Here’s What Happened When We Tried To Play Moneyball Without Any Money


This is an excerpt from The Only Rule Is It Has to Work: Our Wild Experiment Building a New Kind of Baseball Teamby Ben Lindbergh, a writer for FiveThirtyEight, and Sam Miller, the editor in chief of Baseball Prospectus. (The book was published this week by Henry Holt and Co.) In the summer of 2015, Lindbergh and Miller took over the baseball operations department of an independent-league team in California, the Sonoma Stompers, putting their sabermetric beliefs to the test with actual professional players. As spring training approached, they compensated for their lack of connections, tight budget and even tighter time frame by using statistics to scour the country for overlooked talent.

As proud as we are of our Baptista heist, Sam and I still want to prove that we can find players who wouldn’t have been blips on the Stompers’ screen without us. What we want is a source of talent that the other teams in the Pacific Association aren’t already mining: a good, old-fashioned market inefficiency, like the one the Oakland A’s exploited in the “Moneyball” era when they targeted players with high on-base percentages, or the one the Tampa Bay Rays leveraged years later, when they realized that their opponents’ emphasis on power bats made it easier to sign players with good gloves at a discount. The problem is that inefficiencies like these are increasingly difficult to find.

What sets us apart from our competitors in the Pacific Association, and theoretically gives us an edge, isn’t our meager budget, our nonexistent network of contacts in the indy-league grapevine, or our untested scouting skills. It’s our ability to find the significance in statistics, either on our own or through our relationships with the leading lights of the sabermetric community, who can crunch numbers in extremely sophisticated ways. We could search for wrongly released players by sifting through last season’s stats from the minors and upper-level indy leagues, but that method wouldn’t be the best use of our time. Players who’ve had any recent success at higher levels won’t want to sign with us until they’ve exhausted all their other options, which might mean waiting until midseason. For now, we need to aim lower. Baseball has a caste system, and at our level we’re trafficking in Untouchables.

20150902_140634

The manager’s office in the Stompers’ clubhouse at the end of the season.

There’s only one level of competition that combines a decent statistical record with players who’d probably be happy to hear from us: college. Every June, the thirty major league teams cull the best college talent in the forty-round amateur draft, leaving only the undesirables behind. But MLB scouting directors work with the far future in mind: Although they know that only a tiny percentage of their selections will work out, every player drafted — aside from a few nepotism picks — has a backer who believes he has some shot, however remote, at developing into a big leaguer. We aren’t worried about long-term potential; we care only about what players can do this summer. And because our incentives aren’t aligned with those of the big league teams that went back for fortieths from this buffet before we were seated, we might unearth a few “one man’s trash is another man’s treasure” types: guys who can get outs, or avoid making them, right now, regardless of whether they have the physical tools that allow teams to dream. Drafting college players instead of riskier high school stars was one of the hallmarks of the “Moneyball” A’s, but Oakland owed much of its success to the type of amateur talents who never get near an indy-league team unless their careers take several wrong turns. If the A’s were “a collection of misfit toys,” as Michael Lewis wrote, then we’ll be building a team out of toys that got recalled because they were choke hazards.

Nothing about baseball is as easy as it seems, and finding our statistical standouts isn’t as simple as sorting an official NCAA leaderboard. I email an acquaintance named Chris Long, who worked as the San Diego Padres’ senior quantitative analyst from 2004 to 2013 and has since consulted for other major league teams. Chris, who had no professional background in baseball when he was hired and occasionally clashed with traditional evaluators, was one of the first quants allowed in a draft room alongside scouts and grizzled special assistants, and his numbers-based evaluations helped dictate the Padres’ decisions.

To our relief, Chris loves what we’re doing and agrees to help just for fun, even though he charges big league teams sums in the five figures for the same service. “Performance in college has a fairly strong predictive value to how [players] perform in the minors,” Chris tells us when Sam and I call to learn at his knee. Better yet, he says, some teams are still overlooking that link. “Probably starting in the late ’90s, every year it got a little bit better,” he explains. “It’s still not great. Even the Chicago Cubs, for example, have a very traditional scouting department in terms of how they approach the draft. It’s not like they don’t look at performance numbers at all, but they’re not analytically evaluating the performances of the players and then combining it in some sophisticated way with their scouting evaluations.

