Nerves of Steal

The stolen base. Arguably one of the most exciting parts of a game: The runner weighs his perceived speed against the combination of the pitcher’s concern with holding the runner and the strength of the catcher’s arm. Many MLB greats have used the stolen base as an auxiliary part of a highly successful career (Joe Morgan, Roberto Alomar, etc.), whereas other players lived by, died by and built a whole career around it (Vince Coleman).

Evaluating the effectiveness of a stolen base has been a growing process, from Pete Palmer’s claim about 30 years ago that a SB gave a team 0.3 runs, whereas a CS was worth -0.6 runs, to Tom Tango’s more precise 0.18/-0.43 values. The latter is what we’re using here today. These values state that, on average, a CS is 2.39 times more detrimental than a SB is helpful, meaning that a runner needs to successfully steal 70.5% of the time to merely break even.

Unfortunately for the statheads among us, Caught Stealing data was inconsistent, especially in the National League, until 1951. With apologies to Ty Cobb, Billy Hamilton and other early ballplayers, they are not included in the data in favor of robust data.

Below are the Top 10 most successful basestealers of all-time*.

Rk Player SB CS SB% From To Age G SBRuns+ CSRuns- NetSBRuns
1 Rickey Henderson 1406 335 80.8% 1979 2003 20-44 3081 253.08 -144.1 109
2 Tim Raines 808 146 85.0% 1979 2002 19-42 2502 145.44 -62.78 82.7
3 Willie Wilson 668 134 83.3% 1976 1994 20-38 2154 120.24 -57.62 62.6
4 Vince Coleman 752 177 80.9% 1985 1997 23-35 1371 135.36 -76.11 59.2
5 Joe Morgan 689 162 81.0% 1963 1984 19-40 2649 124.02 -69.66 54.4
6 Davey Lopes 557 114 83.0% 1972 1987 27-42 1812 100.26 -49.02 51.2
7 Kenny Lofton 622 160 80.0% 1991 2007 24-40 2103 111.96 -68.8 43.2
8 Ozzie Smith 580 148 80.0% 1978 1996 23-41 2573 104.4 -63.64 40.8
9 Ichiro Suzuki 444 100 81.6% 2001 2012 27-38 1890 79.92 -43 36.9
10 Lou Brock 938 307 75.3% 1961 1979 22-40 2616 168.84 -132 36.8
Provided by Baseball-Reference.com: View Play Index Tool Used
Generated 9/17/2012.

However, many players considered to be high-profile basestealers, contributed little to no, and sometimes negative, value with their thefts, such as Steve Sax (3.38 Net SB Runs) or Brett Butler (-10.91). Charlie Hustle was known for his aggressiveness on the base paths, but would you have guessed he was the least efficient basestealer in our study? Would Buddy Bell be considered a borderline Hall of Fame candidate by non-saberists if he wasn’t abysmal on the basepaths?

Rk Player SB CS SB% From To Age G SBRuns+ CSRuns- NetSBRuns
1 Pete Rose 198 149 57.1% 1963 1986 22-45 3562 35.64 -64.07 -28.4
2 Buddy Bell 55 79 41.0% 1972 1989 20-37 2405 9.9 -33.97 -24.1
3 Alfredo Griffin 192 134 58.9% 1976 1993 18-35 1962 34.56 -57.62 -23.1
4 Chet Lemon 58 76 43.3% 1975 1990 20-35 1988 10.44 -32.68 -22.2
5 Greg Gagne 108 96 52.9% 1983 1997 21-35 1798 19.44 -41.28 -21.8
6 Rick Monday 98 91 51.9% 1966 1984 20-38 1986 17.64 -39.13 -21.5
7 Duane Kuiper 52 71 42.2% 1974 1985 24-35 1057 9.36 -30.53 -21.2
8 Dave Parker 154 113 57.8% 1973 1991 22-40 2466 27.72 -48.59 -20.9
9 Bob Bailey 85 83 50.6% 1962 1978 19-35 1931 15.3 -35.69 -20.4
10 Lenny Randle 156 112 58.2% 1971 1982 22-33 1138 28.08 -48.16 -20.1
Provided by Baseball-Reference.com: View Play Index Tool Used
Generated 9/17/2012.

