Last May, I unveiled a new statistic I’d been working on for a while called SBot, stolen base opportunities taken. I’d highly recommend reading that piece first before continuing with this one, as I will be building on the concepts presented within.
As I stated in the original piece, there are four components that lead a player to attempt a stolen base: speed, knowing when to run, getting on base, and intention. We have statistics to quantify the first three factors: Statcast sprint speed, stolen base success rate, and on-base percentage. SBot (pronounced very similarly to spot) is a rate stat that looks at how often a player decides to attempt a stolen base when given an opportunity, thus quantifying the player’s intention to steal.
In my May piece, SBot was originally calculated by taking attempted stolen bases (SB+CS) divided by times on base (PA*OBP). This worked as a simple base for the idea I was wanting to explore, but it was definitely far from perfect. In that equation, home runs and triples were included, as well as instances when the player got on base with another runner ahead of them on the basepaths. These events were not actually stolen base opportunities, making the accuracy of the stat a bit wanting.
I started searching for more specific baserunning data, a rare publicly available asset. Luckily, I was able to get in contact with Jeff Zimmerman of FanGraphs who was able to hook me up with the baserunning data I needed to properly identify a stolen base opportunity. Jeff is great, and I really appreciate his help with this project. Go show him some love and read his work!
The data provided by Jeff identifies a stolen base opportunity as a player on first base with second base open. A second set of data exists for a combined rate of the runner on first or second base with the next base open, but we’ll focus on the first dataset as the vast majority of stolen bases occur between first and second base. This data includes the yearly and career totals of stolen base opportunities along with attempted stolen bases, which we can use to calculate the year-to-year and career SBot for every player since 2005.
Let’s first take a look at 2018 SB+CS leaders, the men who attempted the most steals, along with the four SB factors (minimum of 250 plate appearances):
|Player||SB+CS||Sprint Speed (ft/sec)||SB%||SB Opportunities||SBot|
Before observations, I’ll add a bit of context to those numbers. An elite sprint speed is considered to be 30 ft/sec with the 2018 leader, Byron Buxton, clocking in at 30.5 ft/sec. As for stolen base percentage, it is generally agreed upon that a player must successfully steal bases at a 70% to 75% rate in order to help his team, while anything above that is considered to be good and anything below considered to be bad. Stolen base opportunities were outlined above and are heavily dependent on OBP and plate appearances. That said, it is difficult to say exactly what constitutes a good or bad SB Opps total, as it’s more of a contextual stat than a measure of skill. SBot, as it’s calculated now, is considered to be good at a 20% clip and elite at anything over 26%.
Now, with that said, here are some observations from the numbers above:
- Speed does not inherently equal success, as shown by Amed Rosario.
- Jonathan Villar is notably slower than I thought he was, yet he has the highest stolen base success rate of the bunch. That must mean his ability to read pitchers and knowing when to run must be elite. A similar case can be made for Jose Ramirez.
- Adalberto Mondesi is absolutely OUT OF CONTROL!
Let’s talk about Mondesi for a bit. A very polarizing player with tantalizing skills and a brilliant, albeit small, 2018 sample, Mondesi hit .276 with 14 home runs and stole 32 bases in only 75 games. That is bonkers. What may be even more ridiculous is his 53.42% SBot, which is easily the highest rate I’ve seen in a sample of this size. To clarify, that means that Mondesi attempts to steal second base in more than half of the opportunities he’s given! What’s even more impressive, he’s quite successful in doing so! Does this mean we could see Mondesi stealing upwards of 60 bases in 2019? Maybe, but I’m not sure.
This table has led me to a new theory: In order to be a reliable source of stolen bases, you must display well above-average marks in three of the four stolen base factors. Whit Merrifield led the MLB with 45 steals in 2018 while displaying nearly elite speed, a very good success rate, receiving plenty of opportunities, and a high SBot. His teammate, Mondesi, has elite speed, the same great success rate, and an apparent pathological need to sprint from first to second; the only thing he lacked in 2018 was ample opportunity. Meanwhile, Rosario stole 24 bases, with elite speed and plenty of opportunities, yet he lacked the success rate of Merrifield and Mondesi, which may have led to a lower willingness to run as we can see in his 20% SBot. Now, 24 steals is nothing to scoff at in today’s stolen base landscape, however, you’d like to see someone with Rosario’s speed have a higher stolen base total.
Now, this begs the question: What do we as fantasy baseball owners do with all of this information? Can we use it to predict future stolen base production?
