The Stolen Base Break-Even Point: How Often a Runner Has to Succeed

By 4Baseball Staff · November 5, 2026 · 9 min read

How run expectancy gives a break-even success rate of roughly 70 percent for a steal of second, why it changes with outs and base, and how the 2023 rule changes moved the league's real success rate to 80 percent.

A runner who steals second base is trying to do two things at once: move himself into scoring position and avoid making an out. The arithmetic of that trade is simple, and it has a famous answer. In a typical modern run environment, a runner on first has to succeed roughly 70 percent of the time before an attempt to steal second helps his team more than it hurts. Below that rate, standing on first and doing nothing is the better play.

This article derives that number from scratch, so a reader can reproduce it, then shows how it moves with the outs, the base and the era. It ends with what happened to the league's real success rate after the rule changes of 2023, using season totals from the historical record. The run values in the worked examples are rounded and illustrative, in the range that published run-expectancy tables give for the modern game; the method is exact, the inputs are approximate.

What run expectancy is

A baseball inning has 24 starting situations that matter, built from three out counts (0, 1 or 2) and eight arrangements of runners on first, second and third base. For each of these situations, the average number of runs the batting team scores from that moment to the end of the inning is its run expectancy, written RE. The number comes from counting. Take every time in a large sample of games that a team had a runner on first with nobody out, and average the runs it scored for the rest of that inning.

Run expectancy turns a play into a price. A runner who moves from first to second with one out lifts the team's expectancy from about 0.54 to about 0.70 in the table below, a gain of 0.16 runs. An out does the opposite: it lowers the expectancy by moving the inning to a state with more outs. The beauty of the idea is that it puts walks, steals, bunts and double plays on a common scale, runs, and a reader who has seen the sacrifice bunt and run expectancy math has already met the same tool.

These values are rounded and illustrative. Real tables vary by season and by source, because run environments change, and the break-even point shifts with them.

The break-even formula

Consider a runner on first with some number of outs, and ask what happens on an attempt to steal second. There are only two outcomes. If he succeeds, the state becomes a runner on second with the same number of outs. If he is caught, the state becomes bases empty with one more out. If he does not try, the state stays as a runner on first.

Let p be the chance he succeeds. The attempt is worth p times RE(second, same outs) plus (1 minus p) times RE(empty, one more out). Staying put is worth RE(first, same outs). Setting the two equal and solving for p gives the break-even rate:

p = [RE(first) minus RE(empty, one more out)] divided by [RE(second) minus RE(empty, one more out)]

An equivalent way to say it is in terms of gain and cost. The gain from success is the increase from first to second. The cost of failure is the drop from first to the empty bases with an extra out. The break-even rate is the cost divided by the sum of the gain and the cost. The larger the cost of being caught compared with the gain from succeeding, the higher the success rate the runner needs.

Working it through with the table

With nobody out, success moves the runner from RE 0.90 to 1.14, a gain of 0.24 runs. Being caught moves the team to bases empty with one out, 0.27, a cost of 0.63 runs. The break-even rate is 0.63 divided by 0.87, which is about 72 percent.

With one out, success moves the team from 0.54 to 0.70, a gain of 0.16 runs. Being caught leaves bases empty with two outs, 0.10, a cost of 0.44. The break-even rate is 0.44 divided by 0.60, about 73 percent.

With two outs, success moves the team from 0.23 to 0.34, a gain of 0.11 runs. Being caught ends the inning, so the team falls to zero for the rest of the inning, a cost of 0.23. The break-even rate is 0.23 divided by 0.34, about 68 percent. The two-out case is the lowest of the three because the runner on first with two outs is worth less to begin with, so there is less to lose.

All three results sit in the range of 68 to 73 percent, which matches the long-held rule of thumb that a stealer needs to succeed about 70 percent of the time. The real answer depends on the table you use, and a different era gives a different table.

Stealing third is a different bet

The same formula gives a very different answer for second base to third base. With nobody out, success moves the team from 1.14 to 1.35, a gain of only 0.21, while being caught drops it to 0.27, a cost of 0.87. The break-even rate is 0.87 divided by 1.08, about 81 percent. With one out the numbers are 0.70 to 0.95 on success and 0.10 after a caught stealing: 0.60 divided by 0.85, about 71 percent. With two outs a runner on second already scores on most singles, so the gain from third base is tiny (0.34 to 0.37) while the cost of being caught is the whole inning, and the break-even rate rises above 90 percent.

That is why stealing third base is rarer and rarely worth it. A runner at second is already in scoring position, so the extra base adds less than the first one did. The catcher's throw is also shorter: home plate to third base is 90 feet, while home plate to second base is about 127 feet.

A season-long example

The break-even rate is only a threshold, so it helps to see what clearing it is worth. Suppose a runner makes 40 attempts to steal second with nobody out, in the table above where success gains 0.24 runs and failure costs 0.63.