“You’re looking for guys that you want to perform immediately,” Chris continues, echoing our thoughts. “That actually gives you more freedom, because you can go for guys that have flaws. High strikeouts, but also hit home runs, for example. Those guys tend not to do as well as prospects, but they’re certainly going to make contributions to the team. Or the scout’s least-favorite player, the short, gritty [batter] that gets hit by a ton of pitches and has a strike zone the size of a postage stamp — those guys are going to be perfect for your team.” His voice, normally low-pitched and nasal, takes on the bright timbre of an infomercial. “Win your independent league with this one weird trick!”

OpeningNight23 copy

Umpire Dean Poteet with Stompers mascot Rawhide on opening day.

Pitchers have a weird trick, too. “I think there’s going to be a place for some junkball pitchers, guys that scouts wouldn’t touch because their fastball velocity is too low, but they did still strike out guys using a variety of off-speed stuff,” Chris says. “Some of those guys are going to do quite well at that level of competition. Put it this way: They did well in college. If your level of competition is at the level of good D1, for example, there’s still room for junkball pitching.”

College stats are difficult to work with, which explains why, until very recently, major league teams all but ignored them when compiling their predraft rankings. College players face dramatically different levels of competition depending on whether they’re in Division I, II, or III, and even within each division there are significant variations from conference to conference. “An average D1 school is going to beat an average D2 baseball team about 70 percent of the time,” Chris says. “It’s not like 98 percent or anything. There’s a bit of luck, and the spread isn’t that huge.” Still, the gap between divisions and schools is large enough that one can get into trouble trying to go on gut feel or surface stats.

On top of the interdivisional differences, there’s enormous variability from game to game: Some hitters beat up on bad weekday pitching that would pale in comparison to the Pacific Association’s, but look overmatched against stronger weekend starters. There’s also unevenness in climate, ballpark dimensions, and playing surfaces. Before players can be compared on an even footing, we have to take into account the diversity in environments that makes raw stats at certain schools far less impressive than the same stats would be elsewhere.

To do his draft work, Chris built a repository of college statistics by “scraping” information from school websites and parsing it into a database-friendly form. By comparing players’ production at each park with their production elsewhere, he can isolate the statistical impact of every home field. And by comparing players’ performance against each opponent with their performance in all other games, he can also assess the impact of each school’s strength of schedule. Apply the appropriate adjustments to each player’s actual stats, and the result is a ranking of every player’s production on the same scale, independent of division and environment. With the confounding effects of location and competition neutralized, we can compare Division III hitters in good pitchers’ parks to Division I hitters in bandboxes, based purely on their play. This should save us from signing a “slugger” who posted gaudy stats against guys throwing garbage in the college equivalent of Coors Field.

“The starting point would be, you really want to identify the best of the seniors that did not get drafted, and also didn’t return to college,” Chris says. He says he’ll send us his adjustments for every school, along with the real prize, rankings of every fourth-year college player from 2013 and 2014. This is our favorite phone call ever.

Listening to Chris, Sam and I realize, much to our relief, that we’ve found an organizing principle for our Stompers star search. Chris’s stats will make the world seem smaller and more manageable, transforming a confusing array of unknown names and unverified claims into objective rankings like the ones we’re used to. We’ll operate under the theory that a player capable of posting elite stats in college could hold his own in the Pacific Association, one of the lowest rungs on the professional ladder. And if the stats insist that someone can play, we won’t rule him out based on body type, facial structure, or fastball speed. We’re the Ellis Island of the indy leagues. Give us your small, your soft-throwing, your huddled middle infielders yearning to play for (almost) free.

Team_Photo_2016

The official Sonoma Stompers team photo.