As you can see, the bulk of the players listed in the above tables played in the 1970s and 80s, an era with more emphasis on the stolen base and “small ball” in general, coming in the wake of a pitching-dominant era where an extra base here or there was quite valuable. Also, when expanding the lists, one will note that current players are more and more successful basestealers than those of previous generations, likely due to better pre-game scouting and paying greater credence to statistics. There are 10 active players among the 40 best basestealers, but only one among the 40 worst (sorry, David DeJesus). Teams like the 2007 Phillies stole at an absurd 88 percent clip, whereas the 1977 Cardinals were successful only 54.5% of the time, poor enough to “earn” -24 runs.

One final note for this piece: The .18/-.43 run values are for the stolen base ON AVERAGE. On a case by case basis, the values will differ, but as a rule of thumb, these are the best figures to use (according to Mr. Tango).

*All-time, for my purposes today, means 1951-present.

Leave a Reply

73 Comments on "Nerves of Steal"

Notify of
avatar
Sort by:   newest | oldest | most voted
MikeD
Guest

Buddy Bell and Duane Kuiper were teammates. Even at that time when the understanding of the value of SBs vs. getting caught wasn’t quite as well understood, it was just crazy that these guys were ever allowed to wander more than two feet off first base.

RJ
Guest

Duane Kuiper: fantastic broadcaster, less than fantastic ballplayer.

Ed
Guest

Tom Veryzer was also a teammate of Bell and Kuiper. For his career Veryzer was 9 for 32 in steals, a 28.1% success rate. It’s the 5th lowest success rate for anyone with more than 20 attempts.

Tmckelv
Guest

And don’t forget Alfredo Griffin (#3 on the second list above). He spent his first 3 seasons (1976-78) as a backup infielder “learning” the art of stealing bases from 3B Bell, 2B Kuiper and SS Veryzer.

Alfredo chipped in with 2 SB and 3 CS in very limited playing time with CLE before getting traded to the Blue Jays where he could REALLY not shine on the base-paths.

Ed
Guest

As bad as he was, Griffin actually improved the Blue Jays base stealing prowess. The year before acquiring him, the Blue Jays were only successful 28 times in 80 attempts, perhaps the worst team effort ever. The only player over 50% was Gary Woods who was successful in his only attempt. The Blue Jays worst offender was the player Griffin replaced, shortstop Luis Gomez. Gomez was 2 for 12 in 1978 and 6 for 28 in his career.

John Autin
Editor

FWIW, the BJHBA says the Bell/Kuiper Indians did a ridiculous amount of hit-and-running, contributing mightily to the team’s awful CS numbers.

Hartvig
Guest

Is there an easy way to determine how often a team ran the hit & run during the course of a season? I read that same thing and assuming that it’s true it appears that it may have spanned at least 3 different managers- Ken Aspromonte, Frank Robinson & Jeff Torborg. That strikes me as kind of odd but it’s hard to imagine another explanation for their astonishingly bad caught stealing numbers.

Doug
Editor

This year’s Pirates aren’t much better than the ’77 Cardinals. The Bucs have only a 56.7% success rate, but at least have only had 111 attempts.

Most attempts this year is 176 by the Marlins, followed by the Padres (175) and Rays (169). In 1977, the Pirates stole 260 bases in 380 attempts. In 1982, Oakland stole 232 times (130 by Rickey) in 319 attempts.

Ed
Guest

Would love to see a follow-up post that looked at what teams gained or lost the most runs via the SB.

Jim Bouldin
Guest

As far as gained runs goes, as a percentage of total runs, it’s got to be dominated by the 76 A’s and the Expos and Cardinal teams of the 1980s.

Doug
Editor

Those 260 steals for the ’77 Pirates actually cost them 4.8 runs.

Hartvig
Guest

Plus they had Omar “The Out Maker” Moreno in their lineup for the entire season yet they did manage to score about 20 runs more than the league average. I’m not sure if those 2 factors alone were enough to cost them the pennant but they didn’t help.

Doug
Editor

It was a slow, gradual process, but Pete Rose did finally figure out how to steal.

