for game 1 and 2 makes the impact of holding out your #1 for game 3 of the LCS (and having him lined up for game 7) not so bad

      Vote -1 Vote +1

  4. Jim Lahey says:

    I wouldn’t use anyone but Halladay/Lee/Hamels unless one got injured for the entire playoff run.

    Vote -1 Vote +1

  5. Puzzled says:

    Yes, with the built in rest, why not the Game 1 starter?

    Vote -1 Vote +1

    • Matt Swartz says:

      They eliminated off-days. It would only be three days rest for all three guys, which Lee & Hamels have never done, and which Halladay has only done once since 2008. It would be really risky.

      Vote -1 Vote +1

      • Telo says:

        Only two pitchers need to go on 3 days rest. The third would be 4. I was half kidding with my Halladay post… half.

        Vote -1 Vote +1

      • Telo says:

        Ah, I take it back, you’re right. It’s 3 days rest for everyonr

        Vote -1 Vote +1

      • hk says:

        The games in the NLDS are scheduled for October 1, 2, 4, 5 and 7 meaning that the Game 1 starter Halladay would be the only one pitching on 3 days of rest in the NLDS if they only go with 3 starters. Therefore, all we are really talking about is who starts Game 4 of the NLCS (if they get there) and Game 4 of the World Series (if they get there).

        With Stutes’s and Bastardo’s recent struggles and with Charlie Manual’s unwillingness to use his closer in a tie game on the road, it is possible that whichever one pitches out of the bullpen could have more value to the Phillies and pitch more high leverage innings during the entire post-season than the one chosen to start, especially if they only use 3 starters in the NLDS.

        Vote -1 Vote +1

  6. Prince Fielder says:

    They will only need three this year.

    Vote -1 Vote +1

  7. Joey E. says:

    so i guess Doc and Lee didnt learn their lesson’s last year?
    3 days rest. be men

    Vote -1 Vote +1

  8. Phils_Goodman says:

    Chuck out Oswalt’s numbers from June when he was clearly pitching through a painful back injury, and you’ve got a 3.09 ERA and a 3.00 FIP over roughly 100 IP. Oswalt has a decade of proven ace performance behind him and 10 starts of playoff experience. In that context, the June results were an obvious fluke and there was a clearly identifiable reason for it (which is no longer present). As long as Oswalt’s health shows no signs of breaking down between now and October, he is the solid choice.

    Vote -1 Vote +1

  9. cuck says:

    why not just go with oswalt, hope “good oswalt” shows up, and keep worley in the pen as long relief? if needed worley can do mop up or spot start. you get all 5 of your possible good starters and all of a sudden you dominate 4.5/5 games

    Vote -1 Vote +1

  10. Jefferey M. says:

    I know Matt you have generally defended Charlie’s decision making over his tenure as Phillies’ manager. Is there any defense at all of him bringing Oswalt back out for the 8th inning in an utterly meaningless game tonight, and then leaving him in to pitch to Morse with two men on, consequently having Oswalt finish tonight at 110 pitches thrown? It would seem to me that we’d want to have Oswalt throw as FEW pitches as possible to keep him fresh for the playoffs. It’s utterly mind-boggling Manuel would keep him in that long and for that many pitches.

    Vote -1 Vote +1

    • Matt Swartz says:

      I’m not sure I defend Manuel as much as Amaro, who I think has a bad wrap for one bad deal despite making up for it on other deals. Though I do think Manuel has a good rapport with his players, is probably an excellent instructor, and I’d imagine that erases his in-game decision making deficit.

      That said, I think Manuel decided to ruin my article yesterday by taking a “Good Oswalt” game and leaving him in long enough that he started allowing runs. I think he figures Oswalt has one start in five days and then a long layover, but I think it’s best to deal with him as fragile at this point. Regardless, yesterday’s velocity numbers from Oswalt portend good things IMO, since they’re gonna use him anyway.

      Vote -1 Vote +1

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Current ye@r *