Although Chris promises results soon, it takes him some time to deliver. That’s understandable — we aren’t paying him, and plenty of others are — but as the days stretch into weeks, I start torturing myself by checking his Twitter feed a few times a day to see what else he’s working on. He’s a sports polymath (emphasis on the math), and he’s on an NCAA volleyball analytics kick. Thanks to a little light stalking, I learn a lot about the best volleyball schools, but Sam and I are no closer to building a baseball team. I send emails asking for updates at what I deem to be socially acceptable intervals. Finally, one of them works. The long-awaited rankings arrive.

Chris cautions me that “99.9 percent of the talent is sucked away in the draft,” but I can almost hear the heavenly choir as my cursor hovers over the file. That remaining 0.1 percent could be the key to the Stompers’ season. I open the list of all players, create another spreadsheet of players who did get drafted, and filter the latter from the former, leaving only the undrafted guys. (When our story gets made into a movie, the spreadsheet-opening montage will make for an exciting scene.) There are hundreds of rows, each of them containing a name, a school, a division, a position, an at-bats or innings total, and a few columns of hitting or pitching stats, both raw and adjusted. Each player receives an “Index” score, based on adjusted on-base percentage and slugging percentage for hitters and those same stats allowed for pitchers (Column R).

After additional begging by me, Chris also sends us estimates of each catcher’s “framing” skill — his ability to catch pitches in a way that makes umpires more likely to call strikes — based on the percentage of taken pitches that are called strikes when he’s behind the plate versus the percentage of taken pitches called strikes when the same pitchers are throwing to different catchers. On the phone, Chris suggests that we “find some senior guy who’s the best framer in all of the NCAA but wasn’t drafted. That would be an interesting guy to invite to a camp.”

He’s reading my mind. “If you could find one who’s left-handed, he might have to get a restraining order against Ben,” Sam says. The majors haven’t seen a left-hander who played catcher as his primary position since before the birth of the American League, and no left-handed catcher has caught even an inning since 1989. The anti-lefty bias seems based more on superstition than sense, with explanations usually citing theoretical impediments (can’t throw to third; trouble applying tags) that don’t stand up to scrutiny. Chris agrees. “I think the bias against left-handed-throwing catchers is pretty stupid,” he says. He’s my statistical spirit animal.

Ben_Kavanaugh

Ben with his clipboard at a tryout.

Every rating Chris sends is based on a smaller sample than we’d like — in many cases, a four-year player’s entire college career comprises fewer innings or plate appearances than a big leaguer records in a single season — but it’s the best we can do. It’s clear that we’re not the first people to look at a list like this. Scanning the top of our leaderboard of undrafted players, we see a number of players who were signed as free agents after the 2014 draft and spent the summer playing for organizations that are known as early adopters of amateur analytics, among them the Cardinals, Astros, and Yankees. This is discouraging in one sense — the college carcass is picked even cleaner than we thought — but encouraging in another: Most of the undrafted free agents played well at minor league levels that are comparable in quality to the Pacific Association. We might be on the right track.

Armed with the names we wanted, our job shifts from ranking to recruiting. Sam and I spend a few hours combing through Google results, YouTube videos, and social-media sites for information on the players with the most impressive stats. At the end of this process, we’ve created another spreadsheet full of players whose coaches we want to call. (Spreadsheet creation is becoming a theme.) Separately, we practice our sales pitches. In our fantasy leagues, it takes us one click to add an available player from the waiver wire, and no one is allowed to turn us down. In reality, we’ll have to talk to human beings and try to persuade them to travel to a place they’ve probably never been, to play for a team that they almost certainly haven’t heard of, for a salary that we’re embarrassed to say out loud. Aspiring pro athletes expect to lead nomadic lives, but this still isn’t the easiest sell.

No matter how persuasive we are as speakers, there’s only one aspect of our project that makes this PR campaign possible: We have the power to make these people professional athletes, with all the cultural cachet and appeal to the opposite sex that this occupation confers. With a few words, spoken like a sacrament, we can give a few young men a line on their résumé that they’ll never remove, an answer to “What do you do?” that makes people perk up. The Stompers are the mangiest mutt of a team imaginable, but a pro team nonetheless. And ballplayers miss being ballplayers: Even two years after graduation, every player’s Facebook photo and Twitter bio is a callback to his college career.