Age 22-28: 50 for 110 (45.4%)
Age 29-43: 137 for 225 (60.9%)
Age 44-45: 11 for 12 (91.7%)

Ed
Guest

Geez, if Rose had been able to play into his 50s, he might have been able to challenge some of Rickey’s records!

spudart
Guest
Debra
Guest

Wow, great insight by this writer! Kuiper’s low/detrimental SB % I would think when combined with zero power contribution really brought about a negative offensive presence for the Tribe!

Ed
Guest

Rick Monday may not have been able to steal bases but he sure was good at stealing flags!

Jim Bouldin
Guest
Of course, Palmer and Tango’s numbers are almost certainly derived from curve fits to various empirical data sets, and so are of limited usefulness in explaining the potential for the stolen base to be a powerful offensive weapon in the right context. Thanks for including the caveat that the two sets of numbers are for MLB overall, and therefore are also of limited utility in comparing players from different teams and different eras. Getting thrown out when you’re on a team that hits lots of home runs, is not going to be the same as Coleman getting thrown out on… Read more »
Brendan Bingham
Guest

Yes, context is important, including consideration of game situation. An extreme example: with two outs in the bottom of the ninth and trailing by one run, the break-even point for attempting to steal second is something like 35%.

AlvaroEspinoza
Guest

I’d love to see how base stealers effect the performance of the hitter. Do they distract pitchers when they’re on first, force pitchers to throw more fastballs and decrease their effectiveness? Or do base stealers force hitters to take more pitches and decrease their performance? Or have no effect at all?

no statistician but
Guest
Alvaro: The whole dynamic can be very complicated and variable, depending on who is involved. Example: for years almost no one dared to attempt a steal on Whitey Ford because his move to first was indistinguishable from his move to the plate. And if someone did, he started late and got mowed down by Yogi or Elston Howard. This situation made the jobs of the Yankee infielders easier, the hit-and-run harder, and the double play more likely. With someone like Adam Dunn at the plate, I doubt if base stealing has much impact, assuming the runners in front of him… Read more »
Howard
Guest

Where does Carlos Beltran rank?

Doug
Editor

Good thought, Howard.

Apparently, just outside the top 10 at 35.3 runs. But, it could be a slow process moving higher – his 13 of 18 mark this year added just 0.19 runs to his total.

Howard
Guest

Thanks, Doug. That’s pretty impressive considering how far behind he is in raw numbers compared to the other guys on the list.

Jimbo
Guest

do the .18/-.43 take into other affects of an aggressive base runner?

I find it hard to believe that a theoretical baserunner who steals 142 bases in a year and is caught 60 times is only breaking even, what with all the mayhem he would be causing.

Jimbo
Guest
I’ve also always been surprised that the bases are just the right distance apart that base stealing is so interesting. And that there has never been a player so fast that he could run on the first pitch every time he got on base, and do so with enough success. Such a player could steal 250+ bases in theory or have 300-400 attempts. I think such a player would be too injury prone and beaten, but I’m surprised no speedster has ever tried, especially considering how overlooked the CS traditionall was. The 142/202 player I theorized above would probably garner… Read more »
Lawrence Azrin
Guest
From wiki.answers: “The Base Ball Convention of 1857 held in New York on January 22, by former Knickerbocker President Daniel “Doc” Adams, was the first convention of its kind. During this convention Adams was elected President of the Convention and headed the Committee on Rules and Regulations. He set the distance between bases at 30 yards. It has been assumed that the distance was 30 paces between all bases although this distance was never specified in the original Knickerbocker Rules. The only mention of distance is listed in Rule 4 and states; “The bases shall be from “home” to second… Read more »
Doug
Editor

Dick Schofield Sr. may be the most consistently bad base-stealer. At least he had the good sense not to try too often. He went 12 for 41 for his career, and not once in 19 seasons did he have more steals than caught stealing.

Ed
Guest

Jay Buhner has the lowest career % of anyone with 20 or more attempts. He was 6 of 30, a 20% success rate.

e pluribus munu
Guest
I think we have to remember that managers have a lot of say about who steals when, and high CS numbers for certain players may reflect poor judgment by managers. Schofield Sr. was one of the longest serving perpetual subs, so an inordinate number of his PAs were probably in late innings of close games as PH, PR, or 1-2 inning defensive replacement, when managers might be more prone to gamble for a small ball edge. (I don’t know this, I’m just speculating from recollections and too lazy to look up a sample of Schofield’s logs.) A guy with Schofield’s… Read more »
Doug
Editor

No doubt, epm. I just thought it mildly interesting that with year after year of low single digits in the SB and CS columns, not once did Schofield manage say, 2 SB and 1 CS. It was always equal or the other way round.

bstar
Guest

A lot of the credit for those record-breaking 2007 Phillies has to go to stolen base guru Davey Lopes, the first-base coach of that team. Lopes had the all-time record for SB% for a long time at 83.01%(557 out of 671) before Carlos Beltran passed him fairly recently.

