In 1968, Bob Gibson had one of the best single-season performances by a pitcher in baseball history. His 1.12 era was the lowest in the modern era. His excellence in a season of pitching excellence (22 other pitchers had a sub 2.00 era that season) resulted in baseball lowering the mound.
Gibson was not only dominant. He was intimidating. He owned the inside half of the plate. Batters, reportedly, feared to step into the box against a heater that he could locate with high velocity and command at their knees or chin.
Ask anyone from that era. Gibson had one of the best fastballs ever.
I can hear my father-in-law now, God rest his soul, talking Cardinals with me. I would quote some sabermetric gobbledygoop about the modern game and he would throw the unprovable past right back at me, high and tight! “You think modern pitchers are good and throw hard? Well, I watched Gibson! If (insert modern inferior player here) can hit 95 mph, then Gibson was over 100! (Mumbling…) Can’t believe my daughter married this guy…”
I can’t either, Erwin.
It’s a debate that has continued for generations because it’s the best kind of debate. With absolutely no empirical evidence whatsoever, anyone can claim anything they want and they get to be right!
On Wednesday, I went to Twitter and asked the Cardinals’ hive-mind “how hard was Bob Gibson’s fastball”? Vote to see the results:
How hard was Bob Gibson's fastball?— Jason Hill (@JPHill_Cards) October 7, 2020
I'll be spending the day calculating the apx. mph of Gibson's fastball for an article that will run Thurs at Viva El Birdos. Before that, I want to crowdsource Cardinals' Twitter & get your guesses!
Vote & reply! (Retweet so others can, too.)
With over 700 votes cast, the majority of fans believe Gibson threw somewhere between 94-97 mph. Nearly 60% of voters agree that Gibson at least threw 94 or harder.
Who’s right? Can we know?
We sort of can.
Using pseudo-science and quasi-math, I think I can estimate how hard Bob Gibson threw his fastball. Specifically, I’m going to focus on Game One of the 1968 Series, when Gibson had (arguably) his most famous moment. Gibson shut down the Tigers, recording a record 17 strikeouts in a complete game, five-hit shutout.
Notice I say “pseudo-science” and “quasi-math”. My method isn’t going to be terribly precise. But it will be fun and I think we can get a result that will leave us satisfied.
Here’s how I’m going to do it. While we don’t have pitch velocity data from 1968, we do have precise velocity information in 2020. Statcast can tell us fastball pitch speed down to the tenth of an mph. It conveniently provides high-quality video of these same pitches.
Using simple video editing we can find the moment that a fastball leaves a pitcher’s hand and disappears into the catcher’s closed glove. Between those two moments is a unit of measurement that you won’t find on the back of any baseball card or advanced sabermetric page. It’s a video metric: frames per second.
Right there we have velocity and time over a relatively similar distance. That’s just enough information to establish a baseline ratio of video frames per mile per hour using 2020 data. We can then apply that same ratio to any other video of any era, so long as the frame rate per second is the same (and there’s not a gigantic difference in player height).
If we want to know how hard Bob Gibson threw his 4-seamer in Game One of the ’68 series, all we have to do is compare the video of his fastballs that day to video of modern fastballs, add in some fuzzy multiplication, and shake vigorously. What could go wrong?
Several things. Video is an imprecise beast. A pitching motion viewed frame by frame includes significant motion blur that’s not evident at full speed. Finding the exact moment a pitch leaves the hand or the glove closes requires some guesswork. We can somewhat mitigate the subjective nature of this by using multiple samples to set the ratio of frames to miles per hour. We will use three different modern Cardinal hurlers to do this:
(Don't mind me, I just need a place to host my videos for tomorrow's Bob Gibson article.)— Jason Hill (@JPHill_Cards) October 8, 2020
Here are the three samples I will use to set my ratio of frames to miles per hour.
1. Jordan Hicks, 2019 – 104 mph
2. Jack Flaherty, 2020 – 94.9 mph
3. Adam Wainwright, 2020 – 89.5 mph pic.twitter.com/Ya3c472m4H
As usual, those three provide a satisfying result. I counted 31 frames between Wainwright’s release point and Molina’s catch. Flaherty is at 29. Hicks is at 26. Now for the quasi-math. Between Wainwright and Flaherty, there is a difference of 2.8 mph (rounded up) per frame. (95 mph minus 89.5 mph, divided by 2). The difference between Flaherty and Hicks is 3.0 mph per frame. Hmm… that’s a little high. But, then it’s right back to 2.8 if I go between Wainwright and Hicks.