• At a 75 percent success rate he succeeds 30 times and is caught 10 times. The gain is 30 times 0.24, or 7.2 runs, and the loss is 10 times 0.63, or 6.3 runs. The net is about plus 0.9 runs.
• At 85 percent he succeeds 34 times and is caught 6 times: 8.16 runs gained, 3.78 lost, a net of about plus 4.4 runs.
• At 65 percent he succeeds 26 times and is caught 14 times: 6.24 gained, 8.82 lost, a net of about minus 2.6 runs.

A common approximation holds that about ten runs equal a win. On that scale the very good base stealer is worth around half a win from stealing alone, the break-even runner is worth nothing, and the poor one costs a quarter of a win. The point is not that steals are unimportant; it is that the steal is a small edge for most players, and the sign of the edge depends on the success rate more than on the attempts.

A shorthand follows from the numbers. If the break-even point is near 67 percent, the runner needs two steals for every caught stealing, so net steals, calculated as steals minus twice the number of times caught, is a quick check. A runner with 30 steals and 12 times caught has 30 minus 24, plus 6, and is above that line. A runner with 30 steals and 18 times caught has minus 6 and is below it. The two-for-one line is slightly generous, since it assumes a break-even rate of 67 percent, and a calculator such as the site's stat calculator makes the exact arithmetic quick.

What the table leaves out

The run-expectancy method uses averages, and a real steal attempt is not average. Several things push the true break-even rate up or down. A runner on first with a strong hitter at the plate risks more by going, because a caught stealing erases a runner who might have scored on an extra-base hit. With a weak hitter at the plate, the gain from reaching second is larger, because the runner is less likely to be driven in from first. The pitcher matters too: a left-handed pitcher faces first base and can hold a runner better, a point the platoon splits article considers from the batter's side, while a pitcher with a slow delivery or a high leg kick gives the runner more time.

The catcher and the pitch matter as well. A fastball reaches the glove sooner than a breaking ball, so a runner prefers to go on slower pitches, and the four-seam fastball and sinker and the slider, sweeper and cutter explain how pitch type changes the time to the plate. Other factors are the score and inning, since late in a close game a single run is worth more than the same run in the third inning, and whether the stolen base also keeps the batter from hitting into a double play.

The table also hides the batter's reaction. A runner in motion pulls a middle infielder toward second, which can open a hole for the hitter, and a pitcher who worries about the runner may throw more fastballs. None of these effects is easy to measure, and none of them is in a simple RE table, so treat the break-even rate as the starting line for judgment, not the whole answer.

What the league actually did, 1962 to 2025

The league's real success rate has been climbing for 60 years, and the table shows it using the historical totals for both leagues combined. The success rate is stolen bases divided by stolen bases plus times caught.

Two features of the table stand out. In 1968, a season of very low scoring, the league's success rate was 61.9 percent. The break-even point is generally lower when runs are scarce, so that rate may not have been as costly as it looks. In 1987 and 1999, when scoring was higher, the league's success rate was about 70 percent, near the break-even point, and steals per team game fell from 0.85 to 0.70 between those seasons.

The 2023 rule changes

Before 2023 the league's success rate had risen to 75 percent, but steals per team game had fallen to 0.47 in 2019. A common explanation is that in a league full of home runs, the gain from moving a runner up one base is smaller, while the cost of being caught stays the same, so fewer runners went.

Then the rules changed. For the 2023 season the bases were enlarged from 15 inches square to 18 inches square, which shortened the run from first to second by a few inches. Pitchers were limited to two disengagements, meaning pickoff throws or step-offs, per plate appearance, and a third that did not retire the runner was a balk. A pitch clock sped up the pitcher's delivery. Each change made stealing easier: the shorter distance, fewer chances to hold the runner and less time to deliver.

The results are in the table. Stolen bases jumped from 2,486 in 2022 to 3,503 in 2023, an increase of more than 40 percent, and the success rate rose from 75.4 percent to 80.2 percent. Attempts rose without a drop in success, which is the signature of rules that moved the break-even calculation in the runner's favor. The 2024 season held at 79.0 percent and 3,617 steals, and the 2025 season had 3,440 steals at 77.7 percent. The rules, as of the 2025 season, are on the official site at mlb.com.

Common questions

What is the break-even rate for stealing second? Roughly 70 percent in modern run environments, a little lower with two outs and a little higher with nobody out or one out. The exact figure depends on the run-expectancy table used.

Why is the break-even rate higher for third base? The runner on second is already in scoring position, so reaching third adds less, while a caught stealing still costs a runner and an out.

Is a stolen base ever worth more than the table says? Yes, in close late games when one run decides the result, and when the runner distracts the defense. The table gives averages across all situations.

Do the numbers change with the era? Yes. The break-even point tends to rise as scoring rises, which is why the 1960s and the 1990s call for different judgments.

Where to go next

The same logic that sets the steal threshold also governs bunting, and the sacrifice bunt and run expectancy article works through that case. Current stolen-base leaders are on the leaders pages, and for fantasy players the stolen-base category is covered on the fantasy pages. The statistics and analytics topic has more reading.