Steal Attempt Grader
Every stolen base attempt carries a break-even probability: the success rate at which the attempt is worth the risk. We grade each attempt against that threshold using empirical success rates binned by runner sprint speed, catcher pop time, and outs.
Key finding: Most steal attempts (2020–2025) were positive expected value decisions. The steals that hurt teams most often came from slower runners facing elite catchers — situations where the success rate fell below the RE24 break-even threshold.
Most efficient team (full sample)
+3.25
COL 2021 — 3.25 runs / 100 attempts
Best individual runner
+3.74
Myles Straw — 93 attempts
Worst individual runner
-2.02
Willy Adames — 52 attempts
Overall good-steal rate
65.9%
Attempts above break-even, 2020–2025
About these grades
Each steal attempt is graded against the RE24 break-even success rate for its base-out state. Empirical P(safe) comes from 27 bins (runner speed tier × catcher pop time tier × outs) computed from all graded attempts in 2020–2025. Pop time data is from Baseball Savant (2020+ only). Runners with fewer than 50 career attempts are flagged Low. Double-steal contexts use the definitional pre-steal base state. 2020 entries are flagged 60g. Entries marked Live are from the current in-progress season. Full methodology