Thursday, August 19, 2010

Differential Madness!

No, not the kind of run differential madness that drives most Jays fans to drink. That kind of differential madness shows the Jays as the sixth best team in baseball (by Third-Order wins), still only good enough for fourth in their own division.

Nope, I'm taking about strikeout-walk differential. We're all familiar with strikeout to walk ratio as a decent measure of a pitcher's worth. The good men at The Book blog (aka Tom Tango) believe K-BB differential has a stronger relationship to both same year success as well as future performance.

One of the biggest troubles with striker per walk is it overvalues eliminating walks relative to adding strikeouts. From the comment section of that very post, we see that if a pitcher strikes out 6 per 9 compared to 3 walks, his ratio is 2. Drop a walk per nine and poof! you're a 3 K/BB and feted as a king. Add one strikeout but keep your walks the same, you're marginally better with a 2.33 K/BB. That just ain't right, especially since guys who miss bats are rack up Ks are harder to come by than guys who limit walks.

Another problem with K/BB is it doesn't stand up to the asymmetry test. If you invert the numbers, the results aren't the same. 3/1 gives a ratio of 3.00, 1/3 gives 0.333. These two numbers aren't the same "distance" from one, which makes them pretty flawed. Another reason home run per fly ball is a much better rate than AB/HR (a ratio.)

Anyway, lets look at the numbers. Which Jays stand to benefit the most for looking at their numbers this way (hint: Brandon Morrow.)

Brandon Morrow10.484.112.556.37
Casey Janssen8.272.193.776.08
Shaun Marcum7.52.063.645.44
Scott Downs7.081.863.85.22
Kevin Gregg9.664.532.135.13
Marc Rzepczynski7.292.692.714.60
Jason Frasor8.
Ricky Romero7.653.262.344.39
David Purcey7.713.512.24.20
Shawn Camp5.671.833.093.84
Brett Cecil6.412.812.283.60
Brian Tallet6.074.551.331.52
Jesse Litsch3.092.891.070.20
Dana Eveland4.235.440.78-1.21

Pretty logical stuff. Guys like Gregg and Morrow get away with much, much more than a Jesse Litsch, and this differential method really shows it. By K/BB, Litsch and Tallet don't lag too far behind Morrow and Gregg. A better valuation of strikeouts makes the differences that much more striking.

Thinking about it another way, consider we flipped the rates used. Instead of strikeouts per 9 innings, we used K per batter faced. A pitcher may walk 12% of the hitters he faces but strike out 25%. His K/BB is 2 to 1, just like a pitcher who walks only 6% and strikes out 12%. Using the differential method, we see a much better representation of the qualities of a good pitcher.

If you click through and read The Book post, you'll see that not only are these "the qualities of a good pitcher" as I just wrote, they're also consistent with better pitching performance. Which is good, right?


  1. Holy cow, Casey Janssen is criminally underused. He's a team leader in all four of those categories. I actually thought he was just one of those guys I liked even though I probably shouldn't, but he's put together a really impressive season so far.

  2. To clarify, obviously by "team leader" I mean one of the best on the team, not necessarily #1.

  3. That his K rate is so high kind of blows my mind.


  4. I was going to say, perhaps this should change the way I look at Casey Janssen. And I think it will. Fantastic post, Drew.

    So, next year, Purcey replaces Downs, Roenicke replaces Frasor, and we pick up the option on Gregg. That gives us middle relief options in Janssen and Camp, and we can find some chump (see: Tallet, Brian) to be the long man.

  5. Eyebleaf: I love Tallet (how can you not with that facial hair and that fashion independence?), but I would also love to see him blowing leads for somebody else next year.

  6. I wonder how appealing Gregg is to other teams right now. He's a Type B so if the Jays let him go, they get a supplemental pick, but if he actually has any trade value, it might be prudent to pick up his 2011 option with plans to trade him in the offseason.

    As frustrating as he's been at times, he's actually put up one of the better seasons of his career this year. After 2008 - a year in which he saved 29 out of 38 games with a 3.41 ERA and a 1.282 WHIP - Florida traded him for a pretty decent-looking pitching prospect, Jose Ceda. This year, his ERA and WHIP are almost identical to his 2008 numbers, but he's saved 27 of 31 so far -- on pace for more saves and less blown saves -- and he's striking out a lot more batters as well.

    I know AA really values the draft as an opportunity to get high-ceiling players, but you've got to believe that if they can get a guy like Jose Ceda -- with a respectably high ceiling himself, while being closer to the majors and more established than a draft pick -- it might be worth picking up Gregg's option and exploring the market.

  7. Tallet is a non-tender candidate.

    Gregg, Camp, Janssen, Roenicke, Purcey, Carlson Mills/Richmond is a decent bullpen.

  8. Yeah, I think I'd prefer somebody like Richmond or Rzepczynski or Sean Hill as the long man instead of Tallet. They'd all be cheaper, at least. And the fact that I've always liked Tallet has made it especially tough to watch him struggle all through this year.

  9. Ty, every time someone mentions Rzep as a long-man I can't help but feel that it's a bit of a waste. Tallet gets mop-up duty because who cares who's pitching then, but Rzep (I think, at least) is a pitcher who can perform better than that.

  10. I think he's better than that too, but if Stewart or Drabek steps up to take the fifth spot in the rotation next year, where else do you put him? I'd rather have an overqualified swingman than an underqualified one. Look at the Rays -- their long man is Andy Sonnanstine, a guy who was also bumped from the rotation by more talented arms and as a result has been a very solid long reliever/spot starter.

  11. Tallet is GROSSLY overpaid at $2 million for the job he's doing. I'd gladly put the soft-throwing Mills in that position. I like to think that RZep has more upside, too, as our potential fifth starter, than to be doing long-man duty next season. But you're right, it depends on the Drabeks and Stewarts of the system. RZep could hold down a spot in the pen until Drabek, if he's our 5th, hits his innings limit. Then RZep jumps in.

  12. I need to look at this again when I'm sober.

    ...or more stoned.

    Can't figure out which one it is yet.

  13. Great piece (as always), love this blog. I recently wrote something on Morrow at my blog, definitely a solid add to this solid read. Check it out if you have time.


Send forth the witticisms from on high