In 2008, Ichiro Suzuki, regarded as one of the fastest players in baseball, stood at first base with second base empty a total of 262 times. From 2006 to 2008, Ichiro has been thrown out by the catcher on only 9.2% of his attempts, one of the best rates in history. Despite this, he only attempted 34 steals, a rate of 13.0%. In 1980, a year in which Omar Moreno of the Pirates was thrown out by the catcher on 22.2% of his steal attempts, Moreno ran an amazing 68.4% of his opportunities. At that rate, Ichiro could have stolen 163 bases last year.
A successful stolen base increases the team’s chances of scoring, but it is paired with the risk of the cost in lost opportunity of a caught stealing. The successful steal is most valuable when it is more difficult scoring by other means. In 1968, when Bob Gibson had a 1.12 ERA the average major league wOBA was .289, a successful steal was worth .027 wins, while each caught stealing cost a team .040 wins. With the high offense of recent years, from 1999 to 2002 the value of a steal had dropped to .018 wins, while each caught stealing cost .043 wins. The return is lower, while the cost of being caught is more. In 1968, a runner had to be safe 60% of the time to break even, but now it’s risen to 70%.
From 1963 to 1968 baseball used a high pitcher’s mound and a larger strike zone, depressing offense to the point that in 1968 the majors as a whole hit only 237/302/340. Although these were rescinded in 1969, large ballparks and expansion also helped to keep offensive levels low through 1976. In such an environment, stolen bases were one of those market inefficencies that some teams used to try to gain a competitive advantage. The major league rate of attempting steals of second was 6.7% in 1971, and then began a steady increase until it peaked at 10.9% in 1976. That year, manager Chuck Tanner‘s Oakland A’s stole 341 bases while being caught 123 times. In 1977 Tanner was traded to Pittsburgh, where the Pirates stole 260 bases in Tanner’s first year at the helm, and over 200 in three of his first four seasons. The Cardinals and Astros were two teams that played in cavernous stadiums paved with AstroTurf, and both turned their lineups away from power and to speed.
From 1976 to 1992, the attempt rate held steady between 9.6 and 10.9%. The premier base thieves were given a green light to steal at will, with Moreno setting the record with a 68.4% rate of attempts in 1980. Rickey Henderson attempted second base 51.5% in his record year of 1982, and again at 50.9% in 1983. Vince Coleman topped 40% six of his first seven seasons, and Tim Raines ran 61.5% of the time in his rookie season of 1981.
The brakes were applied starting in 1993, coinciding with an explosion in production in the batter’s box that continues to this day. There has been much study of the two year jump in offense in 1993 and 1994, including an article last year by Tom Tango at The Hardball Times, where he advocated a juiced ball as the culprit. Whatever the reason, run scoring was up 13% from 1993 to 2008, compared to 1976 to 1992.
Paradoxically, even as the average rate of caught stealing has continued to drop to where it has only been 22% the past two years, the rate of stolen base attempts has also dropped steadily since 1992. With conditions so favorable for power hitting, teams have moved their rosters away from speed and towards power. The higher batter’s production has then decreased the benefit of a successful steal while increasing the cost of a caught stealing. With fewer steals being attempted, teams can risk playing offense oriented catchers with below average arms such as Jorge Posada or Ryan Doumit. Therefore, most steals today are attempted when the runner has a substantial advantage – a superior base stealer matched against pitchers and/or catchers lacking the skill to stop him (Dexter Fowler vs Chris Young, Carl Crawford vs Jason Varitek). Being more selective in attempting steals then further drove down the caught stealing rates.
In the 17 season from 1976 to 1992, there were 52 player seasons of more than 100 opportunities in which the runner attempted to steal second more than 40% of the time, 9 times higher than 50%. In the 16 seasons since, there have been only 5 player seasons of 40% or better, none higher than 50% – Chuck Carr 42.2% in 1993, Alex Sanchez 41.9% in 2003, Corey Patterson 40.7% in 2006, Jose Reyes 40.2% in 2007 and Roger Cedeno 40.1% in 1999. Today’s best base stealers, such as Willy Taveras, Hanley Ramirez, Michael Bourn and Jacoby Ellsbury attempt steals of second base in the 30% range, where as 25 years ago they each would have likely stolen at least half again as often. Ichiro and Carlos Beltran, with two of the lowest rates of caught stealing, only attempt in the 15% range.
It was the lack of offense of the early 1970’s that drove the push for more stolen bases, but the need for speed was cut back nearly twenty years later when suddenly anyone in the lineup had the ability to hit the ball out of the park. Willy Taveras bragged this past winter that he could steal 100 bases with Cincinnati – of course he will need better than a .270 on base percentage, but he also will likely never be given the green light to attempt 50 or 60% of his opportunities as long as offensive levels stay near their current levels. It just doesn’t pay enough.
Rate of attempted steals of second base
Rate of caught stealing by catcher at second base
The play by play information used here was obtained free of charge from and is copyrighted by Retrosheet. Interested parties may contact Retrosheet at “www.retrosheet.org”.
