Fun With Fastballs: The Value Of Velocity

i think that varying speeds definetly is a huge reason why pitchers are sucessful but i think if you vary speeds at a higher velocity as opposed to varying speeds at a lower velocity… the higher velocity will be more sucessful, because as a hitter you have to read the pitches that much quicker at a higher velocity

Because I would like to see the simple bivariate correlation between pitchers’ average 4SFB velocity and their WAR. That would give us a much clearer idea.

Expectedly, the pitchers with the highest average velocity have the highest WAR, and then there is an expected dip. But what is odd to me is that pitchers in the 91-92 mph range do better than pitchers at 92-93. Why would that be?

Velocity is postively correlated with WAR. But what about control? The D-train can throw hard but walks too many. If you are looking long-term, you might also consider whether velocity and durability are inversely related to one another. To take exteme cases, Joel Zumaya and Mark Prior against Tim Wakefield and Jamie Moyer. Also, I would suspect that HR% will turn out to be important. I think pitchers should be viewed as a whole and not on one dimension. Finally, you can’t measure heart under pressure. God has blessed Mariano and David Robertson specially. Conversely, Pat Tabler used to be the most dangerous hitter with the bases loaded. I think the ability to perform under pressure is sometimes undervalued. For relievers, the inherited runners statistic can give a clue.

What was the variance per tier?

that the three most important things in pitching are velocity, placement, and movement, and not necessarily in that order.
Even the hard throwers have to have some kind of movement (if it goes in straight, it will go out straight) and placement, while pitchers like John have been successful w/o much velocity.
I did a little research on the calculation of WAR for pitchers to see if there was an inherent bias in the formula in favor of hard throwers. According to Fangraphs and Big League Stew, WAR for pitchers is based on FIP, adjusted for things like park factors.
FIP is a strange formula. In an attempt to eliminate the possible impact of fielding, it excludes all hits other than home runs. This is sort of like analyzing someone’s driving record but ignoring all multicar accidents because they might be the other drivers’ fault. The formula assigns certain weights to walks and strikeouts and then introduces a normalization constant so that the formula will correctly predict the ERA of the average pitcher in the league.
Given that hits (except for HRs) are not considered, the FIP formula is built around walks and strikeouts, and rewards a high strikeout-to-walk ratio at the exclusion of almost everything else. Pitchers like Tommy John, who pitch to contact successfully, are penalized because they win w/o striking out many people. For example, John’s lifetime FIP by my calculation is 3.88, but his lifetime ERA is 3.34.
To the extent that John’s WAR is based on his FIP, it is understated, and if most of the low-speed pitchers are pitch-to-contact types, the WAR for the group may be understated as well.

Split into relatively even groups, based on velocity.
FBv 93.1+ MPH – 15 pitchers, average WAR = 13.6
FBv 92.0-92.7 – 14 pitchers, average WAR = 10.6
FBv 91.0-91.8 – 17 pitchers, average WAR = 10.1
FBv 89.7-90.8 – 14 pitchers, average WAR = 7.1
FBv 84.3-89.3 – 13 pitchers, average WAR = 7.0

So 93+ amounts to roughly 1 fWAR per season in additional value compared to the 91-92.7 group. The 84.3-90.8 group loses roughly 1 fWAR per season in value compared to the 91-92.7 groups. So the difference between 93+ and below 91 is about 2 fWAR per year in value.