Clone Wars: James Shields and Roy Oswalt

While Roy Oswalt has been the ace in Houston for some time James Shields has just taken the role for the Rays. The similarities in their struggles this year are quite striking. We can take this chance to see what these similarities tell about their 2010 season and beyond.

Florida News - May 14, 2009

Both pitchers had lower strikeout rates last year going from career rates over 7 down to 6.85 per nine innings for both of them. For neither pitcher that is a career low so not a huge concern. These two pitchers are not largely known for their strikeout abilities.

Both pitchers are control specialists with great walk rates. Shields has a career rate of 1.96 and Oswalt is only slightly higher at 2.06. Oswalt was right at his career rate with a BB/9 of 2.08 and Shields was only slightly up at 2.13.

So they declined slightly in K/BB, but they were still very good. Oswalt has had better groundball rates in the past and that seems to have played a role in his drop last year. He went from a 50 percent groundball rate the previous season to a 43 percent. That was the biggest reason for his struggles last year as his xFIP went from 3.55 to 3.88. The xFIP shows his 4.12 ERA was also a bit unlucky, but with a career xFIP of 3.58 and a career ERA of 3.23 it’s surprising to see him trail his xFIP like this.

His LOB% seams to be the main explanation at 72.7 percent this year down from a career rate of 76.2 percent. His struggles of 2009 centered around a slightly lower K/BB, a drop in ground balls and a bit of bad luck. All things we should expect to regress in a new season and be fine. His age is a bit more concerning that things will start to age, but it’s better to expect the Oswalt of 2008 over the one we saw last season.

Shields, on the other hand, does have some bigger concerns. His makeup was built on elite K/BB numbers due to a very low walk rate. His ground ball rate was never at Oswalt’s level and he had to have a good K/BB number to make up for it. So far he has with 5.11 in 2007 and 4.00 in 2008 and while his 3.21 in 2009 was very good, it’s not the level he has shown.

His pitching approach has changed over the past few years with fewer fastballs, which is his worst pitch with a career run value of -0.66 per 100 pitches. At the same time, he has also dropped his number of change-ups, which is by far his best pitch. Going from 30 percent of his pitches down to 23 percent in 2009. It’s unclear why he is making this change, but the pitch has always maintained run values over 1 per 100 thrown.

Much like Cole Hamels, he must rely on his change-up and get back to throwing it 30 percent of the time. If he can return to 30 percent of his pitches being change-ups then he can start to regain the form we saw in 2008 and 2009, but if not he will be the solid pitcher of 2009 without ace level stuff.

While an older option, Oswalt probably makes the more solid and reliable selection in 2010. He didn’t change anything and is dealing with more luck-based changes. His defense is no where as good as Tampa’s, but solid enough for fantasy purposes. If Shields returns to his pitching arsenal that was so successful in the two years before 2009 then he could match Oswalt, but that has more to do with what he does and not luck.

Print Friendly
 Share on Facebook0Tweet about this on Twitter0Share on Google+0Share on Reddit0Email this to someone
« Previous: Should Carlos Gomez hit more ground balls?
Next: What determines batters’ selectivity levels? »


  1. JB (the original) said...

    Two things that worry me about Oswalt (in a fantasy way), and that’s propensity for injury, and pitching (theoretically) half his games in hitter friendly Minute Maid Park.

  2. Troy Patterson said...

    I’m not sure what the change was, but since 2006 Minute Maid Park has played as a nuetral or pitchers park according to baseball reference park factors.

  3. James said...

    Troy, control pitchers, like Shields, can have lower BABIP and HR/FB than league average. Why is Shields going to lose his incredible control at the age of 28? Saber has always been unable to predict breakouts, which is why you are being rigid; however, taking out bad years is a reasonable way to attempt to solve this problem.

    Am I 100% sure he will return, no. However, it’s reasonable to expect a control pitcher to return to his two best years when we realize his rookie year was his rookie year and last year was a tough situational year (coming off a World Series run).

    This is over, but I hope you won’t be a slave to career numbers when that career is only 3.5 years old. Pitchers can develop and thus those career numbers should be taken with a grain of salt – we aren’t talking about a 10 year pro that has multiple years of statistics. Good luck this year!

  4. James said...

    1. Shields is in his prime and is 28
    2. His HR/FB and BABIP were high last year
    3. He pretty much fell apart in the 2nd half, which probably could be explained by pitching in OCT and bad luck. You neglect to mention that his K/BB rate pre-allstar break was fine. 

    Oswalt is almost 5 years older, had a back injury last year, and is probably entering the downside of his career. In my opinion, this is a horrible comparison and conclusion – there is no reason why Shields will not return to form, but I can point to age and injury as legitimate reasons to avoid Oswalt.

  5. Troy Patterson said...

    So I can’t predict a breakout, but you can by removing data?

    I never said he would lose his control, but that he would be the pitcher that CHONE is projecting at an ERA of 3.91 and not the 3.56 ERA we saw in 2008.