Hank G.
Guest

“Lopes had the all-time record for SB% for a long time at 83.01%(557 out of 671) before Carlos Beltran passed him fairly recently.”

At the time Lopes retired in 1987, Time Raines was 511 out of 585 (87.35%), so I don’t see when Lopes had the all-time record at all.

bstar
Guest

I guess you’re right about that, Hank. I don’t know why Lopes is stuck in my mind as once-the-career-leader, but he is. My guess is maybe after 1981, when his career SB% was 83.1, or after ’79, when it was also 83.1%. I always think of that when I see Lopes coaching first base for whomever, but I guess I need to exorcise that thought from my mind.

Doug
Editor

Lopes was the career leader after the ’76 and ’79 seasons.

http://www.baseball-reference.com/leaders/stolen_base_perc_progress.shtml

Hank G.
Guest

So, if Rickey Henderson had never attempted to steal a single base, his career would have been worth only about 95 WAR. Good thing he made the extra effort.

Steven
Guest

Those 1977 Cardinals were nicknamed “Rapp’s Runnin’ Redbirds.” Ran Vern Rapp right out of a job in early 1978.

NatsLady
Guest
If I understand the commenter above, the WPA of stealing second down a run in the bottom of the 9th with two out makes the attempt worth it even if you have only a 35% success rate. If that’s true, why is a SB often ruled defensive indifference? Obviously, it is DI if you look at positioning of the infielders–who are often repositioned after the 2nd out to maximize getting the 3rd out. Is this the best strategy? Apologize if I misunderstood the comment. The Nats have been bit and bit and bit by SB’s and DI in late innings.
Doug
Editor
NatsLady, Tom Tango’s win expectancy estimates by state (from which WPA is most likely derived) read like this, for bottom of 9th, down 1, 2 outs. Man on first ONLY: .104 win expectancy Man on second ONLY: .151 win expectancy Man on third ONLY: .172 win expectancy Using the values above with a 35% chance of stealing successfully, you get: (0-.104) * 0.65 + (.151-.104) * 0.35 = -.0511 . So, with only a 35% chance of success (i.e. expectation is defense will contest the play), the steal attempt actually reduces your chance of winning by about half, down to… Read more »
Brendan Bingham
Guest
Natslady and Doug, Thanks for following up on my comment, and Doug, thanks for citing Tango’s win expectancy numbers in reply. I was not imagining a DI situation. Here’s my (perhaps flawed) reasoning for “guesstimating” a 35% break-even. With a runner on first and two outs, it probably takes two hits (or a walk and a hit) to bring in the tying run and extend to extra innings. This of course is not equivalent to winning, merely to extending play. If you think of the stolen base as “replacing” one of the necessary hits by putting the tying runner on… Read more »
Doug
Editor
Brendan, The other factor is who the batters are (something that WPA and Win Expectancy don’t consider, of course). I would be more inclined to try the steal the weaker the hitters are who are coming up. So, if my best hitter is at the plate, I’m probably staying put, more especially so if I have a low expectation that I’ll succeed with the steal (which make’s Ruth’s choice to steal in 1926 all the more surprising, with Meusel and Gehrig up next.) As to Natslady’s questioning why teams are indifferent to taking the extra base, I personally can’t recall… Read more »
birtelcom
Editor
When I apply the +0.18 and -0.43 values to the total, overall major league-wide numbers for each season from 1951 through 2012, what I get is that across MLB there was an overall NET LOSS in runs from stolen base attempts every single season from 1951 through 2004, with the sole exception of 1996 when there was a tiny net gain of 5.5 runs league-wide. The league wide loss in net runs during this period was often over 100 runs a season (the average net loss over the 1951 through 2004 period was 92 runs), with the largest negative net… Read more »
Jim Bouldin
Guest
birtelcom, Why do you conclude that the “runs gained” and “runs lost” numbers are the root of the problem here? You are placing some kind of complete trust in Tango’s numbers, and I see absolutely no reason to do that. We don’t know what time period he derived his numbers from, nor anything at all about the methodology he used to get them. It’s a complete black box and so you can’t just apply them to any MLB numbers you want. Futhermore, even if we applied them to the “right” time period, the conclusion that SB/CS numbers somehow imply that… Read more »
birtelcom
Editor

