Last year, we decided to do season previews a little bit differently, and instead of running down each individual team, we previewed the league by position. We liked the format, so we’re doing it again this year. For those who didn’t see the series the first time around, let me borrow from last year’s introduction:
This is only looking at the upcoming season and doesn’t account for potential long term value – we’re just concerned with what each team may get from a given spot on the field this year…
The fun part of comparing teams at a given position is that we’re not limited to just looking at one player, but can compare the expected production of an everyday guy against a left/right platoon, or we can note what a team should expect from giving a stop-gap two months of playing time before they call up their top prospect in the early summer. Few teams get an entire season’s worth of playing time at a position from one guy, so by using depth charts to create an expected playing time matrix, we can give a more thorough evaluation of what kind of strength an organization has at a given spot.
We’re following the same structure as last year, but we’ve also made some improvements that should upgrade this year’s version. To create the data that drives the actual placement of each team within each position, we’re combining a few resources to give us a custom projection that I think could end up being quite an interesting forecast. For playing time, we author-sourced depth charts for every team, handing out playing time at each spot based on the best information available at this point in time. We then prorated the ZIPS and Steamer projections to those playing time levels and averaged– with a straight 50/50 split — those forecasts, then ran those inputs through our WAR calculation.
It might sound a little convoluted, but the hope is that we’re taking the things that projection systems do the best and complementing them with playing time projections informed by updated injury information and the kinds of things that humans can know by reading the paper and following a team’s moves with a close eye. Because it’s a hybrid of multiple inputs, and because these are positions specific, these projections aren’t going to match any of the ones you’ll find on a player’s page.
For guys who play multiple positions, you might think they look low, but keep in mind that his value is being split across multiple posts. This system doesn’t hate Ben Zobrist – it just isn’t going to reflect all of his value within any single post. The author of the second base post also doesn’t hate Ben Zobrist, by the way — the data was given to the writers, so don’t get mad at them if you don’t like the placement of your particular team on a given list. The rankings are essentially based on an algorithm, and thus don’t reflect any particular bias against your favorite team. Really. We don’t hate your team. I promise.
Certainly, there’s no way to know exactly how playing time is going to be distributed. We don’t know when the Rays are going to call up Wil Myers. We don’t know who is going to blow out their knees in April and miss most of the season. These are hopefully educated guesses that will probably end up being wrong, but might be less wrong than automated playing time projections. And, by combining two of the better projection systems out there for the rate stats with those playing time forecasts, the hope is that we can somewhat decently reflect what each team has in the organization at a given position. Even if the Opening Day starter is terrible, if they have a legitimate prospect on the door step, this system will reflect that there is some hope for the position at some point during the year.
For the most part, we’ll be publishing two parts of this series each day over the next week and a half, though the pitching staff posts are so ridiculously large that we might end up splitting those into thirds. When the series is over, we’ll also recap the totals and give you a look at the overall picture by team, and hopefully you’ll gain some insight into each team’s strengths and weaknesses along the way. We hope you enjoy this exercise. It’s a decent amount of work, but we find the results pretty interesting, and we hope you will as well.