    I agree a control pitcher can have “some” control on BABIP as shown here:

    But he’s had 774 IP to show that and his average in an extremely large sample is .306.  You can’t cherry pick the numbers you like and remove those you don’t.

    On the other hand I have seen no data to show pitchers have any control over HR/FB.

  6. James said...

    Unable should have read difficulty in predicting breakouts of young players by simply using career numbers – sorry. The reason is you don’t have enough data to be a slave to career numbers. For example, your extremely large sample has two years of extremely high BABIP, which taint the sample.

    Yes, HR/FB control, is much less than BABIP and pretty much nothing, but pitchers can have lower than league average.

    PS – you remove data all the time if you think a player’s season was lucky or unlucky. Shields had two unlucky years, and removing those numbers isn’t cherry picking if we know he was unlucky.

  7. Troy Patterson said...

    1 is a fair argument, but something to remember is pitchers do not have “primes” the same as hitters and are tough to predict in that regard.

    On number 2 you say his HR/FB was high, but I see 11.2% which is right at his career rate of 11.1%.  BABIP was .317 and career is .306.  One is right on and the other is a small step off.

    The last point is confusing.  If he “fell apart” in the second half why would you trust him to throw a solid 2010 for a full year.  His walks were up in the second half, but he still had a K/BB of 2.96.

    It depends on what you mean by return to form.  I think he will return to form, which is his career ERA of 4.01 not the spectacular 3.56 in 2008 when he had a xFIP of 3.82.

  8. Troy Patterson said...

    Well I use all the projections I can find as well.  I think CHONE does a fairly good job of using MLEs to add in his years in the minors to show what he is.

    Lower than league average HR/FB is a factor of park and league not the pitcher.  Just look at the pitching staffs in the NL West.  They all have lower than league average HR/FB, but not because they have any control.  It’s pitching >50% of their games in Petco, Dodger Stadium and San Fran.

    I never remove data from an analysis unless you can prove an injury or external influence.  Your removing “unlucky” years, but assuming all remaining years are “normal”.  I’m assuming those years are “lucky” and then they should be averaged with the “unlucky”.

  9. James said...

    Projection systems are great, but they don’t understand certain circumstances that we as humans must take in account when deciding value.

    Yes, that is true, but Shields having a high BABIP would influence his HR/FB rate, and thus, your average is off. Thus, there is a reason to expect his HR/FB rate to be lower.

    Well, I have reasons to assume my 430 IP sample is better than your 344.1 IP sample, and you haven’t provided one reason why his ‘07 and ‘08 season were frauds. And I have given you plenty of reason why your sample is bad. Thus, mine is better.

  10. Troy Patterson said...

    How does BABIP effect HR/FB?  BABIP removes homers…That sentence made no sense to me please explain again.

    I never said ‘07 was a fraud.  I said he is a 3.90 ERA pitcher and that would mean I think his ‘06, ‘07 and ‘09 are his real or “unlucky” numbers and ‘08 was an example of “lucky”, but I still include ‘08 in my calculations of average.

    Also saying your “reasons” that your sample is better doesn’t make the other 344.1 IP disappear.

  11. Newton said...


    Can you explain why you think a pitcher can control the ratio of flyballs that turn into homeruns?  I would be very interested to understand how you can conclude that a pitcher has any control on what happens at impact.

    A pitcher can control the desired location, movement, and speed of a pitch.  He cannot affect the moment of impact.  That is all dependent upon the batter’s ability.  This is why HR/FB% is easily measured among hitters, and is regarded as purely chance towards pitchers (except taking into account park factors).

    A pitcher cannot physically somehow “deaden” or “enhance” impact.  All he can do is put the ball in a spot to make it harder for a batter to create solid contact(whether via reaction time, angular displacement, etc). 

    For example: A ball low in the zone has a higher chance of being a ground ball because a batter cannot react quickly enough to achieve a squared face with the bat.  A very high pitch (the elevated 3rd strike) can do the opposite, causing most batters to overcompensate and cut under the pitch.  However, in both of these scenarios, if a batter reacts properly and exerts force on the baseball, he can drive it out of the park. Nothing the pitcher has done can change this.

    Once contact is made, the pitcher is removed from the equation.  He has no power to somehow make his flyballs travel less further than any other pitcher without violating Newton’s 3rd law. 

    If a pitcher could somehow make his pitches travel less on impact, he would have to somehow be able to exert MORE force on the bat than the bat to the ball during impact.  That’s physically impossible.

  12. James said...

    - Prime as in he is young and should be able to rebound as opposed to a pitcher who is 33 and is less likely to rebound. I think that is fair, no?

    - 06: 13.7%
      07: 11.1%
      08: 9.1%
      09: 11.1%
    Shield’s was trending downward until last year, which as we know was unlucky for him, why can’t we expect 9.1% in a normal year? Career 11.1% encompasses his short career, which wasn’t pretty two out of four years. Thus, the results are skewed. 

