Jamesd wrote this as a comment on the Bar, and I thought it deserved to be a post on the main site. -AAR

One of the most exciting aspects of this year was RAJ’s base-stealing exploits. With the new rule changes, I don’t understand why more players didn’t run wild. I thought there were good reasons why RAJ shouldn’t run as often as some players: the Braves offense is so good that his CS are more likely to cost the team runs than CS by players on other teams; similarly, the batters behind him are more likely to HR, making his successful steals somewhat less valuable than those by players on other teams; he was prone to minor injuries even before his 2021 knee injury; and he has good but not amazing speed. He refused to listen to me, and he ran often, and it was exciting. However, I doubted that the steals were as valuable as the attention they received implied, and I wanted to take a closer look.

This is the first of three posts I plan to write on RAJ’s steal attempts, which I reviewed on BRef’s game logs. I expected to find that they were valuable but not nearly as valuable as his batting contributions. I looked at this from two angles – the value of each steal attempt at the time it happened (post 1), and its eventual value given how the rest of the inning & game turned out (post 2). I expect to write a third post with a few other points I thought were interesting.

First, I considered the win probability added for each steal attempt and compared that to RAJ’s total WPA for the season. Most of his successful steals were worth .02 WPA or less. He only had five steals that increased the WP by .05 or more:

  • August 20 vs. SF: Down 3-2, bottom 8th, 2 out, he stole 3rd and scored on a throwing error to tie the game. That was worth .29 WPA, I’m sure mostly because of the throwing error & run. Yates decided it would be fun to walk everyone in the 9th, and the team lost 4-3.
  • April 16 in KC: 4-4 in the 8th, 1 out, he stole 3rd. That was worth .08 WPA. The next three batters went K, BB, K and no runs scored, though the team later won 5-4.
  • September 27 vs. the Cubs: Two that I imagine we all remember. At 5-5 in the 10th, one out, he stole second, worth .06 WPA. Albies drove him in with a game-winning single (though if he hadn’t stolen, he would’ve still been on second or third after Albies’s single with just one out and still a good chance to win).
  • Earlier in that game, down 3-2 in the 8th with one out, he stole second, worth .05 WPA, and was driven in by a game-tying Albies single.
  • May 13 in Toronto: 2-2 in the 7th with no out, he stole second, worth .05 WPA. Two walks, a force at home, a strikeout, and a flyout followed, so nobody scored, and the team wound up losing 5-2.

RAJ also had eight CS with a WPA of -.05 or worse.

The costliest was on April 12 against the Reds with the score tied 4-4 in the 7th and no out. He was caught trying to steal second, with a WPA of -0.10. No runs scored that inning, but the team went on to win 5-4. The above includes a little foreshadowing, as it shows that even his five most valuable steals only significantly affected the outcome of one game, and his most damaging CS didn’t prevent the team from winning.

Combining all his WPA gained from SB and lost from CS gives about 1.09. His total WPA for the season is 6.8, so the steals have value, but are nowhere near as valuable his contributions with the bat. (In one place BRef refers to the 6.8 as being offensive WPA, but in another it’s “the contribution of the batter during the game,” so it’s possible that the 6.8 doesn’t include the 1.09.)

The 1.09 combined WPA from attempted steals is an estimate, but I assume the true combined WPA was slightly different, for three reasons: First, BRef rounds WPA numbers for each individual play to the nearest .01, so all the .02 WPA successful steals might have been worth .016 or .024, and the rounding errors might add up to .1 or .2 one way or the other, but probably not much more. Also, BRef’s WPA assumes average batters afterwards, and Albies/Olson/Riley/Ozuna with their power were more likely than average to drive him in even without the steals, so the steals added a bit less value and the outs subtracted a bit more than BRef says. Finally, there were 8 cases in which RAJ attempted a steal while a batter was striking out or walking or was part of a double steal attempt. In those cases, BRef just gave one combined change in WPA for both the actions, so I had to guess how much of the WPA change was due to RAJ’s action, and I might have been off by .01 a few times. None of these issues should change the overall conclusion that RAJ’s steals were nowhere near as valuable as his contributions with the bat.

Next up: How valuable RAJ’s steals and CS wound up being based on what happened afterwards.