Birchwood Brothers 7.1: Let’s Pretend

We were going to write about how all the various statistical projection systems that people, including us, use (or, in most cases, including ours, borrow) to project statistics produce pretty much the same result, in terms of suggested draft position or dollar value, because by definition all projection systems base their projections on what the player being projected has done in the past, and everyone has access to the same statistics from seasons past. And then we were going to note that the most important thing about any given projection by any given projector is playing time, so that the projections for established players who are likely to keep their jobs throughout the season are pretty similar, whereas the projections for non-established players vary more widely. And then we were going to explain that we don’t have much to say about established players that other fantasy writers don’t say, and in fact we probably have less to say, because by and large they’ve looked at at least as much data as we have, and often more. And we were going to explain further how our thing isn’t to project our own performance stats but rather to project playing time, in the belief that, as Bill James, the Odin of Sabermetrics, and many others have said, if you get the playing time right, most of the time you’ll get the projection right.

And then we were going to say that we don’t actually project playing time per se, but rather go trolling (in the good sense) for players who, for whatever reason, aren’t projected to be strong candidates for significant playing time, and who are therefore inexpensive, sometimes visibly so, in your draft or auction, but who we think have a chance to play more. (Although sometimes we just think the projections are misguided.) And then we were going to say that our self-appointed mission, which we pursue with a relentless passion that borders on lunacy, is not just to identify those players for our own competitive purposes, but also to share with our readers the results of our fluorescent diving into the Challenger Deep of utility infielders, grade-B prospects, setup guys, fifth starters, hitters with speed but no power, hitters with power but no speed, and backup catchers. But then we discovered that when you translate the statistical projections of the two state-of-the-art Fangraphs projectors who generously share their projections with the world—those being the ATC projections of Ariel Cohen and the BAT-X projections of Derek Carty–into dollar values or draft positions, there really are significant differences between them, and then we found that the differences themselves vary to some extent according to which system—Standings Gain Points or Z-Scores, which we can’t stop to tell you about now—you use to calculate the value of any given package of statistics, which in itself is kind of interesting, but it left us without anything really to say, since our entire first article was going to be about how all the values you come up with are about the same, and…

But we have perhaps opened too abruptly. Allow us to start over and (re)introduce ourselves. We are the Birchwood Brothers, two genuine, increasingly geriatric siblings, who, despite a significant difference in age, bonded over baseball stats (and baseball, by the way) many decades ago and have never unbonded. We’ve been Fangraphs contributors since 2015, took a sabbatical last year, and have now returned, refreshed and clear-headed, to opine and entertain (ourselves, if no one else). We play fantasy baseball as an excuse to write, and write as an excuse to play fantasy baseball. We frequently do the kind of perverse thing we did in the paragraph preceding this one, to the delight of a few and the annoyance of many.

Why put up with us at all? Because—if we may boast for a moment—we sometimes know what we’re doing, as evidenced by our performance in the three years we’ve played in The Great Fantasy Baseball Invitational (it’s been around for four years, but we missed the first one), a competition confined to members of what is ludicrously called the fantasy baseball “industry”—writers, bloggers, podcasters, Twitterers, whateverers–and widely recognized as the place where the Elite Meet to Tweet and Compete. We finished 7th of 315 teams in 2019 and 14th of 435 teams in 2021—and, uh, 57thh of 390 teams in 2020. Only five other players have managed two top-20 finishes in those three years. In fact, our average finish in those three years is 26th, which is the second-highest average finish, behind only Rob Sherwood of Fantrax, whom we hope we don’t face in our TGFBI league, and his 25.33 average finish among the 214 fantasy junkies who participated in TGFBI each of the last three years. And if you’re looking at average percentile finish, the story’s the same: ours is 93.31%, behind only Rob and his 93.33%.

