Pitcher Aging Curves: Maintaining Velocity

Bill Petti published the first two parts of a series on pitcher aging. Bill’s first article focused on pitchers, in general, and the second was on the difference between starters and relievers. For the third installment, I’ll look at aging patterns for pitchers who maintain a relatively constant velocity from year-to-year.

From the previous articles, the average pitcher loses about 4 mph from their fastball from ages 21 to 38. In essence, most pitchers’ stats degrade as their fastball speed drops. Using the same methodology, I wanted to know how pitchers age when they don’t lose velocity on their pitches.

I examined pitchers who had their fastball speed within +/- 0.5 mph, compared with the previous season. I ran into a problem in that 60% of the samples saw a decline in pitch speed — while 40% stayed the same or increased. The initial curves mimicked the ones of the general population. To get a sample that had an equal number of samples above and below zero, I limited the number of samples below zero to a random sample equal to the number above zero. Those results were better.

Even with this change, though, I have two biases in the data. First, there were more declining values for older pitchers than for younger ones. This can be seen in the up and down of the velocity curve. The second problem was that there are larger velocity drops than velocity increases. For example, there are more -0.5 mph declines than +0.5 mph increases. When I tried to correct for each of these problems, the data sample size became too simple.

With that in mind, here are the fastball aging curves for pitchers who maintained a constant velocity:


— From ages 21 to 38, a pitcher’s FIP will basically remain constant with the pitcher’s peak years being from 25 to 30 years of age.
— Pitchers improve their walk rates until age 24 and then the number stays relatively constant.
— BABIP goes up steadily during a pitcher’s career, possibly because hitters become more familiar with pitchers.

Most of these curves are somewhat similar to the curves for all pitchers. The main difference is in the amount of change in values. Here are some select curves when the velocity is kept constant, compared with the curves for all pitchers.

(“= Velo” is for data with constant velocity)


— Even though the problems stated earlier show a decline in velocity, it’s a total of only -0.3 mph from ages 21 to 38. The entire population loses 3.7 mph.
— The constant velocity FIP initially drops and eventually goes back up to the age-21 value. For the general population, their FIP increases by 1.7.

Roy Halladay has been able to maintain a relatively consistent fastball speed over several seasons. Since 2002, his fastball speed has ranged between 90.3 mph and 92.7 mph with an average value of 91.9 mph. Over that same time, his FIP has a range between 2.20 and 3.79 and averaged out at 3.11. Here’s how is how Halladay’s K/9, BB/9 and velocity has aged when compared the average of players with constant velocities:

– While Halladay has been able to main a relatively constant fastball velocity over the years, it has bounced around a little from year-to-year.
– Whenever Halladay’s velocity moves up, his K/9 goes up and his BB/9 goes down. Likewise, when his velocity drops, his K/9 and BB/9 increase. Crazy how that works, right?

As long as a pitcher is able to maintain a certain velocity, like Halladay has, the player can generally pitch with the same results year after year. It’s only when the pitcher begins to lose velocity that he sees his stats degrade at a higher rate. Sure, vvelocity isn’t everything with a pitcher — but it’s important. Pitchers who maintain their velocity don’t really “age” — they stay the same.

Print This Post

Jeff writes for FanGraphs, The Hardball Times and Royals Review, as well as his own website, Baseball Heat Maps with his brother Darrell. In tandem with Bill Petti, he won the 2013 SABR Analytics Research Award for Contemporary Analysis. Follow him on Twitter @jeffwzimmerman.

11 Responses to “Pitcher Aging Curves: Maintaining Velocity”

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

    Jamie Moyer?

    Vote -1 Vote +1

  2. Johnhavok says:

    I have a Jamie Moyer baseball card from 1885.

    Vote -1 Vote +1

    • Antonio Bananas says:

      Think at how much higher his counting stats would be if he didn’t serve in WWI, WWII, Vietnam, and Korea.

      +12 Vote -1 Vote +1

      • Choo says:

        And just try to imagine how lethal Jaime Moyer’s fastball was before God took his rib to make Eve.

        +7 Vote -1 Vote +1

      • MC says:

        Jesus’s whole water to wine trick was his way to excuse striking out when facing Moyer on the sandlot.

        Vote -1 Vote +1

  3. James Piette says:

    Great stuff, but I worry about all the noise involved. In the fourth installment, I think it would be very informative if you published some of the sample standard deviations, both overall and by age. That way, we can really determine if any of these things are deviating by a meaningful amount.

    Vote -1 Vote +1

  4. Stats says:

    The legends on your graphs are almost impossible to follow. In some cases, like the last graph, the example line for multiple curves is exactly the same. I recommend you follow the advice of countless visualization masters and stick the curve labels next to each curve and eliminate the legend entirely. If not, at least improve the size of the example lines so that we can see what’s going on.

    Vote -1 Vote +1

  5. Frag says:

    Great stuff!

    Vote -1 Vote +1

  6. Excellent work indeed, it seems the high caliber baseball analysis, as per usual, attracts the fewest comments.

    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>