Arms in the Pennant Race Running out of Fuel?

Good starting pitching is always important, but it seems to take on added significance as the pennant race heats up. And often it seems that the teams with the freshest, strongest arms are the teams able to emerge out the grind of a 162 game season into the playoffs.

This year fatigue could be especially important to monitor since there are several pitchers pitching deeper into the year and/or logging more innings than ever before whose performance could have big implications on the pennant race.

To examine whether any of these pitchers are showing signs of tiring, we’ll use an imperfect, but I believe reasonable, measure of pitcher fatigue– the change in fastball velocity from start to start.

I retrieved all data from and only considered pitches classified as four-seam fastballs.

In the American League West, the Rangers enter play tonight holding a 3 game lead over the Angels, but after surrendering an average of more than 8 runs per game over their past 6 contests, there are legitimate questions as to how well the Rangers’ pitching will hold up over the rest of the season. Adding to those concerns is the fact that two important members of the Rangers’ rotation, Matt Harrison and Alexi Ogando, have already surpassed their innings total from last season. In fact, with 151 innings, Ogando has already doubled the 72 innings he threw last year, and Harrison needs only 17 more innings to double his 2010 total of 86 innings.

Considering the dramatic increase in innings this season, both pitchers are candidates to suffer from fatigue over the final month of the season.

Let’s take a look at Ogando first.

Given the the fact that he had never started in the big leagues before this season (not to mention his recent rough outing against the Red Sox), Ogando seemed to be a likely candidate to have lost some velocity over the course of the season. Surprisingly, we see that the exact opposite has happened! Ogando started the year throwing smoke, and his fastball has only gotten faster as the season has progressed.

Despite two recent outings in which he got hit around, Ogando’s statistics don’t suggest there’s has been a change in the quality of his stuff. From April through July, Ogando had a (Strikeout – Walk) per plate appearance mark of .12; in August that mark has fallen to .11, but considering that he faced both the Red Sox and Tigers over that stretch, I’m not sure whether we should look at that as a drop in performance.

Given that his velocity has improved, and that there are no signs of a sizable drop in performance, there don’t appear to be any warning signs that Ogando’s performance will deteriorate over the final month of the season.

The picture isn’t quite as encouraging for Matt Harrison.

Harrison’s velocity has fluctuated from start to start, but for much of the season his velocity hovered right around 93 mph, since the mid-way point of the season, his velocity has begun to decrease, with his most reacent start against the Red Sox being especially concerning (although he threw few four-seamers in that particular outing).

The Rangers also seem concerned, since they have elected to skip his next start. Harrison told’s Louie Horvath: “It’s in my best interest to skip a start. I haven’t rested since the All-Star break. [The Rangers] know it has been a while since I’ve thrown this many innings, and they’d rather have me good and ready to go for the last five starts than to keep pounding away and making it worse.”

The silver lining is that Harrison doesn’t seem to have regressed statistically. His (strikeout – walk) per plate appearance mark entering August was .08, and in four August starts he’s increased the rate to .10.

Give the Rangers credit for being mindful of the situation, but I don’t think there is any reason for Ranger fans to panic.

Outside of the AL West Race, Bartolo Colon, Josh Collmenter, and Ryan Vogelsong all find themselves entering uncharted waters this August, and each figures to play a role in shaping the course of the pennant races.

Although Collmenter is currently 28 innings short of last year’s total. The rookie has never pitched in September as a professional, and because he started the year in the bullpen, he has already appeared in more games this season than last.

Fortunately for the Diamondbacks, there aren’t any signs that they should be especially concerned.

The trendline does slope down at the end of the graph, but some of that can be attributed to Collmenter’s 16th start, which seems to be a bit of an outlier, but even the bottom of the slope only represents a drop of about a quarter of a mph. Nor is there any statistical indication that Collmenter’s performance has begun to tail off. Before August Collmenter had a (strikeout – walk) per plate appearance rate of .11, and in August that rate has increased to more than .13.

If the Diamondbacks are able to clinch the AL West before the end of the season, it may be a good idea to ease off Collmenter a bit down the stretch, but he appears to be strong entering the playoff push.

Entering tonight’s action five games back, it will be difficult for the defending world champs to catch the Diamondbacks without Ryan Vogelsong continuing his story-book season.

Unfortunately for the Giants, Vogelsong is one of the more concerning case we’ve looked at.

