Mailbag: Are the Padres Stealing Wins?

Bad timing for the headline, but this is a fun one…

Hi Geoff,

Here’s a question for your mailbag:

I hate it when the Padres steal a base. Why? Because Mark Grant and Dick Enberg (whom I love) inevitably bring up that stupid correlation between stolen bases and wins this year for the Pads. Is this correlation outside of the norm for the league? What about doubles? Do the Padres tend to win as often when they hit at least one double?

It’s kind of like saying that the Padres are more likely to win when they do things well. Well… duh?


Thanks, Andy, for the question. It’s funny you should mention this; I was just wondering the same thing a few weeks ago. I actually started running numbers and drafting an article, but then stopped because it seemed heavy on the duh factor and I wasn’t sure anyone else would find it interesting.

First, a brief philosophical interlude… People love stories. People love when things makes sense. One of the key narratives coming into the 2010 season was that the Padres would be aggressive on the bases this year. We heard the story from the beginning of spring training and came to believe it.

In fairness, there is truth in the story. The Padres have been aggressive on the bases. Sometimes it has helped them, other times not so much. Either way, this is part of the narrative and so it is what we have.

We also have a team that is winning games with surprising regularity. We have to attribute the Padres’ unexpected success to something, right? (Because, you know, we love when things make sense). Why not stolen bases? After all, they are a manifestation of aggressive baserunning. The idea that stolen bases is leading to success fits the narrative well and therefore makes us happy.

Okay, fine. How about actual baseball? Well, here are the numbers through Friday, August 27:

SB?  G  W-L   Pct
Yes 60 49-11 .817
No  67 27-40 .403

Wow, all that running sure helps.

Well, maybe. Although it’s true that the Padres win more often (twice as often, in fact) when they steal bases than when they don’t, it’s also misleading. Here’s the bigger picture:

SB?  G  W-L   Pct  R/G   BA  OBP  SLG
Yes 60 49-11 .817 5.38 .276 .359 .420
No  67 27-40 .403 3.64 .227 .289 .341

It would be more precise to say that the Padres produce more runners when they steal bases than when they don’t, thus leading to more runs and to more wins. That’s still not quite right, as really the reverse is true: They steal more bases when they produce more runners. In graphical form, this is what we’ve been hearing:

Stolen bases -> Runs -> Wins

And this is what’s been happening:

Baserunners -> Runs -> Wins

In other words, stolen bases are a by-product, not the cause.

A couple other oddities caught my eye:

  • The Padres are 11-0 when they steal three or more bases in a game.
  • Their 27-40 record when they fail to steal a base breaks down as follows:
    • 0 CS: 16-29
    • >0 CS: 11-11

That is not what I would have expected. All I can think is that maybe the Padres are getting caught more often because they have more baserunners in those games, which as we know leads to more runs… You can make yourself crazy thinking about this stuff.

As for how the Padres’ current success in games where they steal a base stacks up against other teams, I don’t know. Someone with a good database might be able to answer that question. (For the curious, I’m just dumping game logs into a spreadsheet and then executing simple operations. That’ll do for a team or two, but it gets rather tedious after a while.)

Anyway, Andy also asked about doubles. Here’s that last table re-run with doubles:

2B?  G  W-L   Pct  R/G   BA  OBP  SLG
Yes 96 60-36 .625 4.88 .265 .336 .403
No  31 16-15 .516 3.19 .202 .281 .300

This makes my head hurt. First off, how do the Padres have a winning record with a batting line of .202/.281/.300? Well, of those 31 games in which the Padres failed to hit a double, five coincided with shutouts thrown by one (or more) of their pitchers. They allowed just one run in four more. So, yeah, great pitching has helped.

Second, why is the Padres’ record in games where they steal at least one base so much better than that in games where they hit at least one double? This would require more rigorous testing than I’m prepared to do right now, but I have my suspicions:

  • Stolen base attempts are dictated to a degree by game situation. If you’re losing by a lot, you’re less likely to run. There’s a selection bias at work. With doubles, on the other hand, there is never a question of intent. Batters presumably always try to make hard contact, which sometimes results in doubles.
  • Doubles are relatively common. The Padres have hit at least one in roughly 72% of their games this year. Even still, they rank last in the NL as of this writing, which means that whatever team they are playing on a given night hits doubles with even greater frequency. If we expect doubles to lead to runs (and wins), this should hold for both teams. In other words, if the Padres gain an advantage by hitting a double in a game, so also does their opposition by doing the same… and they do so more often.

Like I said, though, these are just untested suspicions.