Jim: The gist of your comment is entirely consistent with mine. Dalton’s original post applied the Tango numbers (.18 and-.43) to all players in all seasons since 1951. I simply wanted to see what would happen if we applied those same numbers on a league-wide scale, and I reported the results. And I added that I thought the results suggested some skepticism about whether applying the formula so uniformly resulted in an accurate reflection of real value. So I think we are agreeing.

Jim Bouldin
Guest

My fault birtelcom. I completely failed to read your last sentence, or else I didn’t get what you were saying there, can’t remember which. So thanks for explaining.

The problem with this kind of stuff is people start taking these numbers as some kind of gospel and applying them willy-nilly.

birtelcom
Editor

No worries, Jim. I wasn’t particularly clear in my comment, plus the gist does sometimes get hard to follow in a winding comment thread. (Which is a perfectly acceptable price for having such an enthusiatic and articulate community of commenters.)

Lawrence Azrin
Guest
My problem with your conclusion is that the 70.5% success-rate (derived from the +0.18/-0.43 run values for SB/CS) is _not_ static over time, and varies considerably over various eras. In particular, when the runs-scored level is well below-average (say, from 1963-68), the break-even point goes down considerably. At the height of the recent offensive explosion (say, 1996 or 2000), the break-even point may have been 72%-73%. Conversely, in 1968, “the year of the pitcher”, it was probably below 60%. The break-even point may also vary with the park, or the team’s offensive style (someone mentioned the 80s Cardinals). We don’t… Read more »
Jim Bouldin
Guest

Offensive strategy will have an enormous impact Lawrence. In addition to being very questionable as *overall* numbers (as birtelcom shows), it’s almost certain that they don’t apply to a good base stealing team, particularly if that team depends on high OBP and stealing for its offense. That’s why numbers derived from empirical analyses are limited in what they can tell you.

birtelcom
Editor

Tom Ruane of Retrosheet did this article a while back that has value-added numbers for each year from 1960 through 2004: http://www.retrosheet.org/Research/RuaneT/valueadd_art.htm Based on his numbers, it looks like the major-league-wide break-even point in 2000 was about 75%, and about 65% in 1968. Tom Ruane, BTW, is one of the great forces behind Retrosheet, the amazing source of the basic baseball box score and play-by-play historical data in digital format on which we all ultimately depend. There can never be enough kudos to Tom and to Retrosheet.

Artie Z.
Guest
> That MLB as a whole was pursuing an entirely voluntary strategy that was on the whole continuously and consistently counter-productive for at least 50 years does seem odd. While I don’t disagree with any of Jim’s or Lawrence’s comments about the particular numbers for the value of stolen bases, I want to comment on this statement about MLB as a whole doing things that don’t seem to make a lot of sense. You can start with teams using fast guys with low OBP in the leadoff spot. There are still teams, playoff caliber teams even, that do this ***cough,… Read more »
James Smyth
Guest

Another feather in the cap for the Hall of Fame candidacy of Tim Raines! http://raines30.com/

Dan McCloskey
Editor

Hear, hear!

mosc
Guest
I don’t agree with the numbers. I know it’s not feasible for historical players but I’ve always wanted to see RE24-based numbers for steals. It matters on the outs and the other base runners. Look at the run expectation before and after each SB attempt. There are lots of situations where a SB is of great value, others where it is of low value. If you steal second with a guy on third and two outs, that’s not adding nearly as much value as if there were no outs in that situation. Going first to second with one out is… Read more »
spudart
Guest

I’d like to see a list of the top 100 players. I subscribe to Play Index, but the link in the post doesn’t go directly to the results.

spudart
Guest

Here’s the top ten Cubs of all-time (since 1951)
Ryne Sandberg 15.91
Davey Lopes 7.76
Tony Campana 7.39
Bob Dernier 6.8
Eric Young 6.27
Corey Patterson 5.59
Alfonso Soriano 4.17
Brian McRae 4.15
Delino DeShields 3.1
Ryan Theriot 2.95
Ced Landrum 2.71

Artie Z.
Guest

When you realize that Davey Lopes did that in just 174 games and he was around 40 years old at the time … wow.

