Happy birthday, rotisserie baseball!
It's been 40 baseball seasons since Dan Okrent recruited 11 friends and colleagues to play in the very first rotisserie baseball league, as detailed in 1981 in Inside Sport magazine. There are several claims as to who invented "fantasy" sports, but there is no doubt who invented rotisserie baseball, the scoring system named for the New York eatery frequented by Okrent and his friends, La Rotisserie Française.
To mark the occasion, I invited some industry friends and colleagues to play a little game of my own. I called it Project G.O.A.T., and the goal was simple: to assemble the greatest rotisserie baseball team ever, using seasons from 1980-2019. Forty years of baseball greatness on one roster.
But that seemed way too simple. Drafting any fantasy team is an exercise in compromise. If you take Ronald Acuna Jr. with the first pick in your draft, it's not reasonable for you to assume that Mike Trout, Christian Yelich or Gerrit Cole will still be waiting for you next turn. Therefore, parameters had to be set to ensure that every player had a cost associated.
First, a player could be used only once, regardless of position. So if one were to use Alex Rodriguez as a Mariners shortstop, one could not also have him as a Yankees third baseman.
Second, only one season was allowed per calendar year. In order to use Pedro Martinez's outstanding 1999 or 2000 seasons with the Red Sox, one has to forego Ivan Rodriguez's MVP 1999 season (the only 35/25 100/100 season by a catcher in the history of baseball) and Todd Helton's 2000 campaign (the year he hit .372 with 42 homers and 147 RBIs).
Third, each franchise can be represented only once. Pick a Braves pitcher or a Rockies hitter, but choose wisely. Nobody gets to rebuild the Rockies' lineups that terrorized pitchers in the pre-humidor years or the Big Three that won seven Cy Young awards for the Braves in the 1990s.
And finally, to represent all four decades, a max of six players per decade was allowed, which spread them out in some combination of 6/6/6/5 across the 1980s, 1990s, 2000s and 2010s.
Simple enough, right? Rosters must include two catchers, one of each infield position plus a middle infielder and a corner infielder, five outfielders, a "UT" (which can be any position) and any nine pitchers for a total of 23 players, just like the original game.
While the original league was 4x4, we played this one 5x5; batting average (spare me the sermon, OBP truthers), runs, home runs, RBIs and stolen bases for hitters, and wins (you too, "never-winners"), strikeouts, saves, ERA and WHIP for pitchers.
Position eligibility was determined by either 20 games played the previous season or at least 10 games played there during the season in question. For players who were traded during the season, the franchise with which they ended the campaign is recognized, so Mike Piazza 1998 was a Met, not a Dodger nor a Marlin.
Got it? Think you could handle it? Well, download the worksheet and the complete set of rules right here, and give it a try before you read on.
I dare you.
The contestants
Once the rules were set, 12 participants were needed to play it out. I reached out to some of the best and brightest fantasy baseball players in the business. Of course, I started at home with ESPN's own Eric Karabell, Tristan H. Cockcroft, AJ Mass and senior baseball writer Dave Schoenfield, whom I've lost to more than once over the years.
Mastersball's Todd Zola and FanGraph's Paul Sporer, both frequent contributors to ESPN Fantasy, were next. Then, I could think of no better fantasy baseball mind than Ron Shandler's to tackle the project. Add to those Rotowire co-founder Jeff Erickson, The Athletic's Eno Sarris, Yahoo's Andy Behrens, USA Today's Steve Gardner and former baseball exec Brad Kullman, currently with BBHQ, and you have a formidable ensemble of fantasy thinkers with multiple expert-league titles and easily two centuries of playing experience between them.
Each of them graciously said "yes" right away upon being issued the challenge; I'm not sure they fully grasped exactly what they were about to put themselves through, but even if they had, I'm sure the answer would still have been "yes." It's just the kind of selfless, dedicated, fantasy baseball masochists they are.
Gaming the system
I'll admit now that I had an ulterior motive when formulating this contest. No, it wasn't to torment my unwitting victims, but rather to prove a point. Our game is terribly dependent on rankings and projections. We're trying to predict the future -- an impossible task -- and we assign a huge amount of value to what we think will happen. Best guesser wins, right?
Project G.O.A.T. completely eliminates the projections aspect, as well as the chaos of the draft or auction. Every one of the 12 participants had the same opportunity to acquire any player at the price of forfeiting every other player from that season or wearing that team's uniform. Every batting average point or batter struck out is a known quantity. And yet, with nearly endless permutations available, I expected I'd get 12 very different rosters, and per rotisserie scoring rules, someone would finish first, and someone would finish last. Could there be a grand revelation to be had from how that came to be?
Among the other rules were that they were not to discuss or compare their rosters with each other and no algorithms were allowed. Spreadsheets, sure. Dollar values? Knock yourselves out. Anything that emulates draft preparation was fair game. Ron Shandler has a project of his own on his site, a Roto Hall of Fame.
"I already had what I thought was a head start because I've been keeping track of these peak seasons," he said. "I sorted all the players and immediately slotted the $50 seasons into my roster, then scaled down from there. Then the 80/20 rule took over. The last 20% of the players took 80% of my time."
Each one of our participants had a variation of that method; start with great seasons, picked by system or by instinct, and then try to solve the puzzle by moving the pieces around until you had a 23-man roster that fit the parameters, fitting as many stat-packed seasons as possible into every slot and researching players from the past 40 years who may have had one of those magical seasons.