As we work our way down our short lists like political candidates calling potential donors, it becomes clear that college coaches — at least the ones we’re trying to reach — aren’t great at returning calls. I consider trying to catch their attention by claiming to be with a big league team, then decide that misrepresenting myself might not be the best way to persuade players to sign with the Stompers.


Ben Lindbergh discusses his summer with the Stompers on our podcast What’s The Point.

 

Subscribe: Apple Podcasts | ESPN App | Download | RSS | New to podcasts?

Mindful that each passing day might remove someone we want from the market, I bypass coaches who don’t call back and contact some players directly, gambling that they won’t be bad guys. But ballplayers aren’t award-winning communicators, either: Unlike a lot of people in our profession, they aren’t constantly connected to email, and they aren’t notified when someone sends them a tweet. Whenever possible, I call a player’s parents, banking on the fact that if his mother is like mine, he’ll know no peace until he replies.

The more players I have trouble tracking down, the more I expand the search, and the longer my short list looks. I form attachments to strangers in our spreadsheet almost instantly: All it takes is a name, a stat line, and a head shot, and I’m mentally penciling a player into our lineup and announcing his name over the public-address system. If you’ve played fantasy baseball, or even rooted for a real team that’s one upgrade away from a well-rounded roster, you know the feeling of fixating on a particular player: refreshing MLB rumor sites until a deal is done or off the table, or sending several permutations of the same trade to a leaguemate and hoping that one of them will land your white whale. This is the same impulse, turned up to 11. I’m perplexed by the players who aren’t curious or courteous enough to respond, but the close calls are especially agonizing.

Take Andrew Kelley, a 2014 graduate of Grinnell College, a Division III school known for its “rigorous academics and tradition of social responsibility.” (I see nothing about its undervalued athletes, but maybe that means we’re ahead of the curve.) Kelley, who’s fourth on our pitcher list, is listed at 5-foot-7, which is a concern given the sport’s prevailing preference for skyscraping pitchers. But everyone on our list is bound to have some physical flaw. And hey, he was 5-foot-6 as a junior — maybe he’s a late bloomer with 5-foot-8 in his future.

In 48 innings, mostly in relief, Kelley struck out 48 batters and walked only 4. His LinkedIn account says he’s had a full-time job as an “Integration Engineer” since a few months after his final semester, but I dig up his email and message him anyway, just to make sure. His response, which arrives in under an hour, mentions small samples, machine learning, predictive statistics, a previous internship with an expert on the physics of baseball whose articles I’ve edited at Baseball Prospectus, and the fact that he possesses a “pretty good knuckleball.” He can’t quit his job to play baseball, but he feels bad about it. “It would have been awesome to have been able to play this summer and chat about all sorts of statistics,” he says. I come close to shedding tears. This feels like finding out that your biggest college crush, who met someone else and settled down soon after school, would have wanted to date you if you’d only asked her out sooner. Alas, Andrew is too smart and well compensated to be a baseball player. He’s in a better place, and it’s time to let go.

Andrew Kelley isn’t the only one who gets away. He’s not even the only engineer: Our best shortstop option, with the improbable porn-star name Billy Steel, has just joined Northrop Grumman. Then there’s George Asmus, a standout pitcher at Sonoma State with hometown-hero potential. Unfortunately for us, he’s happy in his current role at Triple-A — as in, the American Automobile Association. There’s also Arismendy Nuñez, a well-intentioned but indecisive senior starter at Old Westbury, who strings me along over multiple calls (including a conference call on which Theo makes a passionate appeal) before breaking up with me by text. He says something about family obligations, but odds are he’s just not that into us.