T-Bone
Guest

Tony Campana did it in just about the same number of games, 179, but he is a bit younger than 40. He doesn’t have a whole lot of value other than that but it is fun to watch him run, whether for an SB or a hit.

Mike Felber
Guest
Over 70% success rate merely to break even-really? I am aware of run scarcity differentials, but I thought that the break even rate averaged around 2/3. This is about 4% higher than I figured. Is everyone confident that this both applies as a historical average, & in this post-PED era? I can imagine that folks consistently underrated the damage from CS though. And true, if the base paths were just 5% closer or farther apart, it would render the challenge to easy or hard. Either not an effective tool for most all, or so easy that it is not much… Read more »
tag
Guest
It’s interesting to speculate when baserunners may simply become too fast to be thrown out. It’s hard to think that Trout is faster than, say, Bo Jackson or Deion Sanders were. Which makes his success rate all the more remarkable, because theirs wasn’t near his. And it’s not that, like, say, Lopes, he’s been in the league 15+ years and has had time to learn all the pitchers and intricately refine his approach. I imagine that at some point certain players might become too fast to be thrown out by all but pitchouts / perfect pegs. The Cubs’ Tony Campana… Read more »
no statistician but
Guest

tag:

I think you’re over-simplifying a little.

Stolen bases depend quite a bit on getting a good start. Speed helps, but we’re talking maybe 27 yards here, so the initial takeoff is crucial, and that depends on the pitcher to a degree.

tag
Guest

Probably oversimplifying, but I think it’s interesting that we’re getting young guys like Trout and Campana who, in a significant number of attempts, are being successful around 90% of the time now. Sure they’re good baserunners, but so was Lou Brock. These guys are just flat-out faster than in the past, and at some point other guys make become too fast.

Jim Bouldin
Guest
You’re asking exactly the kinds of questions that are of the greatest interest to me in baseball discussions–questions of optimal baseball strategy relative to the constraints imposed by the rules of the game and the geometry of the playing field. I’ve always wondered how the dimensions of the diamond were set exactly, but it’s clear to me that there must have been some experimentation with different distances between the bags for the typical stolen base attempt to typically be a very close play at both second and third. Given that they set it at d = 90′ some long time… Read more »
Mike Felber
Guest
I am skeptical that 70.5 % is not a bit too high as a historical AVERAGE for a break even point. Or that today, when run production has cooled to ~ historically average post dead ball levels, it is also not around 2/3. Is anyone here confident that the break even point IS over 70%. IF it is lower, then that would obviate the need to believe that the net effect of stolen bases was a negative, wholly effort for most, for 1/2 a century! Though when players were slower when the distances were greater, the run environment tended to… Read more »
Jim Bouldin
Guest

Not sure I’m following what you’re saying there Mike, can you give it another go?

Mike Felber
Guest
Me no speak-a so good? 😉 I am inquiring whether 70.5% is a reasonable break even point for SB/CS value, today & for a historical average. From what I have read elsewhere, including at this site’s progenitor, that is about 4% too high. The rest is positing that if so, then the Tango’s pessimistic view of the efficacy of thievery for 50 years from the mid 50’s may be show as wrong. And the available info saying that for close to a decade value has been created by SB attempts may be even stronger. Plus I posited earlier that this… Read more »
Mike Felber
Guest
Once Again, with Feeling: can anyone answer my query & musing? I never saw 70% + listed as a SB break even point, any opinions or proofs? I asked: I am inquiring whether 70.5% is a reasonable break even point for SB/CS value, today & for a historical average. From what I have read elsewhere, including at this site’s progenitor, that is about 4% too high. The rest is positing that if so, then the Tango’s pessimistic view of the efficacy of thievery for 50 years from the mid 50′s may be show as wrong. And the available info saying… Read more »
wpDiscuz