There’s our ratio and margin of error. We can set 2.8 mph per frame as our baseline ratio and 95 mph and 29 frames as the starting point. (We will end up with a +/- of .1-.2 in our multiplier.)
Now, let’s get to the fun part. Enter Bob Gibson!
If you want to (and you do want to), you can watch the entirety of the 1968 World Series Game 1 on YouTube. For our purposes, the video quality isn’t nearly as good as the modern-day starters but it is going to work well enough. I pulled three fastballs from the first few innings of Gibson’s outing and looped them like the ones above:
Video two: three fastballs from Gibson's '68 Series Game One masterpiece.— Jason Hill (@JPHill_Cards) October 8, 2020
1. Gibson - 29 frames
2. Gibson - 32 frames**
3. Gibson - 29 frames pic.twitter.com/Hep0vHWUM1
If we want to know exactly how hard some of Gibson’s fastballs were in the ’68 series, that’s a good start. But, three pitches from one video of one game doesn’t make much of a sample, even if that one game is one of the best of all time. Let’s expand our reach a little. I looked at five additional fastballs from throughout Gibson’s career. Unfortunately, only two of these fastballs were good enough (in terms of video quality) to couple with the ’68 samples, though all will be considered in our analysis:
Video three: two fastballs randomly sampled from Gibson's career.— Jason Hill (@JPHill_Cards) October 8, 2020
1. Gibson - 30 frames
2. Gibson - 31 frames pic.twitter.com/PoUDKf7EXk
Now we’re getting somewhere! We have multiple data points from multiple sources and a pretty consistent spread – 29 frames to 32 frames. I did the math when you weren’t looking and graphed the results so they look all scientific and official. Here are the frames per mph and estimated velocity data for Bob Gibson’s fastballs:
Fastball Velocity by Frames (2.8 mph per frame ratio)
|Gibson 68 1||95||29|
|Gibson 68 2||87||32|
|Gibson 68 3||95||29|
Bob Gibson’s average fastball velocity was 91.9 mph! Let it be so decreed!
Is that the end? The final answer? It’s probably a pretty good answer. But, there are some problems with the data that I think impact our bottom line.
First, there’s at least an outlier in our data set. One of Gibson’s fastballs in the ’68 series charted out at 32 frames – 87 mph. That’s the slowest of any of the pitches I looked at, including a bunch that didn’t end up in my sample. How do I reconcile this with the rest of the data?
There’s a good chance that this pitch is not a 4-seam fastball. It’s entirely possible – if not 100% certain – that Gibson regularly varied his fastball grips and velocities. Most pitchers of Gibson’s era did so. “Throw it as hard as you can all the time” is a very modern way to pitch. Maybe this pitch is a 2-seamer. Maybe he just took a little something off.
The videos themselves support this conclusion. The batters’ swings progress further into their follow through on the slower pitches (31-32 frames). That’s just smart pitching. Gibson speeds the hitter up and then cuts down on the gas. The batter is out in front. Swinging strike.
Second, there’s the aforementioned group of pitches that I looked at but could not include. At least three of Gibson’s fastballs scored between 26 and 27 frames – 100-104 mph. Wow, right?! Don’t get too excited.
These pitches were left out of my sample because they had 2-3 duplicate frames buried in them. This was expected and it is the result of the conversion process with modern video editors. When converting videos of different frame rates (for example, 25 fps to 29.7 fps), editing software will duplicate random frames to keep the timing right. Unfortunately, those duplicate frames make it impossible to determine what exactly happened on the pitch.
I just can’t trust these videos. But they do exist! That’s something to note.
Now we have outliers on both sides of our data set but clear explanations for both. I think that allows us to narrow our field and offer a revised final answer.
How fast was Bob Gibson’s Fastball?
Bob Gibson’s 4-seam fastball “sat” between 92-95 mph in the sample I considered. It’s likely he actively varied his grip or intended velocity, producing a high velocity range, measured at 87-95, with numerous indications that he regularly exceeded 95 mph.
In the 1968 World Series, one of the best single-game pitching performances of all-time, Gibson’s 4-seam fastball was routinely 95 mph. There is inconclusive evidence that Gibson touched the upper 90’s that day, perhaps even reaching triple digits.
Maybe I should do Lou Brock’s sprint speed next? I think so!
Let me know in the comments if you have a better way to approach this question! I hope only to spark conversation and creativity in others. I’m certain there are more precise ways to get to the answer, but I’ll leave those up to the collective! Have fun with it.