February 21, 2012

“Gleeman and The Geek” #29: Pitching, Pitching, and More Pitching

This week's episode of "Gleeman and The Geek" is ridiculous and was recorded in front of an audience at Smalley's in downtown Minneapolis, with more beer than usual and several special guests. Topics included a pitcher-by-pitcher look at the rotation and bullpen, bleeping out John Bonnes' filthy mouth, projecting Francisco Liriano's future, arguing Carl Pavano versus Scott Baker, on-air hugs and popping collars, and the new mega-addition to the Twins blogosphere.

Gleeman and The Geek: Episode 29

In addition to the direct download link above you can also subscribe to the podcast via iTunes.

  • ben

    I like the new intro music a lot

  • Jeff

    The whole Pavano thing isn’t hard to figure out. The statistics are there. Let’s use, say, last year’s ERAs (acknowledging that ERA is imperfect and you might want to use more than one year)

    Pavano: 4.30 ERA for 222 IP
    Baker: 3.14 ERA for 134 IP

    and a suitable “6th” starter guy, say, Swarzak:
    5.03 ERA for a career 161 IP (we’ll consider last year to be a bit fluky-good)

    3.14*(134/222) + 5.03*((222-134)/222) =
    3.89 ERA. Which is less than 4.30.

    I mean, you could squabble of the particulars, exactly what stat to use, etc., but the take-home message is that Aaron is right: if you start with 2/3 of a good year pitched by Baker, he’s enough better than Pavano that even if you fill the last 1/3 year with someone replacement level, Pavano’s so barely above replacement level anyways that you’re still better off with Baker’s good innings.

  • JS

    Bonnes gives me a headache.

  • chris

    Pavano’s fWAR the last 2 seasons: 3.1 and 2.9. Baker’s: 2.5 and 2.7. Not sure exactly what all the components are, but you’d have to assume Pavano’s innings have something to do with that.

  • http://fieldoftwins.blogspot.com Shane

    This was actually a good podcast.

  • Jon

    I did a similar analysis to Jeff, but used the Bullpen ERA and FIP to make up the difference in innings between Baker and Pavano. I actually thought Aaron was being a bit obtuse in dismissing the idea that Pavano’s innings are valuable, but it worked out much more in Baker’s favor than I would have guessed…

    Pavano: 222 IP, 4.30 ERA, 4.10 FIP
    Baker: 134.67IP, 3.14 ERA, 3.45 FIP
    Bullpen: 4.51 ERA, 4.43 FIP

    Baker + 87.33 Innings of Bullpen: 222 IP, 3.68 ERA, 3.84 FIP

    I really thought, given how bad our bullpen was, that it would be much closer and possibly even in Pavano’s favor, but nope…

  • tjw

    You can do the same comparison with WAR.

    Scott Baker: 134.67 IP, 2.7 WAR
    Scott Diamond: 39 IP, 0.2 WAR
    Liam Hendriks: 23.1 IP, 0.3 WAR

    Total: 197 IP, 3.2 WAR
    vs.
    Pavano: 222 IP, 2.9 WAR

    In other words, the combined contribution of Baker, Diamond, and Hendriks last year was 30% more valuable than Pavano, despite throwing 25 fewer innings.

    The numbers look even better if you substitute Swarzak’s starts for Diamond’s.

    Baker/Diamond/Swarzak: 221.7 IP, 3.9 WAR
    Carl Pavano by himself: 222.0 IP, 2.9 WAR

    Pavano is a full win worse in essentially the same number of innings.

  • Jeff

    Right, so you could make the argument that Pavano provides more value by himself by virtue of pitching more innings – 2.9 vs. 2.7 – at the end of the day that value is less useful that Baker because there is basically an unlimited amount of Diamonds, Hendriks, and Swarzaks in the system.

  • Alex

    Pavano has had an ERA+ over 100 only once since 2007.

  • toby

    @0:33:55: “You will not find a statistic that does not show that [Carl Pavano was less valuable than Scott Baker.]”

    Wow, Aaron. How about, y’know, WAR.

    It’s bizarre that Bonnes is actually arguing a narrative-version of WAR theory and he generally finds narrative ways to dismiss SABR stuff.

    His mistake (because I think he thinks Pavano is way more valuable than he was – it’s just that Baker wasn’t worth what you seem to imply) seems to be that he underestimates the value of replacement level performance – he thinks scrubs are worse than they are. (It’s worth noting that IF the replacement performance actually available to a team is worse than average replacement performance, “innings eatings” Pavano-style becomes more valuable.)

    The thing is, Pavano WAS worth more than Baker last year, purely by virtue of the massive gap in innings. But it’s close: 2.7 (Baker) v. 2.9 wins per fWAR. Baker throws 175 innings and it flips. If he starts 32 games it’s not even close.

    (Replacement level for starting is probably a bit under 5.50 or so last year – it was 5.63 in 2008 – and that seems about right for Swarzak, etc..)

    $8m is certainly WAY too much to pay.

    Bonnes wants to know what the odds of winning are when each guy starts, which demonstrates that he’s WAY off the mark in estimating the value of 7 innings of Pavano v. 6 innings of Baker. If that were the gap, Baker is the easy choice. Not even close.

    The reason Pavano was (a little) more valuable than Baker isn’t because of 7 innings v. 6 innings when he starts, it’s because of a massive 90 inning gap.

  • John Sharkman

    This all gets rather mathy.