Anyway, if you’re back from looking at Fantrax: As you’ve probably noticed, it’s 2022, and we’re going to pretend that, from a Fantasy Baseball standpoint, nothing’s wrong–that’s why it’s called “Fantasy,” of course– and that (1) the Omega Variant, or whatever they wind up calling the next one, doesn’t render this entire discussion, as well as these two entire writers, moot; (2) the baseball season starts approximately on schedule; (3) both leagues use a DH, but otherwise there aren’t any rule changes that can affect Fantasy drafts; and (4) all the available free agents have been signed. This last is of course demonstrably false, and it messes us up considerably, because, as we say, our specialty is finding guys that no one but us thinks might play a lot, which is a function of whom they’re competing against for a job. So, for example: Tyler Wade is listed, at the moment, as the Angels’ starting shortstop. If that remains the case, we think that Michael Stefanic has a good chance to supplant him. But if we draft Stefanic late in our ultra-deep draft, and then the Angels sign Carlos Correa, we’ve wasted a pick.

Nonetheless, we will plunge ahead and, over the next couple of weeks, do a sweep through all of MLB and identify the guys who (1) will either cost you $1 or be available in the reserve rounds, and (2) have a decent chance of doing significantly better than the market thinks they will. Right now, though, we’ll report the results of an investigation we conducted into a subject that has long interested us: what is the premium one pays for a hitter with multi-position eligibility?

This was, as you’ll see, a down-and-dirty calculation, but it’s still kind of suggestive. Because dollar values rather than draft position show any differences most clearly, we looked at NFBC draft positions and then used Jeff Zimmerman’s method to convert them to dollar values. We started with the 210 hitters that ATC would draft, using Z-score pricing. We then reduced the sample by (1) eliminating catchers, because the markup due to position scarcity has all catchers underdrafted; (2) taking out DH-onlies Franmil Reyes and Nelson Cruz; and (3) omitting players who were among the top 10 hitters drafted, because the way Zimmerman’s pricing model works, the prices for very early draft choices come out too high. This left 168 hitters. Of those, there were 135 who qualified at a single position. Comparing their Z-score dollar value to their ADP as translated into dollars, we found that they were, on average, “overpriced ” by ATC Z-scores by $1.12.

In other words, relying on Z-scores, which don’t care about multiple positions, you are typically paying about $1 more than the market, which does. Conversely, the eight guys who played two positions (e.g. 2B and SS) that offered you only three positional options (2B, SS, MI) were undervalued by ATC by $0.75, and the more flexible group (e.g. guys who qualified at 2B and 3B, and thus had four positional options, i.e 2B, 3B, MI, CI), were undervalued by $0.51. Overall, the multipositional group was undervalued by an average of $0.57. We know what you’re saying: what about outfielders?

They qualify for five roster spots even if outfield is all they do. Interestingly, if you take out the first two rounds, where ADP pricing is wonky, Z-scores actually overestimate ADP pricing by $1.14, suggesting (at least to us) that people are undervaluing that flexibility of filling 5 positions. This may be because they think that there is no shortage of OFs to get at any given point of the draft. If so, we’re skeptical. There are 90 starting OFs (30×3) in MLB, and you draft 75 of them, or 80%, whereas there are 60 Starting MIs, and you draft 45 of them, or 75%, so you actually need a higher percentage of starting OFs.

In any event: Among multipositional players who played OF and something else, ATC underpriced them by $0.44. Among multipositional guys who didn’t play OF, it was $0.65. But the samples are small enough that we’d just say that there’s no statistical difference. So ATC Z -scores overvalue one-dimensional guys by $1.12 and undervalue multidimensional guys by $0.57. Big deal, right? Except it is. That $1.69 net difference is, in the middle of the draft, about two rounds. If, say, you’ve got a one-position guy whose ADP is 180, that (using Zimmerman’s conversion technique) is about $7.92. Magically bestow a second position on the guy and his value is $9.61, which is an ADP of about 148.

What do you do with this? Well, if you’re using a pricing system that doesn’t attend to whether guys are monopositional or multipositional, upgrade the latter by about $2. And pay special attention—herewith our first recommendations of the season—to Ryan Mountcastle and Jorge Polanco, who are both multipositional and (according to ATC) undervalued by the market.