No Honour Among Thieves

On a subjective level, the stolen base is one of the more enjoyable aspects of the baseball game for me. The cat-and-mouse of the runner’s lead and pickoff throws (although that can sort of get out of control), pitch outs, and the actual run-and-throw and “did he make it” is all fun to watch.

Objectively speaking, in the current era of relative baseball enlightenment, we know that the stolen base is not all that valuable. Linear weights reveal that, on average, a stolen base is less valuable than any of the “good” batting events: not just any kind of hit, but a walk, too, is more valuable than a stolen base.

That is not to say that stolen bases are not good. On average, a stolen base in modern baseball is worth just under .2 runs. Of course, that is just in an average context. Obviously, different steals have different run values given the base and out situation, and different win values given the relative score and inning (Win Probability Added is one metric that tries to measure the value of such events). The value of Dave Roberts’ famous steal in the 2004 ALCS obviously is not best summarized in “0.2 runs,” even objectively speaking.

On average, though, base stealing failure is much more costly than success is beneficial. In modern baseball, while the stolen base is worth less than 0.2 runs, getting caught costs more than 0.4 runs. Getting caught stealing hurts a team more than twice as much as a successful steal benefits them, on average.

Looking at a player’s seasons stolen base total by itself is thus not much help in telling us about his value in that respect without looking at how many times he got caught in the act. So today, I am going to use linear weights to take a look at the five individual seasons in “modern baseball” with the most steals that were actually net losses.

[The next few paragraphs are about methodology. If you are not interested in that, you can skip down to the numbered list, although it will make more sense if you at least glance at this stuff.]


If you want simple generic weights for stealing and caught stealing, +.2 for steals and -.4 for caught stealing work pretty well (although they slightly understate the usual gap of the caught stealing) for most of what I call “modern baseball,” or the “Retrosheet Era” starting in about 1955. That designation is based on when we start having accurate records of caught stealing and other events that were not tracked as carefully in previous seasons (although that situation is slowly being rectified).

If you want to be more careful, linear weight values do change from year to year. In the case for base stealing, the “break even point” goes down the more stingy the run environment becomes. When fewer runs are scored, there are fewer runs per out, and thus outs are less valuable. The difference on a league and season level is not huge from year-to-year in the modern era (although on a game level, it tells us why attempting a steal with Roy Halladay on the mound is a better play than attempting a steal with Kyle Davies staring the batter down).

For this post, I am using custom linear weights for each season that I generated using Base Runs. I will hopefully get a chance to post about that in more detail down the road. For now, I simply decided to use the custom weights to reflect the base runners’ “wisdom” relative to their seasonal run environments.

Methodological excursion over — let’s get to the players.


I searched for seasons since 1955 with the most steals and a negative total run value of the base stealing. Here are the top five base stealing seasons that actually hurt the teams the players for which the players played according to linear weights.

No contemporary players had a season in the the top five, but everyone’s favorite Dodgers’ contract of recent years — Juan Pierre — made the list with the 2004 Marlins, when he had 45 steals but got caught 24 times, costing -2.05 runs according to my linear weights. While he had fewer steals than anyone in the top five, he did cost his team the most runs, so he has that going for him.

5. Luis Polonia, 1991, 48 SB, 23 CS, -0.54 runs

Polonia had three different stints with the Yankees, and played on World Championship teams with the 1995 Braves 2000 Yankees. His first stint with the Yankees came in 1989, when Oakland sent him to New York as part of the trade that sent Rickey Henderson (of all people) to the As. Polonia was primarily known as a speed threat throughout his career, but while he did steal many bases (321 for his career), he was not especially efficient (caught 145 times). From 1991 to 1993 with the Angels, Polonia stole more than 150 bases, but was caught almost 70 times. He managed more than than 50 steals in both the 1992 and 1993 seasons, but in 1991 he managed only 48, and cost the Angels more than he gave them in thievery.

4. Miguel Diloné, 1978, 50 SB, 23 CS, -0.17 runs

If I were a betting man, I would put money on being the only contemporary blogger to write two posts prominently involving Miguel Diloné. Diloné was another speed merchant with very little else to offer. He did have a marvelous 1980 season with Cleveland (including stealing 61 bases while only getting caught 18 times), but in 1978, he offered nothing but speed, and as you can see, even that was pretty much pointless.

3. Eric Young, 1999, 51 SB, 22 CS, -0.2 runs

Young had a nice career as a speedy second baseman who could get on base, and the Colorado Rockies are currently hoping that junior can offer something like that (they shouldn’t hold their breath). Dad was a very efficient base stealer most seasons, but in 1999 with the Dodgers he was not.

2. Billy North, 1974, 54 SB, 26 CS, -0.33 runs

North may be best known fighting with Reggie Jackson in the 1974 As’ clubhouse. Ray Fosse came over to break it up, and both Jackson and Fosse ended up getting hurt. Jackson was not exactly anyone’s favorite person, so who knows who started it. North was no Reggie Jackson on the field, but he was a good player. He did not hit for average or power, but he took plenty of walks, stole a lot of bases, and was considered as a good defensive center fielder in his prime. He stole more than 50 bases in four different seasons, and led the league with 75 steals in 1975. He also led the American League in steals in 1974, but as you can see, got caught enough to cancel that out.

1. Steve Sax, 1983, 56 SB, 30 CS, -2.00 runs

You don’t know when to stop running, do you, Saxxy Boy?

