You may have thought it was going to be different, but deep down, you knew the outcome wasn’t going to be favorable. Again.
PECOTA Day was Tuesday, when Baseball Prospectus unleashes their projections. Last year, those projections caught the wrath of Royals fans with an estimation of 72 wins. Bulletin board material! Surely, the best record in the American League in 2015, coupled with a second consecutive trip to the World Series – this time with a championship – would find the Royals in a better frame of projection, right? It turns out, computers just don’t give a damn.
PECOTA says 76 wins and last place in the AL Central.
That makes the Royals three for three in coming in at under .500. In addition to PECOTA, ZiPS and Steamer both have the Royals finishing in the second division, well out of reach of the postseason. While all projection systems have their strengths and weaknesses, one particular system doesn’t tend to separate itself from the pack. And while it’s too early to call their 2016 projections failures (pitchers and catchers haven’t even reported yet) we can certainly get the ammo ready.
These computer models don’t like the Royals. Fact. Last year, that negative sentiment was shared by the national media, who routinely partake in the exercise of making postseason predictions in April. This year, I have a feeling there will be more of a disconnect between the machines and the writers. Why? Because individuals can’t avoid but to pick with their heart, their gut, and their own built-in biases.
Let’s take the Royals from last year as an example. Heading into the 2015 season, the staff at ESPN made their predictions and of 15 individuals, only two had the Royals winning the Central. And just two more had the Royals winning a Wild Card spot. That’s four out of 15 picking the then defending American League champs of simply returning to the playoffs. Drill down a little further none of the four who predicted playoffs for the Royals thought they would return to the World Series.
Why? If I may be so bold as to remove my Royals hat and place my dime store psychologist cap upon my dome, it’s because after 29 years of futility, there wasn’t a groundswell of support for the Royals to continue to capitalize on their success. They were too new, rank outsiders who happened to rally from the brink in a one-game Wild Card contest, and rode their good fortune 90 feet away from tying the seventh game of the World Series. In other words, they caught everyone off guard and there was considerable doubt they could pull the same trick again. Understandable. Besides, the Detroit Tigers had won their fourth consecutive Central Division title that year. Even though they had been pushed aside by the Orioles in their division series, their pedigree meant they remained the team to beat. The confirmation bias held that the Tigers owned the division and were the leaders until someone got it together to knock them from the top of their kingdom.
That leads to another facet to these preseason predictions and that was the fact the Central last year was the home to the Hot Team. The Hot Team is the annual phenomena where a club gathers enough support for one reason or another and become the sexy pick. The dark horse. It’s the opposite of the confirmation bias. In other words, the Tigers couldn’t possibly continue their hegemony, so let’s look to see if an up and coming team has the pieces to pull the surprise. Only, because everyone is picking them, that’s not really going to count as a surprise. Welcome to the expectations saddled upon the 2015 Cleveland Indians. Indeed, even though the Tigers were the defending champions, the majority of the ESPN panel tabbed the Indians for AL Central glory.
This year, I’ll assume there isn’t much doubt behind the Royals. They now have the benefit of having national (i.e. a lot of ESPN) exposure behind them. Writers and broadcasters who will make the predictions are familiar with the team, and with the core intact, the individual players as well. The familiarity harbors a certain sort of selection bias. The national media is so familiar with the Royals, their individual stories, and obviously their successes, they will strongly lean to favoring Kansas City in their preseason predictions. The Royals are a likable team and with mostly everyone returning, they’re easy to like again. And again.
Yet last year’s success isn’t enough. The Royals needed to make an off season free agent splash and they did by re-signing Alex Gordon. That keeps them on the radar of the national media, not just because they retained one of their own, but because they signed one of the consensus top 10 available free agents. For the Royals, this is huge. They are financially relevant. Their payroll is middle of the pack. They have national credibility. There is no Hot Team in the Central this year. The Tigers traded a couple of stars and replaced them with lesser players. The White Sox are, despite what USA Today thinks, still the White Sox. The Twins are a couple of years away as their young talent marinates much in the same way the Royals young talent did in the first part of this decade. And the Indians… Haha. You don’t burn the national media and get a second chance.
In six weeks time, when the prognostications are made at ESPN, I would suspect at minimum 10 of the 15 will fall for the Royals to win the Central. The rest will pick them to win a Wild Card. Most will peg them to advance to the World Series. You will be happy.
On the other hand, computers are cold bastards. Circuits and tubes and microchips. They harbor no such biases. Nor do they play favorites. They parse the facts, formulate, and crank out a conclusion. Damn your emotions.
Computers don’t get defense. They don’t get bullpens. They don’t get heart. They don’t understand soul. And they certainly can’t comprehend joie de vivre. Such is PECOTA, who on Tuesday impassively hung 76 wins for the Royals in 2016. What. The. Hell.
Should you be surprised? Probably not. This is largely the same group of players who were projected to win only 72 games the season before. I’ve found that projections seldom seem to buy in to individual breakout performances or improvements. So while Dayton Moore and Ned Yost can contend that “everyone will get better,” projections don’t hear those declarations so much. Hell, you’re lucky PECOTA popped for 76 wins this seasons.
Why are the projections so wrong when it comes to the Royals? As Sam Miller pointed out at BP, the systems don’t ignore the aggregate. When looking at Wade Davis, we see him as a cyborg who features an upper 90s fastball and a deadly cutter. The computer sees Davis as a failed starter. Harsh, but true. The two most recent years on his resume accounts for plenty, but there are still three to five years of struggle behind that. It happened. So when you have a player like Mike Moustakas who suddenly decided to hit the ball to all fields, how does a computer handle that? With skepticism, that’s how. Last year, Moustakas posted a .291 TAv, which was a career high by nearly 40 points. This year, PECOTA projects a .254 TAv, which is one point higher than his career average. (TAv is “True Average.” It’s a measure of total offensive value scaled to batting average. Adjustments are made for park and league quality, as such the league-average mark is constant at .260.) Then you have Eric Hosmer. One year he’s quite good. The next year, he never really fires on all cylinders. How does a projection system handle that? By splitting the difference, but leaning more toward the most recent outcome.
The previous paragraph is just a long-winded way to say you can explain the projections away, especially if you find the disagreeable. Yet the projections can sound the alarm. In the case of the Royals, this is far from a complete team. Sure, it’s felt like one the last two years, but it’s difficult to ignore the abyss at second base. Or the platoon in right. And the poor production behind the plate. While the raw number may be surprising, it’s understandable why the projections dislike the Royals again.
So why pay attention? Let me throw that question back at you: What else are we going to talk about in mid-February? It’s a diversion. Something to kick around before the games start. Something to prime the pump to get us ready for the season.
We’ve been through this drill before. There’s value to be found in the projections, but you don’t want to hang your hat on them. We know that the computers are inexact and the prognosticators deal with their own sets of biases. Neither one is perfect, and thankfully so. If we knew the outcome of the season before the season, exactly how would that be fun. The joy is in the journey that often uncovers the unexpected. Take those projections and predictions, file them away, and let the season begin.