I believe we can.
If we work from the assumption that my hypothesis is true, we simply need to identify players who have one or two of these four stolen base factors and the ability to improve the others in order to fully unlock their bag-swiping potential. Speed is a factor to look for in potential improvement candidates as it is the least likely of the bunch to drastically improve. With that in mind, let’s take a look at the fastest players from 2018 (minimum 25 competitive run opportunities*):
|Player||Team||Sprint Speed (ft/sec)||SB%||SB Opportunities||SBot|
|Ronald Acuna Jr.||ATL||29.6||76%||147||14.29%|
*I’ve included Garrett Hampson here despite having only 21 opportunities because of his likelihood of regular playing time in 2019.
First off, please forgive the poor formatting of the SB% column. Some of the players listed did not have very many attempted steals, and putting 100% for Hampson’s two steals didn’t feel right. The above players were the 15 fastest men in MLB in 2018. Now, excluding injury, we should assume regular playing time for Billy Hamilton, Trea Turner, Harrison Bader, Mondesi, Mallex Smith, Trevor Story, and Ronald Acuña Jr. The addition of consistent opportunities to steal should improve the values of Bader and Mondesi from their 2018 levels, assuming their 29% and 26.5% respective strikeout rates do not get out of hand and drastically lower their OBP next season.
Taking a look at the remaining names, we can reasonably speculate a scenario where Buxton, Delino Deshields Jr., and Hampson could see regular at-bats. As of April 1, 2019, Buxton and DeShields are both in a position to be their teams’ regular center fielders, while Hampson now seems to have a path to playing time as the Rockies’ second baseman.
This second group, the players who are likely to see a large increase in their playing time (and subsequently their stolen base opportunities) are a good group to target as potential steals sleepers for 2019.
Buxton is such an interesting case — the undisputed top prospect who simply has not panned out at all. Buxton played only 64 games total in 2018, the majority of which took place in AAA. Despite not receiving a September call-up, the Twins do seem prepared to move forward with Buxton as their starting center fielder at the major league level. This will at least give him the opportunity to prove he belongs in the big leagues and the opportunity to steal bases for your fantasy team. Buxton is the fastest man in MLB, and he’s proven he can steal very efficiently with his 29-for-30 stolen base record in 2017. The key here is whether the 25-year-old former top prospect can hit well enough to warrant a spot on the 25-man roster, let alone the starting lineup. We may have seen Buxton’s floor last season, as the Twins couldn’t justify keeping him on their big league roster. However, the ceiling may very well be a .250 to .260 batting average with 15 home runs and 35 to 40 stolen bases. That steals total could climb even higher if the Twins decide to let him run wild, yet it seems unlikely that we will see this ceiling in 2019. I’m not sure we’ll ever get to see it, and that’s honestly a shame for baseball as Buxton could have been (and may yet be) one of baseball’s biggest stars.
DeShields has been a popular steals target late in drafts for what feels like half a decade. He was a popular breakout candidate in 2018, but injuries hampered him throughout the season. Still, he managed to steal 20 bases in just 106 games with a paltry .216 batting average and .310 on-base percentage. Unlike most steals specialists in today’s environment, DeShields is more than capable of taking a walk. This gives him elite speed, a high success rate, and plenty of opportunities. We can only hope that the Rangers will be willing to give DeShields the green light more often to get his SBot higher than 19.2%
As for Hampson, we have heard all offseason about his tools. The man has 70-grade speed that was further verified by Statcast upon his arrival in the majors and an impressive hit tool that will only get further played up in Coors Field. What’s more, Hampson has stolen 125 bases out of 148 attempts in his professional career. Now, all but two of those steals were against minor league batteries, but it still shows he knows when to take the jump. Now, with Daniel Murphy shelved for a while, Hampson ought to have ample opportunity for at-bats. Hopefully we can see some of that speed put to use on the basepaths.
I foresee SBot continuing to evolve as we get more clear and accurate data and more opportunity to put it to use in the fantasy baseball context. I’d love to see SBot become a Pitcher List custom stat staple on the level of VPR, but we may have a ways to go before then. The development of this statistic is very much an ongoing process; I would love to answer any and all questions and inquiries that pertain to it.
Whether you just want to know Starling Marte‘s career SBot or how it could be applied at a team level, I probably want to know as well but haven’t gotten around to asking that question yet! So please, by all means, leave me comments and let’s continue working on this new tool together.
Graphic by Justin Paradis (@freshmeatcomm on Twitter)