How accurate are community playing time forecasts for pitchers?

On Friday, I took a look at how the fans did in last season's community playing time forecast for White Sox hitters. Now it's White Sox pitchers turn. And at the end of this post, I'll apply some of what we learned from the small sample size of last season's survey, to the projections for White Sox players this season.

As I noted in the last post, in the aggregate, fans did a good job projecting pitching: 1471 IP projection vs. 1460 IP actual. Let's see if that bodes well for accuracy on individual pitchers. First up, the starters:

The White Sox did a somewhat unique thing in the first half of 2011, experimenting at times with a 6-man rotation. The departure of Edwin Jackson at the trade deadline also opened up some starts for other pitchers [Jackson's White Sox IP is listed first in the 'actual' column]. Outside of this group of six, Zach Stewart (8 GS) and Dylan Axelrod (3 GS) were the only others to make a start. Neither of these players were an option in the White Sox survey (Stewart was an option in the Blue Jays survey, where their fans projected him for 20 IP).

The actual results for Mark Buehrle, Jackson and Gavin Floyd were all within the standard deviation of their projections, so the fans did a good job with them. Those projections may not have been all that hard to get close to right, since they were essentially for full, healthy seasons. The projection for Jake Peavy was, in my estimation, a bit more difficult to get right - and the fans pretty much nailed it. The obvious miss was Philip Humber. With him we're again seeing the difficulty in projecting break-out performances by players without much of a track record.

Now the relievers:

The pitchers who got relief innings other than those on the survey were the aforementioned Stewart, the trade deadline acquisition Jason Frasor, phenom Addison Reed and surprise Hector Santiago.

Starting at the top, Tony Pena was a big miss because of his early season elbow injury. Because of his 3 starts and 100.1 IP in 2010, fans were expecting something of a repeat, likely with him getting some starts when Peavy was unavailable. A lot of his innings went to Humber.

After Pena, the next highest innings projection went to the five expected stalwarts of the bullpen - Chris Sale, Jesse Crain, Matt Thornton, Sergio Santos and Will Ohman.

The projections for Sale, Crain and Ohman were all within the standard deviation. Thornton fell well short of expectations. Perhaps fans were expecting a bounce-back to his workload from 2010 but the projection of 74 IP was 2 IP more than his career-high from that season. That was probably a dubious prediction for a 34 year old, particularly one who was slated at the start of the season to be the closer.

Santos was always going to be a tough projection. He pitched his first full season as a pitcher in 2010 and his 51.2 IP represented judicious usage by Ozzie Guillen. This can probably be said for Santos in particular but perhaps also for the others who were expected to be "full-time" relievers: on the survey, the range one must choose for those is 60-99 IP. If everyone piles their votes into that category, the survey will spit out a higher innings total. Considering there are few full-time relievers (i.e., those who don't make spot starts) who get above 75 IP, that range might not accurately portray the expectations fans have for those players.

Jeff Marquez and Lucas Harrell were expected to be the minor league starter depth. Marquez was claimed by the Yankees before he could pitch for the White Sox and Harrell only made a few appearances before being claimed by the Astros. If you take their projected innings, as well as Pena's, and combine them with Humber's, you pretty much get to the actual IP of Humber. Or you can take their projected IP and get to Stewart's starting IP. Marquez and Harrell illustrate how quickly the minor league depth chart can change.

Gregory Infante, Jhonny Nunez and Charles Leesman also show how that depth chart can change on the relief side. None of them pitched for the White Sox in 2011. Instead, Jeff Gray, Josh Kinney and Shane Lindsay - three guys with 0 IP projections - basically ended up with those innings. Brian Bruney was one of the last cuts in spring training so fans thought he'd get called up at some point. We're again seeing the difficulty in figuring out how to allocate playing time to minor contributors.

Applying what we learned to this year's forecast

This year's survey is ongoing but, after it's gotten about 15 ballots, the projections don't fluctuate very much. So let's see if there's anything interesting in them. The aggregate plate appearances projection is lower than last season's survey and closer to what will be reality. The aggregate innings pitched projection is again basically dead-on what will be reality.

We again see Alexei Ramirez leading the way but, based on last year, we may want to adjust those plate appearances upwards a bit to compensate for the games to plate appearances translation. He got 4.33 PA/G in 2011. If we apply that to the current 152 games projection, his PA comes out to 658 PA instead of the survey's 613 PA. Similar adjustments should probably be made for most of the regulars to more accurately reflect where they'll hit in Robin Ventura's lineup.

We see a projected reduction to 124 games for A.J. Pierzynski, probably because of age and/or the presence of Tyler Flowers, . The fans did pretty well last season with him. Tyler Flowers' 81 game projection certainly syncs with that. Non-Pierzynski catchers played in 67 games last season, so we may be seeing an expectation that Ventura will substitute in Flowers late in games or otherwise get him into the lineup.

Despite a poor rookie season, Brent Morel garners a projection of 145 games, up from a projected 133 games (and an actual 126 games) in 2011. That's a big vote of confidence from fans, who may believe his September was a sign of things to come.

And despite being shut-down in their 2011 projection of Jordan Danks getting a call-up, fans are doubling down on him appearing in a 2012 White Sox uniform and the projection of 20 games suggests an expectation of more than just a September call-up.

As for the starting pitching, Danks and Floyd are expected to shoulder a full season of starts. Humber gets a somewhat lower projection. Sale's projection of 153 innings is in line with what the White Sox have intimated he will be limited to.

The most interesting projection is Peavy's. Fans expect much more from him this season, with an expected 148 IP. That translates to about 24 starts, a number Peavy hasn't even approached since 2008.

Regarding the other pitchers, the overall projections showed that fans expected Stewart and Bruney to claim the last two bullpen spots. Since Nathan Jones, not Bruney, claimed a spot, the collective wisdom of the fans failed there. Axelrod was expected to be the minor league starting pitcher depth.

Interestingly, despite the presence of an addition lefty in the bullpen, the fans expect Ohman's role to slightly increase.