# FanGraphs Baseball

1. I have a question about the Acceptable SB% in the 2nd graph. Is this the break-even point? Because if so, it suggests to me that teams were stealing too much. An Acceptable SB% equal to the actual SB% would imply that net, teams derived no value from SBs.

Comment by LK — February 7, 2013 @ 12:10 pm

2. Tom Marr: “Bill Whitehouse, Earl, from, uh, Frederick, Maryland, wants to know why you and the Orioles don’t go out and get some more team speed? ”

Earl Weaver: “Team speed, for crissakes, you get [expletive] [expletive] fleas on the [expletive] bases, getting picked off trying to steal, getting thrown out, taking runs away from you. You get them big [expletive] who can hit the [expletive] ball out the ballpark and you can’t make any [expletive] mistakes.”

Comment by rbenchley — February 7, 2013 @ 12:37 pm

3. It’s not that they were stealing too much, but they were “too successful”. Implying that they could’ve ATTEMPTED MORE steals (thus using up some of the more “risky” situations) which would drop their success rate to the break even point

Comment by Matt — February 7, 2013 @ 12:41 pm

4. -sorry this was meant to be a reply to “LK” above

Comment by Matt — February 7, 2013 @ 12:42 pm

5. Right, but my point is that isn’t the break-even point where you derive exactly 0 value from stolen bases? If not, then I’m not understanding what is meant by break-even.

Comment by LK — February 7, 2013 @ 12:42 pm

6. Correct. It seems that only recently have teams been able to extract any net positive value from SBs. My guess is that the success rate will continue to improve as teams eliminate attempts where the probability of success is <66%.

Comment by Jason — February 7, 2013 @ 12:49 pm

7. Right that’s what I was thinking. Intuitively, you want to keep stealing while the *marginal* success rate is above the break-even point, but the marginal rate will always be lower than the average rate. If the average rate equals the break-even rate, that would imply teams were attempting steals with a probablility below the break-even rate that were pulling the average down to break-even.

Comment by LK — February 7, 2013 @ 12:55 pm

8. Excuse me for being ignorant, but what was the major rule change in 1951 referred to in figure 1?

Comment by Dave (UK) — February 7, 2013 @ 1:05 pm

9. This is precisely why we want to find some method of guessing marginal steal success rates. No, you don’t want to have a net 0% (a la Mets and Cubs in 2012 http://www.fangraphs.com/blogs/index.php/the-changing-caught-stealing-calculus-2/), but keeping 5% or so above should, theoretically, keep you in the most productive zone. I look at the late 1980s as the ideal era of base-stealing efficiency.

Comment by Bradley Woodrum — February 7, 2013 @ 1:10 pm

10. http://en.wikipedia.org/wiki/Stolen_base#Evolution_of_rules_and_scoring

Comment by Bradley Woodrum — February 7, 2013 @ 1:11 pm

11. Thanks

Comment by Dave (UK) — February 7, 2013 @ 1:20 pm

12. *curtsy*

Comment by Bradley Woodrum — February 7, 2013 @ 1:20 pm

13. Gotcha. So your contention is that the marginal SB% is currently above the break-even point. I certainly could see that being true; as you allude to in the piece that’s a very difficult thing to analyze.

Comment by LK — February 7, 2013 @ 1:30 pm

14. Perhaps this has been done before and I missed it, but I wonder if one would want to look at whether an increased number of steal attempts is correlated with a higher injury risk as well.

Comment by KM — February 7, 2013 @ 1:43 pm

15. Seems to me the “team SB%” is heading down the wrong path (akin to “GDP” stuff in macroeconomics). Teams don’t steal bases. Individual baserunners do. Further, that player skill is a highly context-dependent skill – and sabermetric measures have done everything possible to try to eliminate context as a measure of skill. So the attempt to measure it using existing measures is akin to seeing everything as a nail because all you have is a hammer.

From a managerial perspective, your first post was right on – each batting spot in the lineup has a particular probabilistic “acceptable SB rate” based purely on the HR rate of the batters following. And the challenge is then to a)identify the individual players who can best take advantage of those spots and b)adjust a batting lineup to maximize the SB opportunities in a game given the starters one has in that game.

Longer-term, I suspect that flexible batting orders may well be the key. Not because the different lineups produce different game outcomes ceteris paribus (or using static analysis). But because those different lineups provide differing opportunities for players with different skill sets to maximize (and ultimately be rewarded for) the skills they have. And that sort of incentive can definitely change game outcomes.

Comment by jfree — February 7, 2013 @ 2:43 pm

16. Wow. I was enjoying that until halfway through, then it started becoming much to complex for me to figure out and I was just reading without understanding. Guess I’ve got a long ways to go.

Comment by Swfcdan — February 7, 2013 @ 3:19 pm

17. Ha, love the Neverending Story reference. But does that make you the dragon or the Empress?

Great article. I really think it will take scouting data to answer the riddle of your GREAT MISSING DIMENSION, though. You don’t really know where each individual player is on their base-stealing efficiency curve. I think at the very least, you’d want details on the speeds and the kinds of jumps individual runners have, so that you can come up with some probabilities of how good they *should* be at base-stealing.

Comment by Steve Staude. — February 7, 2013 @ 3:19 pm

18. What was up with 2009 Ryan Howard? How did he steal 8 bases and why did he attempt 8 steals?

Comment by Ruki Motomiya — February 7, 2013 @ 3:49 pm

19. The Giant Dog.

Comment by Baltar — February 7, 2013 @ 6:50 pm

20. That’s how I feel on Fangraphs a lot of times. Love the site, just makes me wish I would have paid more attention in my Statistics classes back in HS/college.

Comment by Ericpalmer — February 7, 2013 @ 8:59 pm

21. Because steals lead to RBIs!

Comment by Ericpalmer — February 7, 2013 @ 8:59 pm

22. Scouting data sounds like a decent way if the author wants to further test his line of thought. Maybe try Tango’s fan scouting report as a starting place.

Comment by Walkoffblast — February 8, 2013 @ 12:29 am

23. It takes less than stolen bases to make breakfast delicious. BACON!!! Anyway, an interesting read. thanks!

Comment by Nats Fan — February 8, 2013 @ 3:10 am

24. Balk or no balk:

http://mlb.mlb.com/video/play.jsp?content_id=20836363&c_id=mlb

http://mlb.mlb.com/video/play.jsp?content_id=22067929&c_id=mlb

I think when he bends that front knee he is heading home. It freezes the runner and his quick move gets the runner picked of. I think the reason he is known as Mr. Johnny “10%” Cueto is that he effectively disguises a balk.

Comment by McAnderson — February 8, 2013 @ 10:18 am

25. Great article! For the record, the Brewers led the league with 158 SB.

Comment by McAnderson — February 8, 2013 @ 10:53 am

26. Just for accuracy’s sake: I think the big drop in SB% in 1951 is just a record-keeping artifact, not the result of a rule change. The National League didn’t start tracking caught-stealing numbers until that year, so I’m guessing your SB%’s before that are reflecting SB’s from both leagues but CS’s from just one. If you look at just AL numbers, there’s no dramatic change in SB%.

Comment by Mississippi Matt Smith — May 31, 2013 @ 1:07 pm

Line and paragraph breaks automatic, e-mail address never displayed, HTML allowed: `<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong> `