Steve Sax was the 1982 National League Rookie of the Year. In his New Historical Abstract, Bill James opines that despite the impressive performance, one could perhaps divine from Sax’s “class clown” personality that he was not destined for true greatness. It is not as silly as it sounds — James is just speculating in a “writerly” way. It is worth checking out.

Sax is currently 51st all-time in steals. For most of his career his was fairly efficient, but in some season, like this season and the one that followed it, he got caught too many times for his own good. Getting caught with a hand in the cookie jar is enough to get a guy six consecutive life sentences.

Comments (14)

  1. One issue with the caught stealing stat is that it doesn’t really tease out times which might not have been the runner’s fault – specifically, a batter failing to execute on a hit and run. That might only happen a couple times in a season – but to the first three guys in the list, that would’ve nudged them into the net positive category.

    Also, how are pickoffs counted? I always thought certain ones were marked down as caught stealing (EG, if the runner gets caught in a rundown)

    And there’s no way to account for an extra base gained if a steal forces an errant throw (again, infrequent, but aside from Sax most of these guys are one or two extra bases away from breaking even).

    Basically, there’s some undocumented factors that complicate the picture far more than a strict success / failure analysis would reveal. And that’s without going into the situational effects you already mentioned.

    • I think Matt sufficiently states this in the body of the article, your just going into more detail with more examples. Either way it is what it is, a snapshot to be anaylyzed with a grain of salt.

      • Yeah. The details in this case are more the icing on the cake. Point remains that aggressive base running alone isn’t anything near what it’s cracked up to be.

        It certainly is exciting to watch though.

        • “Point remains that aggressive base running alone isn’t anything near what it’s cracked up to be…”

          … if you’re bad at it. These guys all were.

          Counter example: Henderson, White, Molitor and Alomar were a combined 133 / 158 for the 93 Jays (partial season in Rickey’s case, obviously). They were generally aggressive base stealers, but efficient at doing so.

          Even ignoring the extraneous effects noted above, and using the 0.2 number (that’s likely under-valuaing speed guys, as a prime effect of stealing a base is to move yourself into a position where a fast player can score on a single; a good speed guy would be better off at second than the average player) those guys added 16 runs to the Jays’ season total by stealing bases alone, or a couple of wins (expected: 91; expected minus 16 runs, 89) based on the Pythagorean calculator I used.

          Two wins per year isn’t shabby.

      • That’s the point. People want to think that baseball is a simple analysis. It’s more fluid than a lot of statistics give it credit for, especially when we’re dealing with actual game situations (stolen bases) rather than just pitcher / batter interactions.

        It’s like how two singles are theoretically more valuable than a double – but a double has a lot of unrecorded positive consequences (no exposure to a double play on a ground ball, potential to all but guarantee a run with a single if there’s two outs, etc.). When calculating expected outcomes, these things have to be factored in.

  2. Neat article. I wonder if Matt can give us some thoughts on the 1985 St. Louis Cardinals (where Vince Coleman had over 100 steals?). Base-stealing was such an integral part of their offensive game plan.

  3. running should be exclusive to those who can steal bases at higher than an 80% rate.

    the “average” base stealer only adds .2 of a chance for a run, but that includes a wide range of players.

    I guarantee that if that same process was looked at Rajai Davis, that he contribute a lot more than .2 .

    • Take a look at Coleman’s WAR….it was only 1.9 (Baseball Reference) for 1985 with 110 steals.

      He had a higher WAR in 1990 stealing only 77 bags due to his ability to reduce outs with a higher OBP (albeit only 20 points).

  4. The lineup at Getting Blanked is insane: Fairservice, Parkes, both (!) Platoon Advantage guys, Creamer, Klaassen and Robinson. (Although I have no idea if they alternate lefty-righty, heh.) Throw in some Stoeten on the podcasts and a capable weekend editor and dayummmn, this is a hub of goodness.

    It’s a goddamn shame that when Stoeten at DJF sharts out a post about the Blue Jays signing Aaron Laffey it gets 400 comments (and quite likely the concomitant page views), while really analytical pieces like this barely scrape double digits. Also, that’s no knock on Stoeten, I love DrunkJays, too, as equally as I would love two children I don’t have, but it makes me sad that debating reactionary Jays’ fans seems to bring more traffic than advanced baseball perspectives. Not that it’s surprising that appealing to the lesser elements of our culture generates traffic… TMZ, Perez Hilton, HuffPo slideshows, ugh, even the effing newscasts. Infotainment forever!

    Anyway, I suppose I will take solace in the fact that DJF has probably converted more wins-and-RBI luddites to modern baseball thought than any other place in the Jaysosphere.

    Still, like, WTF people?!? You can be a Jays fan and a baseball fan. Srsly! It might even make you a better, more perspicacious Jays fan.

    [Aaaaand scene.]

  5. Just a method question. Obviously there are always qualitative details that are left out when you quantify, but the overall question is whether – other things being equal – you buy the form of quantification and the subsequent analysis as telling you something worthwhile.

    However, on this:
    “When fewer runs are scored, there are fewer runs per out, and thus outs are less valuable.”
    you are referring to differences in raw value, but not necessarily the relative value of base-running – am I correct? So at the end of the day, we could still assume these run values to be a standardized value (e.g. value in a base year) regardless of the year in which they actually occurred. It seems to me like a caveat that doesn’t really need to be provided. In fact, if you accounted for the year (or team), you’d also have to provide the average number of runs in that year (or team), wouldn’t you? The analysis has its weaknesses, but maybe this statement isn’t one of them?

Leave a Reply

Your email address will not be published. Required fields are marked *