Thank you for reading
This is a free article. If you enjoyed it, consider subscribing to Baseball Prospectus. Subscriptions support ongoing public baseball research and analysis in an increasingly proprietary environment.Subscribe now
Brian, just a suggestion for an extension (and not a negative comment on this article at all): maybe quantify the tradeoff between catcher defense (throwing out runners) and offense that's evolved over time. Very interesting stuff.
One thing I'd love to see (and Will might hate to see) is some tables. In particular, the top N player-seasons, player-career, and team-seasons for stolen base attempt rate. Maybe that's available on Retrosheet somewhere, but I couldn't find it easily. :)
Writing: B - Brian's cleanest job, yet, but there were a couple of points that required me to go back and re-read it to make sure it made sense. And, yes, it was dry, but not enough to lose my interest.
Any thoughts on if the bigger / stronger players (due to PEDs and increased weight lifting) lead to less speedster types playing prominent roles on teams? Guys that used to be 'fast' got bulkier and slower
(I think of Bonds and Sosa as guys who obviously slowed down as they grew).
One critique - I would've positioned the charts in the article as opposed to at the end. I expected something after the charts and had a "oh, that's it" reaction.
I liked it, and it gets a thumbs up, but I wasn't wowed by it.
Unfortunately, the way I have the data extracted I can't tell how they reached base. I can look at the database structure and see how easily I can add that info.
And there should have been at least a little acknowledgement that if Suzuki had tried to steal 68% of the time he had an opportunity, like Moreno, his caught stealing rate would certainly have shot up. There's no doubt that Suzuki picks his spots wisely.
Unfortunately, this week's instructions were a day later than usual (Tues 1 pm) which pretty much gave a two and a half days to research and write, then the first idea had to be abandoned because MySQL decided to run like a tortoise and eight hours later hadn't given an query results. Then thursday night to find out that Dan Turkenhopf at day did a historical stolen base article at The Hardball Times.
Two main things should be considered when deciding to steal or not - how important is the run to winning the game, and what are the chances of being successful? I want to add win expectancy table a to my steals database. As Matt Swartz showed in his article this week, WE gives a weight on each event based on how important it is to winning the game, looking at inning, outs, bases occupied and score. Then look at the ability of each pitcher and catcher to prevent a successful steal. The combination of the two would give the expected win probability added with an attempt in that particular situation.
It would be a safe assumption that most steal attempts would be in the situations with the highest eWE (and this can be tested, looking for the average Att% at different levels across time). If the Mariners decide to run Ichiro at a 50% rate instead of 15%, we could look at the 50% of the times he was on 1st that offered the best eWE, and then sum the expected steals, caught stealing and pickoffs given those runner, pitcher and catcher matchups. Just not enough to do in a day or two, but it is now on my to-do list.
How about an assignment that limits the writers to one chart or table?
He'd make a killer backend guy, so long as you let someone else actually choose the topic and write the article.
There was simply nothing wrong with this article. I loved it.
Otherwise, we'll never learn of great new analysis if it happens to be discovered by someone who can't write well.
I haven't looked at the constraints placed on you for this week's effort, but that data could have been explained more clearly by placing a relevant chart/graph or two early in the article, then identifying the elements of the chart that you found most interesting. This approach leans on the visual to help tie together what you are saying.
The other piece I was left wanting is judgment. Has the base stealing pendulum swung too far to the conservative side? Oakland has moved into the middle of the league in stolen base attempts, where they used to trail the league significantly. And that's with a AL worst OBP. Is this shift something the A's are trying to exploit? Jack Cust picked up the second stolen base of his season and his career last night. Kurt "the slower" Suzuki has 3 SBs this year, bringing his career total to 5.
What is your opinion? I'm Seattle's manager. I have Ichiro "the speedy" Suzuki on first in the bottom of the first with no one out. I have a league average pitcher going up against a league average pitcher, with a league average catcher behind the plate. I don't know if the game will be decided by one run or ten. Do I steal? If the break even mark for value is 70% success, shouldn't I steal almost every time in this situation? If I think Ichiro is successful 80% of the time in this situation (hypothetically invented number), don't I have to go if your 1999-2002 numbers approximate the state of the game today (the value of a steal had dropped to .018 wins, while each caught stealing cost .043 wins)? How many wins did Seattle leave on the table last season by not stealing more aggressively with Ichiro? This is what I was seeking at the end of your article. Synthesize the numbers and tell me not just what managers have done, but what they should be doing.
This has to violate some type of word count limit.
There were also several points that were made that seemed to beg for added comment. For example, the notion that stealing was at one point a market inefficiency demands that one contrast the value vs. the cost. Maybe this was implicit, but I didn't see it.
A second potential point where the analysis could have been deepened was the reference to the "cavernous" stadiums in St. Louis and Houston. Doesn't that just beg for a park-adjusted measure of the value of a steal?
I'd be curious to see a chart of park-adjusted steal "break evens."