    - Because he didn’t throw in OCT last year and isn’t suffering from any fatigue as a result. How is that confusing?

  13. James said...

    Sorry I wasn’t clear, but I have time constraints – I was generalizing.

    Jeez – you are rigid. I never said it disappears, but I would expect a reasonable person to say, “OK – I can accept ignoring those stats because they were unlucky, you might have something here, and we could put those bad years to the side.” However, I cannot ignore Oswalt’s age and back injury, which are undeniable, and the reason I thought this comparison was off. Thus, I would expect you to give Shields the benefit of the doubt over someone whose “problem areas” are much more cut and dry.

    I enjoyed the conversation – enjoy the season!

  14. James said...

    I forgot to mention BABIP – .332 .292 .292 and .317, which average to .306, but two of those four years were pretty high right? As a result, you can’t use his career numbers as a barometer. I guess this is why you didn’t mention these numbers.

  15. James said...

    Newton, Troy and I were jumping around two different points on HR/FB rate (league average and a pitchers career rate). I stated before the pitchers have little effect on HR/FB rates. The point I was making before was more of a generalization that pitchers lose their location/stuff when things aren’t going their way. For example, they leave the ball up instead of keeping the ball on the ground, which is one way to control HR/FB rate. In Shield’s bad BABIP years, his GB% was lower, which would support this theory on him.

    Yes, it’s not science and I wouldn’t make this my best point, but it’s a valid conclusion.

  16. Troy Patterson said...

    That is not correct.  Leaving the ball up in the zone will raise the flyball rate, but not the HR/FB rate.

    “I never said it disappears, but I would expect a reasonable person to say, “OK – I can accept ignoring “


  17. Newton said...

    “For example, they leave the ball up instead of keeping the ball on the ground, which is one way to control HR/FB rate”

    It’s a way to control their flyball rate, not their HR/FB rate.  I think that’s the difference you’re missing.  A pitcher can only control the number of homeruns he gives up by keeping the ball on the ground.  The location, speed, and movement of the pitch cannot exert any extra force on the bat to somehow limit possible distance traveled.

    It may make it less likely for a pitch to be a flyball, but it cannot make it less likely for a flyball to be a homerun.  The amount of force exerted on the ball is only controlled by the batter’s ability to identify, analyze and react to a pitch.  The pitcher just needs to hope he provided enough deception and difficulty to make this harder on the average player.

    That’s the nuance you need to remember with what HR/FB actually represents.  Even if the pitcher throws the textbook “perfect pitch”, a batter who does his job properly can still react and analyze it properly and tattoo it as much as he could a ball at rest on a tee.

  18. Newton said...


    I think I missed your repsonse at 2:29, but my concern is when you said:

    “Thus, there is a reason to expect his HR/FB rate to be lower.”

    His HR/FB rate shouldn’t change dramatically from year to year.  If his FB% changes, his HR allowed should change, but the HR/FB rate should always be about the same whether he gives up 50, 100, or 200 flyballs.  That’s why if their is a big jump in either direction, its a cause for scrutiny.

    If that’s what you’re trying to say, that’s fine.  But you made it originally sound like he has control over how his HR/FB can change, which is what I was hoping to clarify.

    I just get frustrated when the world seems to think the 3rd law doesn’t apply to baseball or in my neck of the woods, that Dice-K magically prevents homeruns by walking batters. Someone actually said that to me once!


  19. James said...

    Troy – you knew what I meant. Just because I ignore something doesn’t mean I don’t know about it. Disapear would mean it never there, which are two different things. Would discount be easier for you to understand? Disapear is spelled disappear by the way. I can be rude as well.

    In addition, if Shields is throwing beach balls since his lucky isn’t where it should be, his HR/FB rate will rise.

  20. Troy Patterson said...

    James, no one is being rude.  I am defending my article and questioning your methods.  I would like examples of why I should ignore or weight different data sets the way you are suggesting.

    I understand you are trying to say he was worn out from a long 2008 season, but I need evidence that his only response to fatigue was an increase in BABIP.  His velocity was at career levels and he was even getting a K/BB of 4.71 in Sept/Oct.

  21. James said...

    What about his BB/9? That went up in the 2nd half.

    Plus, what evidence do you have that Oswalt, an older option as you mention, has healed from a back injury that sidelined him for the last two months? In my opinion, there is more to encouraged about with Shields than Oswalt because his back is a COMPLETE unknown. Is that fair?

  22. Troy Patterson said...

    HR/FB and BABIP do not work that way.  They are not skills and you can’t assume he’ll repeat his best numbers.

    The hitters he faces control these numbers and you have to assume he’ll be near the league averages.

  23. Troy Patterson said...

    That is fair to say that the injury is a concern, but so far this spring he looks fine.  I would like to see a velocity reading to give a final say, but with ADP of 126 for Shields and 155 for Oswalt you get better value for Oswalt.

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>