clock menu more-arrow no yes mobile

Filed under:

How accurate are community playing time forecasts for hitters?

With the release of this season's Community Playing Time Forecast Survey at Tom Tango's site (go complete the survey now), it seems like a good time to look back at last season's forecast for the White Sox. While I don't know for sure, I suspect many of the entries came from SSS readers, as we link to it every year and an uptick in entries immediately follows. In any event, the purpose of the survey is to have people who are knowledgeable about the teams in question project how much playing time the players on that team will get. You know, the wisdom of the crowds theory. The idea is to then compare it to the Algorithm to see whether the fans are better at projecting playing time than a Computer. Here, I'm just going to compare the fans' projections to the actual results.

One thing that jumps out is that the aggregate innings pitched projection was pretty darn good: 1471 IP projection vs. 1460 IP actual. The plate appearances projection was not so good: 6673 PA projection vs. 6159 PA actual.

We'll get into individual players after the jump.

Since it appears the hitter projections weren't that good, I'm going to start there, first with the 2011 regulars.

Player Projected PA Actual PA Projected Games Actual Games
Ramirez 618 684 153 158
Beckham 599 557 150 150
Pierre 588 711 148 158
Dunn 583 496 147 122
Rios 585 570 147 145
558 639 142 149
Morel 509 444 133 126
Quentin 484 483 128 118
Pierzynski 486 500 128 129

    The misses are Alexei Ramirez, Juan Pierre, Adam Dunn, Paul Konerko and Brent Morel. For Ramirez, his games played were understated by five (well within the margin of error) but, even crediting him for those, the projection still would have fallen short - so we're looking at a quirk of the survey. Pierre's projection better illustrates that quirk, so I'll discuss him next.

    The Pierre miss is understandable to a degree. He was the leadoff hitter and, therefore, would get more plate appearances than anyone. But the survey doesn't know where in the lineup someone will bat so when it translates games to plate appearances it would spit out the same PA number for a leadoff hitter and #9 hitter who are projected to play the same number of games. Pierre and Alex Rios, for example, were projected for the same number of games and, therefore, the survey gives them the same number of plate appearances. But that doesn't compute because Rios usually batted 6th. Pierre was also projected to play ten fewer games than he actually did. I guess that was some wishful thinking by fans.

    The reason for the Dunn discrepancy is obvious: no one expected him to have such a terrible season that would necessitate benching him frequently.

    Konerko's projected games was off by seven but, even if one were to credit him with those, the projection still would have fallen rather short of his actual plate appearances. So this was a mix of the quirks of the survey and Konerko being a bit more durable than expected.

    White Sox fans were confident that Morel would play a bit more regularly than he did. But again we're looking more at a survey quirk for a guy who batted at or near the bottom of the order.

    Now for the hits. Carlos Quentin's projection was almost dead-on for plate appearances, though it overstated his games played. White Sox fans, from experience, knew he is a china doll and adjusted accordingly.

    Gordon Beckham's projection was also pretty close on plate appearances. The discrepancy again is the survey not quite getting the games to plate appearances translation correct as the games projection was spot on.

    Rios' was quite close, as well. I'm not sure if we should give credit to White Sox fans for guessing that Ozzie would play a terrible performer as much as he did. I imagine the projection assumed he'd get that much playing time because he actually deserved it.

    A. J. Pierzynski's was about as good as you can get. White Sox fans correctly projected that the slight reduction in games played that he had in 2010 compared to prior seasons would continue in 2011.

    Now to the bench players and call-ups.

Player Projected PA Actual PA Projected Games Actual Games
Teahen 300 177 89 78
Vizquel 270 182 82 58
Milledge 255 4 78 2
Castro 288 75 71 23
Viciedo 154 113 51 29
Lillibridge 102 216 36 97
Lucy 92 11 33 6
De Aza 80 171 29 54
Flowers 71 129 26 38
Danks 40 0 15 0
Escobar 31 7 12 9
13 15 5 11

    At first blush, this looks pretty bad. But the standard deviation for these is much higher, meaning there's a wider range around these numbers. Of course, you're not going to be able to defend Lastings Milledge on those grounds. Fans thought he'd play a much larger role - perhaps taking away some of those plate appearances that Pierre ended up getting. Whoops.

    Mark Teahen was also expected to play a larger role [note: those 'actual' numbers include his Toronto numbers]. He spent less time as an occasional starter and more as a pinch hitter - something at which he wasn't good, as his 4 for 26, 1 HR, 1 BB line illustrates. Like Milledge, who didn't even bother to try to get a major league job, Teahen won't be playing in the big leagues on opening day this season.

    Brent Lillibridge also seized some of those plate appearances that were projected for Milledge. His surprise performance demanded more playing time and he got it.

    Fans also thought that Dayan Viciedo would be called up a lot sooner than he was. Again, one can't always accurately project the irrational things people will do. Attempting to project that irrationality is probably why fans assumed Veteran Vizquel would get more playing time than he did.

    Alejandro De Aza, on the other hand, was called up sooner than fans thought he would be and started a lot more.

    Ramon Castro, of course, got hurt and his playing time went to Tyler Flowers and, to a lesser extent Donny Lucy. Fans projected White Sox catchers to collectively play more than was possible. Presumably, fans knew that Lucy could do the impossible and voted accordingly.

    I think the take-away from this is that the fans did a pretty good job of projecting the games played by the regulars in 2011. If one is trying to translate that into plate appearances, one should tweak Tango's community projection a bit and take into account where in the batting order a player will bat. Unsurprisingly, for the non-regulars, the projections were a lot less accurate.

    I'll take a look at the pitching projections over the weekend.