If you’re a Royals fan and you’re reading this blog, you probably already know that the baseball forecasting system PECOTA recently projected that the 2016 Royals will win 76 games and finish last in the American League Central Division.
And you also probably already know that the same system projected that the Royals would win 72 games last season … and missed by 23 wins.
As KC Star beat writer Rustin Dodd recently pointed out, PECOTA is the official projection system of Baseball Prospectus. Last year, its editor-in-chief, Sam Miller, explained why the system has a tough time with the Royals.
PECOTA struggles with “elite” bullpens, great defense and “any team that can outperform its raw stats.” Miller also said the Royals were “a hard team to decipher in 2015 because many of their core players had not performed well in 2014 or in previous seasons.”
For now, let’s concentrate on that last reason for inaccurate predictions.
If you’re making predictions based on past performances and someone changes his approach — if someone gets better — it’s going to throw off those predictions.
And players do change — at least, the smart ones do.
In 2014, opposing teams were using dramatic infield shifts on Mike Moustakas. He didn’t want to change his game — Moose considered himself a pull hitter — so he stubbornly kept hitting the ball into those shifts, and hit .212 overall.
In 2015, he made a change and began using the opposite field more often, laid down a few bunts and that made some teams quit shifting him — he hit .284.
In 2013, Edinson Volquez pitched for the Padres and Dodgers and his combined ERA was 5.71. According to Volquez, when he pitched for those two teams he was emphasizing pitching away. But when he got to the Pirates in 2014, they wanted him to spend more time pitching in.
That change got results and his ERA dropped to 3.04. With the Royals in 2015, he continued to pitch inside and his ERA was 3.55.
In 2015, Wade Davis was pitching to the Minnesota Twins and had one down, a runner on second base and Joe Mauer at the plate. Coming into the at-bat, Mauer had hit .429 off Davis; the two guys on deck — Torii Hunter and Trevor Plouffe — had hit .167. The answer seemed obvious: put Mauer on, set up the double play and pitch to Hunter and Plouffe.
But Davis pitched to Mauer and struck him out.
When I asked Wade why he went after a guy who had hit him well, he responded by asking when Mauer got those hits. Since moving to the pen, Davis has been a different pitcher — he had more velocity because he wasn’t pacing himself — and he felt he could get Mauer out if he came in on his hands. After he got Mauer swinging early to keep the fastball and cutter off his hands, Davis struck him out with a curve.
The numbers would tell you not to let Wade Davis pitch to Joe Mauer, but Davis had made and adjustment and wasn’t the same pitcher anymore.
If you look at what a player has done in the past and assume he’ll do more of the same, a lot of the time you’ll be right. But if a player makes an adjustment — if a player gets better or worse — those predictions are going to be off the mark.
This is why I avoid predictions: I have no idea who spent the winter improving his game and no idea who’s been wearing out the asphalt in the Taco Bell drive-through lane.
People can be unpredictable, and players are people. If a player isn’t having success, a smart player will adjust, and that’s going to throw off the predictions. But predictions give us something to talk about, so we’ll probably keep making them.
Just don’t take them too seriously.