The y-axis makes the change in velocity look more extreme than it is, but Vogelsong has still lost close to a mph since the middle of the year. Considering that the 153 innings Vogelsong has thrown this year is a significant increase from the 95 he threw last year, there is some cause for concern that Vogelsong may be tiring. You can always get yourself into trouble by cherry-picking examples, but in his last two starts, both against the Astros, Vogelsong has struck out only 7 and walked 8.

The loss of three quarters of a mile an hour certainly isn’t cause for panic by its self, but combined with back-to-back disappointing performances and the increased workload, the decrease in velocity is a worrisome sign.

Finally, and arguably the most improbable story of all the improbable stories on this list, Bartolo Colon.

After throwing fewer than 75 innings in 2008, fewer than 90 innings in 2009, and not pitching at all in 2010, Bartolo Colon is back. Some have attributed his comeback to a procedure performed last year in the Dominican in which stem cells were reportedly injected into his arm. Whatever the reason, Colon has pitched exceptionally well this season, sporting a SIERA 3.37. Still, given how few innings he has thrown over the past four years, many in the Bronx are concerned about how he’ll hold up into September and October.

The good news for Yankee fans is that Colon’s velocity has been pretty stable.

I think the prognosis for Colon is similar to that of Josh Collmenter. With no real signs of fatigue, there’s no reason for the Yankees to shy away from Colon. That being said, there has been talk that the Yankees may go to a 6-man rotation for at least a few weeks in September, and Colon could be a beneficiary of that move.

Print This Post

15 Responses to “Arms in the Pennant Race Running out of Fuel?”

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

    Rangers are in trouble. Their staff looks like they are toasted.

    Vote -1 Vote +1

    • Adam D says:

      It doesn’t help that they actually are toasted by the 100+ degree temps in Arlington every night this summer. I expect they’ll feel a little fresher once the temps drop back below 90.

      Vote -1 Vote +1

  2. AndyS says:

    This is a decent study, but…are curves like this the exception or the norm?

    Vote -1 Vote +1

  3. Eminor3rd says:

    *Pennant. I generally don’t point thing like that out, but it’s in the title.

    Vote -1 Vote +1

  4. Bryan Simpson says:

    Good break down of these pitchers. Now that I have your source,, I’m ready to waste every free moment today finding the stamina of the rest of the pennant staffs. Thanks for the source.

    Vote -1 Vote +1

  5. Shawn says:

    I am a little concerned about the braves as well. With Hanson likely out for a while, minor is going to be a regular cog in the rotation. I do not think that he is going to go highly over his innings total from a previous year, but I do not know if he has thrown as many high stress pitches as he has in the past. Along with that, Vizcaino is going to be relied upon heavily, and while he is healthy, he is just recovered from a partial ligament tear. Again, it is not really the innings count, just the fact that rookies arms being subjected to throwing high stress pitches in very important scenerios for hopefully the next two months.

    Vote -1 Vote +1

  6. gdc says:

    When you get to showing tenths of a mph, is there a pitching park effect, maybe thinness of air, where in the delivery it is registered, different radar equipment, or even temperature? Then being pulled down by one “slow” start might just be an infrequent visit to the park with the “slow gun”, as opposed to regular visits through the year.

    Vote -1 Vote +1

  7. AndyS says:

    Wow I just noticed how arbitrary some of these “best fit curves” are…

    Vote -1 Vote +1

    • Barkey Walker says:

      Yeah, on some of them, remove one point and you get a whole new curve. They should be removed or get standard error bars to show how BS (I mean useful) they are.

      Vote -1 Vote +1

  8. Andrew says:

    Seriously, I know you put some work into making these charts, and I’m sure you worked hard to fit your curves with acceptable R/R^2 values, but, as stated by another commentator, with data set deltas which are, apparently (I say apparently, because I’m too lazy to run the numbers) within normal regressed parameters, I’m not sure the data is actually telling us anything, except that pitching velocity is very noisy, even over the course of a season. Now, if, for example, I saw pitches consistently between 86-88 for 4/5 starts and then a sudden drop to say 83-84 (well outside standard deviation) then perhaps we could speculate/attribute such a change to a change in the pitcher. I just am not sure the data tells us much.

    Vote -1 Vote +1

  9. TwoPointsMakeALine says:

    Hmm. Add another term to your polynomial fits, and you can create complicated in-season dramatic stories to go with apparent speed fluctuations.

    Has someone done a study to estimate error bars on every stadium’s speed gun? Then you could put error bars on each point.

    Vote -1 Vote +1

  10. garik16 says:

    Umm, you are correcting for seasonal changes right? A parabolic velocity curve is kind of norma.

    Vote -1 Vote +1