And then there’s the unicorn, the spreadsheet player who has great stuff. His name is T.J. Fussell, and he’s a 6-foot-4, 220-pound right-hander who pitched for Western Carolina, a Division I school. Fussell had a high ERA, but he struck out 81 batters in 64 1/3 innings, and we suspect he got somewhat unlucky. I contact his coach, Bobby Moranda, who gives Fussell as enthusiastic an endorsement as I’ve heard so far. “He is the best!!!” Moranda writes. “He was up to 94-95 with a plus breaker and change. Really should have been drafted!!!” It’s rare enough for us to see any velocity numbers that start with nine, so even after levying the 2-3 mph exaggeration tax, I’m still salivating.

“I’m kind of in a peculiar situation,” Fussell says when I call. He explains that while he “hasn’t left [his] love of baseball behind,” he has enlisted in the U.S. Air Force. The good news is that he doesn’t have to head to basic training until August 25. The bad news arrives a second later: “I’m supposed to be getting married at the end of May.” And there’s even worse news than that: “We had a cruise planned right at the start of June, and it’s for seven days.” He sounds a little uncertain: Supposed to be? Had a cruise planned? Is the call of the mound so strong that he’d think about being a runaway groom? He says he’ll talk to his fiancée and let me know. The next day, he does: Shockingly, she’s pretty attached to the whole “honeymoon” thing. There goes my four-pitch flamethrower.

Fortunately for the Stompers (and my sanity), we don’t always strike out. Our first “yes” comes from Kristian Gayday. (“He’ll be popular on LGBT night,” Theo says.) Gayday, an Indiana native, played shortstop in his final season last spring at Indiana University-Purdue University Fort Wayne, a Division I school where he’s still working as a student assistant. Our spreadsheet says he was the best D1 hitter among 2014 fourth-year players. Not the best D1 hitter among undrafted fourth-year players; the best D1 hitter, period. (In fairness, most true prospects get drafted after their junior year.) He hit .358/.472/.653 with 12 home runs in 166 at-bats, and Long’s adjustments hardly hurt him. And unlike a lot of our targets, he’s 6-foot-3 and 220 pounds, blessed with a prototypical baseball body.

Granted, Gayday’s senior season was the exception: In his first three years, he didn’t hit for power, never slugging above .327. But it still doesn’t make sense that a strapping player with his senior stats and a modicum of defensive ability didn’t get drafted. Before talking to him, I email his college coach, Bobby Pierce. I ask him the questions that the spreadsheet can’t answer. Why the huge improvement in his senior year? Is he a good guy? Can he play defense? (A big blind spot for us: College fielding locations aren’t tracked, so Long’s method evaluates only offense.) They’re all ways of reframing the indelicate question at the root of all our inquiries: “What’s the catch?”

Coach Pierce sets my mind at ease. More than that, he makes me excited, as if we’re party to a secret no one else knows. He tells me that Gayday’s newfound fourth-year “spray approach allowed him to really handle the breaking ball/off-speed really well,” and that he hit at least half of his homers — more than he’d hit anywhere in his first three seasons combined — to the opposite field. That change in approach gives us a plausible explanation for the extreme uptick in production.

Even more encouragingly, Pierce says that Gayday was drawing interest from scouts and advisers until he suffered “severe lower back issues” for six weeks at midseason, playing one day at 50 percent, resting the next day, and pinch-hitting the day after that. “This middle part of the season was when all our local guys came to see him play, and they either saw him play at 50% with below avg draftable tools/skill/performance/ etc, or they didn’t even get to see him play,” Pierce writes. But by the end of the season — after his back had blown his chances — he’d recovered and gone deep in each of his last three weekend series.

Despite Gayday’s breakout, the Mastodons went 19-34, and for them that was a good year. “We’re a small Div I that gets little respect and scouts never come to see us play,” Coach Pierce writes. “We haven’t been very good and I do understand that scouts are very busy and they can’t afford to waste a weekend afternoon on us, but Kristian was plenty good enough to be a 20-30 round guy. If he stayed healthy, he would definitely be in organized baseball.” Pierce also reports that Gayday would fit in fine in the clubhouse.

