I’ve been publishing projections in some kind or different since 2019, making painstaking enhancements to my course of alongside the way in which. To borrow an expression from Dan Szymborski, my projections have now reached a degree of “non-craptitude” such that I’m content material — although no projector is ever actually content material — to share them with you right here at FanGraphs. This text introduces OOPSY, your pleasant neighborhood projection system.
OOPSY goals to summarize all the info you see on a participant web page — an entire slew of element statistics from completely different years, leagues, ranges, and groups, compiled at completely different ages — in an try and make it simpler to guage gamers. I’ve at all times discovered it troublesome to account for all of this info in my head with out the assistance of a projection system, and now I’ve one.
Like many fashionable projection techniques, OOPSY takes MARCEL as a place to begin, including methodological improvements to account for added complexity (MARCEL tasks all rookies to be league common, for instance). OOPSY makes use of its personal strategy to account for all the typical components captured by fashionable projection techniques: league scoring environments, growing older results, main league equivalencies (with inspiration from Clay Davenport) to account for (minor and main) leagues and ranges that boast differing high quality of competitors, park results (each minor and main league), historic efficiency weighted by recency and, maybe most significantly, regression to the imply, with statistics topic to extra random variance regressed extra closely. As an alternative of regressing each participant to the identical imply as MARCEL does, OOPSY regresses gamers to completely different means primarily based on their chance of creating the majors, which is assumed to be a perform of their age relative to degree (primarily based on historic knowledge), with complex-level gamers regressed to a worse imply than Triple-A gamers, for instance.
Taken collectively, my strategy makes it simpler to match participant efficiency at completely different ages, from completely different leagues. It has made the peak (late-20s) model of the projections an insightful strategy to rank prospects, as an example, producing rankings that align pretty nicely with conventional trade prospect lists. (You will discover this offseason’s RotoGraphs projection-based prospect rankings for hitters right here and for pitchers right here.)
OOPSY goals to maintain up with the trade’s unending onslaught of information by incorporating vital new metrics as quickly as is possible. For pitchers, OOPSY accounts for Stuff+ (the system makes use of revised Stuff+ figures, which might be accessible at FanGraphs quickly); you will have seen Eno Sarris referencing these projections beneath the moniker of “ppERA” at The Athletic since 2023. Right here I’ll observe that whereas Stuff+ strikes just a few projections quite a bit, the normal statistics proceed to weigh closely within the pitching projections. OOPSY additionally accounts for velocity and the same old element statistics, equivalent to strikeout price, stroll price, and groundball price.
Final 12 months was my first 12 months (quietly) publishing a slimmed down model of my hitter redraft projections primarily based on conventional parts. OOPSY now accounts for the brand new swing velocity knowledge revealed at Baseball Savant, in addition to barrel price per batted ball occasion, since barrels have constantly been discovered to be the finest indicator of high quality of contact (wOBAcon) ability. OOPSY additionally accounts for the same old hitter element statistics, e.g., Okay%, BB%, HR%, etcetera. The standard parts and barrel price are the muse of the hitting projections, however swing velocity does have a considerable affect on just a few gamers. For example, Giancarlo Stanton has finished his finest to interrupt the system, with a 20-point projected wRC+ enhance due to his elite swing velocity. Even when the way in which I account for the swing velocity metric overrates Stanton — which is certainly doable — early indications recommend it helps enhance the projections general (e.g., swing velocity helps predict offensive efficiency when splitting 2024 into two halves). As extra knowledge turns into accessible, the projections will have the ability to seize the affect of swing velocity extra exactly.
For each arms and bats, incorporating extremely dependable indicators of expertise — like Stuff+, barrels, and swing velocity — may clarify why OOPSY seems to be bolder on the extremes. For example, OOPSY tasks Aaron Decide for a 199 wRC+ in 2025, which is analogous to the place he has landed over his previous couple of monster seasons, however might be extra bullish than you’ll see him projected by different techniques. OOPSY is equally extra-bullish on Yordan Alvarez, Juan Soto, and Shohei Ohtani (observe: OOPSY and Steamer assume comparable league averages). For arms, OOPSY is relatively bullish on the ERA projections for its high two starters, Paul Skenes and Tarik Skubal, and tasks 14 relievers for a sub-3.00 ERA in 2025, in comparison with six for Steamer.
OOPSY doesn’t seize the next components, though I hope to ultimately add these in: novice efficiency, international professional league efficiency (DSL, KBO, NPB, and many others.), and minor league Statcast knowledge. OOPSY borrows defensive worth projections from Steamer and enjoying time projections from the FanGraphs Depth Charts (I like that these are available and up to date usually). For what it’s value, I don’t a lot care which enjoying time projections you utilize with my price projections, even when distinguished annual opinions of projections could ultimately decide me for it.
Whereas my hitting and pitching projections have held their very own relative to different distinguished techniques by way of forecast accuracy, I don’t declare that my system is revolutionary, nor that it’s essentially the very best among the many varied options — not that you must imagine me if I did, given the tendency of forecasters to search out proof confirming the prevalence of their very own techniques. Projection techniques are the results of a whole bunch of subjective choices by the analyst, with many alternatives to do issues barely (or very!) in another way. For example, I exploit my very own taste of growing older curves and main league equivalencies, and embrace variables like Stuff+ and swing velocity that won’t but be captured by different techniques. Methodological variety could also be why averaging the distinguished projection techniques tends to be an efficient strategy.
OOPSY is a faux acronym, an homage to creating errors whereas enjoying the infield rising up, and to creating errors extra usually, the fallibility of humankind. It goals to remain humble, to be clear about the truth that projection techniques are constructed upon completely different assumptions. My hope is that it provides one other worthwhile perspective to the analyst’s toolkit — you may at the least take consolation in realizing I depend on it closely for my very own baseball-related choices.