Oh, and here’s a fun small-sample quirk:

2B  W-L
3  11-1
4   4-0
5   3-0
6   0-3

Yep, once the Padres reach three doubles, they are in great shape… unless they hit that sixth. Clearly the solution is to hold at first with a single in that situation. Then steal second, thus increasing the chances of winning and further validating the narrative. Or something like that.

Bottom line: The assertion that the Padres win a lot when they steal a base is true, but it doesn’t tell the whole story. Thanks again, Andy, for the excellent question. If nothing else, it’s comforting to know I’m not the only one thinking about these things.

Tagged as: , , , , , ,

5 Responses »

  1. I wish they would attempt fewer steals because it kills me when they give away outs. According to EqBRR and a couple other stats, the Padres have cost themselves runs with their baserunning this year.

  2. Good article and good point. All these stats can be summed down to a John Madden statement: The team that scores more points has a better chance of winning! Stolen bases, doubles, triples, singles, walks, HBP, errors are all about getting that guy on base and plating him.
    Even with a 7-game slide, SD is still 3 up in the division. Given that vs. a winning streak and being 3 games down, I’ll take the former, thank you very much.
    The Padres need to recharge today (off day) and get ready for a big homestand. SD has a .528 winning percentage against the division and Colorado has a .529 winning percentage against the West, only slightly better. However, the Padres have a .575 winning percentage at home, while Colorado has a .382 winning percentage on the road. Based on those numbers, I like SD to win this series and break this losing streak, esp with Garland and Richard on the hill Sat and Sun.

  3. I am all for aggressiveness, but I sometimes get the feeling that Bud has a requirement of at least one voluntary out on the base paths per game.

  4. There’s a time and a place for steals. In fact, there are situations where I demand that the Padres DO steal a base.

    Expected runs table for base-out situations:

    Bases   0 out    1 out    2 outs
    ---      .455     .239     .090
    x--      .820     .490     .210
    -x-     1.054     .650     .314

    With a man on first and nobody out, the Padres have 0.82 expected runs in the bag. When they attempt to steal 2nd, they are risking 0.581 runs (man on first, 0 out – no runners, 1 out) to win 0.234 runs (man on second, 0 out – man on first, 0 out) In other words, they’re betting $581 to win $234 at a craps table at The Opposing Battery Casino. That’s 1.48 to 1. It’s a winning bet if the success rate is 67.4% or higher (1/1.48).

    The following players are above that number (plus a few for discussion):

    Player/    SB/  CS/   Pct/
    Durango      3   0  100.0%
    Stairs       2   0  100.0%
    Venable     26   4   86.6%
    S.Hairston   7   1   85.7%
    Gwynn       17   4   80.9%
    Headley     16   5   76.1%
    J.Hairston   9   6   66.7%
    Torrealba    6   3   66.7%
    Cabrera      8   5   61.5%

    Now, we’re going to sacrifice Durango and Stairs to the Cawnteckst, God of small sample sizes. We could do the same to Scotty, but let’s let him stick around. Other than that, we can project the EV of each steal attempt, by multiplying the success rate by 0.234 runs, then subtracting the failure rate x 0.581.

    Venable +0.125
    S.Hairston +0.117
    Gwynn +0.078
    Headley +0.039
    J.Hairston -0.038
    Torrealba -0.038
    Cabrera -0.079

    Multiply each of these by total attempts to see how many runs each of them have contributed:

    Venable 3.750
    Gwynn 1.638
    S.Hairston 0.936
    Headley 0.819
    J.Hairston -0.570
    Torrealba -0.38
    Cabrera -1.027

    Not much, since the Padres have scored 576 runs this year. It’s not that much of a swing in the grand scheme of things. But, Venable and Gwynn have both turned a tangible profit, and Scotty has done okay. Headley’s on the edge though… the rest are all losers on the basepaths.

    But that doesn’t change the fact that Venable, Gwynn, Scotty and Headley are all winners. If I were Bud Black, I’d order Venable and Gwynn to steal second with nobody out at will. Especially with Venable’s patented “delayed steal” trick, where he’s been very successful at it. Hairston and Headley less so… I’d have them do it if they were sure they can, but to not attempt it unless they were sure. The rest I’d ban from stealing, unless we’re losing by a three runs or more. We have to get back into it somehow, and that means taking some risks. But, if we’re tied or ahead, it’s a red light for everybody on the team save four, and Scotty and Headley’s light is yellow.

    As for Venable… GO! GO GO GO!

  5. Thanks, 8-East; nice job. If only Venable would get on base more often. He’s pressing, and so is the whole team.