On the phone, Kristian tells me that his back feels fine. I ask about his plans for the summer. “I was just gonna go to a couple tryouts,” he says. “And if nothing happened, I was just gonna call it quits.” In my eagerness to sign him, I tell him we’ll pay for a one-way plane trip, which I’m supposed to use as a bargaining chip in exchange for a lower salary. My unauthorized largesse works: Kristian consults with his family and commits to sign the contract as soon as Theo sends him a copy. I can’t believe our luck: We’ve signed one of our top targets, and he doesn’t even look like a runt. It seems as if we’ve stolen a march on the majors, using stats to take the long view on a player whom others might have missed because of an ill-timed injury.

Sam seems content once we get Kristian to commit — at least we have something to show for our spreadsheet — but the taste of one transaction makes me hungry for more. I turn my attention to pitchers, tag-teaming with Theo on a series of deal-sealing conference calls. Jeff Conley is a skinny, 6-foot-2 lefty out of Alderson Broaddus University, a Division II school in West Virginia so obscure that most scouts haven’t heard of it. In eleven starts and six appearances out of the pen, the southpaw recorded a 1.96 ERA with 89 strikeouts and only 15 walks in 78 innings. Moreover, he recorded those stats while also playing outfield for 49 games, batting .345 with 21 walks and 11 hit by pitches against only 22 strikeouts, which helped him post a .446 OBP. His left arm was also an asset in the outfield, where he racked up 17 assists. Conley hit only 3 homers, so his adjusted offense isn’t good enough for him to appear on our batter spreadsheet, but even if he’s only an emergency outfield option, we’re happy to have the flexibility, given our restrictive roster size. Mr. Conley, come on down.

We also recruit Sean Conroy, a 6-foot-1 right-handed pitcher from Rensselaer Polytechnic Institute, an engineering school in upstate New York. In 259 career innings at RPI, Conroy struck out 223, walked 49, and allowed only 4 homers, posting sub-2.00 ERAs in his last three seasons and going the distance in more than a third of his starts. He’s also interested in evaluating clubhouse chemistry, an obsession of ours. A psychology major, he’s working on a thesis entitled “How Perception of Teammates’ Ability Affects Personal Ability,” and we envision him as a like-minded mole on the inside who can be our eyes and ears while advising us on avoiding missteps. We also envision him as an effective arm.

“I’m a sidearm pitcher,” he says. “Just recently I’ve added an over-the-top curveball for an out pitch, which is working pretty well for me.” He tells me how he dropped down in his senior year of high school, and how he’s alternated his arm angle ever since to maximize his deception. “The slider would be the pitch I throw most often, like more than 50 percent of the time,” he adds. No wonder the guy didn’t get drafted: He uses an atypical repertoire from an unorthodox angle, and according to his coach he tops out at about 85. All of the oddities that make him undraftable endear him to us.

And then, of course, there’s Paul Hvozdovic, our on-paper ace and shining spreadsheet star. (The first “v” is silent.) When he signs with us shortly before spring training, I pump my fist, just like Jonah Hill in the “Moneyball” movie when he gets the approval to trade for Ricardo Rincon. I haven’t met Hvozdovic — haven’t watched him, haven’t even talked to him. There’s no way I should be this excited about someone I know so little about. But any hidden doubts that might be buried within me will have to talk to the hand, because the limbic system ain’t listening.

People who write about prospects often speak disparagingly about “Google scouts,” wannabe evaluators who “scout the stat line” instead of seeing players in person or talking to experts who have. Sam and I are guilty of these sins — not because we wouldn’t welcome the input of a seasoned on-site observer, but because we have no time, no travel budget, no scouting staff, and next to no video. Stats are our specialty, but they’re also our only resort.

Gayday, Conley, Conroy, Hvozdovic. At this point, they’re names and numbers, not fully fledged personalities. But they’re our names and numbers. Thanks to us, they’ve got golden tickets to spring training. And thanks to them, we won’t feel ashamed to show up.