With the 2021 regular MLB season finished, I know
everyone my loyal readers my six friends you(?) were looking forward to my yearly article on players with more walks than strikeouts. While I have updated my dashboard with 2021 data, that article is going to have to wait, but will almost certainly be out sooner than the one for the 2020 season. I was once again intrigued by a query posed by Grey, of Razzball.com, in his article Top 20 3rd Basemen for 2021 Fantasy Baseball, a Recap. In his recap of Manny Machado’s season, Grey provided an interesting tidbit I will quote below: “Finally, and I don’t know how to check this, but he must have the slowest Sprint Speed with that many steals ever. Sprint Speed is “hustle” in his case, excuse me, I meant to say “head case.” As I wrote last time, “Normally when writers pose these type of questions, I think about answering them, but never do“. If that last article and this one are a current indication of things, then the previous statement is not longer true and luckily for Grey I had some time on my hands, curiosity, and the ability to answer this question. Was he right? Have all my years of researching baseball players paid off? Let’s find out.
I would argue that the most difficult part of this exercise was out of the way, writing a question worth asking. The question defined the parameters as: 1) A player finishing an MLB regular season with at least 12 steals and 2) it needing to be in a season we have sprint speed data for. The good news is I knew how to acquire both of these things. The bad news was I did not know how to get them in one single source.
I first went to baseballsavant.com, and visited their leaderboard for sprint speed. In the past, I used to have to manually copy, paste, and format data from their leaderboards, but thankfully they added a handy button to download a csv file of the data. I did this for the seasons 2015 to 2021 as 2015 was the earliest season they began recording sprint speed. Therefore, I will only be looking at the past 7 season of data. Next, I went over to fangraphs.com, my go to for most MLB data, and downloaded the previous 7 years worth of player data from their leaderboards. As far as I can remember, Fangraphs has made it easy to download their data by providing an Export Data button on their tables.
When you download data from either site, they provide a player ID field that is the unique identifier associated with each player for that website. Player IDs are the cleanest way to join similar datasets for baseball players as names can 1) be different for a single player across different sites (example, Ronald Acuna vs Ronald Acuna Jr. and 2) some players have the exact same name that would mess up the joins (one current example is Will Smith, C, LAD and Will Smith, RP, ATL). Fortunately, there is one brave soul who dares to map all the MLB players to all their players IDs across the various websites. His name is Tanner Bell and he runs the website Smartfantasybaseball.com and also coauthored the web book The Process with Jeff Zimmerman. I have read it and it provides some great insights for fantasy baseball players. Anyway, Tanner provides a free downloadable file on his website on that makes these type of projects super easy. I usually download a copy of it a few times a year to get updated player information.
Anyway, armed with the sprint speed CSVs, the steals CSV, and the player ID mapping, I moved all the data into a single excel file, with each piece of data on separate sheets. Starting with the Fangraphs steals data as the main source, I used the functions index and match (equivalent of a vlookup) to create a column to determine the MLB ID associated with each player. From there, I created a nested IF function to check the season for the particular player and then look up their average sprint speed for that year. This way, I had one master average sprint speed column to make comparisons easy.
With my dataset finally complete, I was ready to answer the original question. To figure this out as fast as possible, I decided to use two filters on the datasets based on Machado’s 2021 data: 1) stolen bases for a single season must be greater than or equal to 12 and 2) an average sprint speed must be less than or equal to 26.4 ft./s. This made the next part very easy. After applying the filters, I saw there were a total of 13 records that met my criteria. You can see an image of the records below:
As you can see, while Grey’s initial hunch was correct, Machado in 2021 is only the 13th slowest player to steal 12 bases in the past 7 seasons. While there are 12 other names below Machado’s, there are actually only 8 other players as some players have done this more than once, including Machado! Machado actually stole 14 bases in 2018 with a slower average sprint speed. That was the year he played for the Orioles and was traded to the Dodgers. I was wondering if he was not trying with the Orioles and that is why his sprint speed was lower as they were not a competitive team (If someone knows how to look up sprint speed by team for a player in a single year, please let me know).
It is only a difference of 0.9 ft./s that separates 2021 Manny Machado from 2018 Alex Gordon, the slowest player from 2015 to 2021 to steal at least 12 bases in a single season. It was Gordon’s 3rd to last season in the majors and the last time he had double digit stolen bases in his career. In Gordon’s defense, he was 34 that year. That said, it makes it more impressive that Brandon Phillips, 3rd and 7th on this list, stole more bases and had faster sprint speeds in his age 34 and 35 seasons, respectively, than Gordon. The 2015 season was Matt Kemp‘s only season with the Padres and his final season where he had more than 1 stolen base.
Viewing the Dataset
You can see a link to my complete final dataset here. I looked at the slowest average sprint speed for 12 steals, but using this file, you should be able to look at it for any number of stolen bases. Note, Tanner’s player ID map is not perfect at the moment as it did not include some 2021 rookies. If I wanted to make the dataset 100% complete, I would manually look up the data and put it in myself. That said, I made the assumption the 2021 rookies were faster than Machado and this should not have a real impact on the analysis.
Thanks for reading and have a great day!