"I generally just tried to get a bunch of 30/30 and 40/40 seasons," said Eno Sarris, "but near the end, it turned into a puzzle, where I just had to find a good second baseman from the 80s, and Juan Samuel fell into my lap."
In 1984, Samuel stole a career-best 72 bases and scored 105 runs. Nice find.
But how did any of the contestants know they were finished?
"It was a bit like the SATs for me," said Eric Karabell. "I trusted my original instincts, then whenever I wanted to fiddle, I decided it was a bad idea."
Some took a much longer route to that conclusion. Brad Kullman had methodically built variations of rosters prioritizing either batting average, power or pitching.
"By the early afternoon of the Sunday deadline day, I had built close to 11 teams and planned to take all I had learned to combine my vast knowledge into one contest-winning super-team! Then, I had a great idea. Since this contest involves 12 teams, why not plug the totals from my 11 teams into a 'standings' spreadsheet, approximating various strategies that might possibly be employed by my competitors? Then I will build a super-team that ranks toward the top of each category and win the championship! I was so proud of myself.
"As day turned to night, however, I came to a rather startling -- and disheartening -- conclusion. No matter how hard I tried, my new super-team was consistently finishing around third or fourth in my mythical standings, no matter how much I tried to adjust it. And what team was sitting on top of these mythical standings? The very first team I had thrown together, of course. As unthinkable as that conclusion may be to accept, my eyes began to glaze over as I tinkered late into the night before I finally cried "uncle!"
Of course, sometimes, a little outside perspective helps.
"After three iterations over four days," admits Shandler, when asked how he knew he was finished, "my wife asked, 'Are you finally going to have dinner with me tonight?'"
As much as Kullman's experiment proved to not bear fruit, he was on to something. By running his own little derby and trying to anticipate what his competitors might do, he gave himself context by which to assess his own efforts. And if he'd been able to mirror the strategies of the 11 other teams, he may well have had the one team to rule them all.
But as it turns out, the biggest variable in this contest was how each individual chose to address the most volatile of all fantasy stats: the save.
The closer quandary
With nine pitching spots available to chase five categories, this is where strategy really came into play. Sarris earned 12 points by efficiently gathering 147 saves, six more than Kullman and Shandler, who tied to earn 10.5 points. They were among the five who allotted three pitching slots to closers.
"It's a classic game of chicken. Some will punt the category entirely," explains AJ Mass. "Some will throw in a token closer to avoid a zero and attempt to steal a few standings points. My feeling was that if I could identify a trio of solid 40-save seasons with solid ratios, close to 100 K's and a handful of wins, the risk/reward would be worth it."
Five others went with a two-closer approach.
"I knew I wanted to go with an innings-based approach and that I didn't want to flat-out punt saves. I was going to go light on that category in the hopes of scraping by on 3-4 points, loading up on wins and strikeouts, and going with the early-era closers who would push 100 innings with great ratios," explained Tristan Cockcroft. "My calculations simply said that guys like Willie Hernandez provided a premium that'd help my strategy, while at least giving me a fighting chance for a couple roto points."
(Hernandez notched 32 saves, striking out 112 while posting an ERA of 1.89 and WHIP of 0.94 in 140 1/3 innings for the '84 Tigers).
Cockcroft's podcast partner, Eric Karabell, earned the most saves points with his two closers, 1992 Dennis Eckersley (51 saves) and 2008 Francisco "K-Rod" Rodriguez (62 saves).
"I dedicated two slots to saves, opting for the one fellow with the most saves but not necessarily other numbers warranting his selection," he said, figuring correctly that this approach would allow him to compete well in the category with others who employed two closers.
As for the final two squads, Andy Behrens and Dave Schoenfield both opted to punt saves completely.
"I'm not someone who typically punts categories, but it was a pretty liberating choice here," said Behrens. "At some point early in the process, I decided nearly everyone would be focused on a very small number of individual closer seasons [and that] everyone who went with two relievers would total either 114 or 119 saves. (Maybe that's wrong, but it's what I concluded.) So how many roto points would that really get me? Choosing three closers would almost certainly mean I couldn't compete in wins or K's, and I didn't want to sacrifice two categories in order to be great in one.
"Another benefit of punting saves is that I didn't feel compelled to choose every starting pitcher's top strikeout season. That is to say, I felt I could still finish at or near the top of the standings in W's and K's without rostering Randy Johnson's 372-strikeout year or Clayton Kershaw's 301. Those were obviously both absurd seasons, but not either pitcher's best in terms of ERA and WHIP.
"I should also mention my concern for the fact that another manager would also dump saves, so I was this close to rostering Orel Hershiser's 23-win 1988 season," Behrens admits. "He earned one save that year (and the prior season), which would have given me an edge over any other punter. But the cost of using Hershiser was just too great; it would have meant no Kershaw and no '88 Jose Canseco."
Using this no-closer-all-starters philosophy, Behrens and Schoenfield tied for wins at 187, finished 1-2 in strikeouts as the only two teams to average at least 250 Ks per pitching slot, and gathered 18 and 19 points, respectively, in combined ERA and WHIP (out of a possible 24) to tie with 43 pitching points each, one ahead of Kullman's three-closer squad, which posted a league-best 1.84 ERA and 0.90 WHIP to add 24 points to the 10.5 earned in saves, but he lagged behind in wins (4.5 points) and strikeouts (3.0).
Offense would now determine the winner.
1981 Donruss Tim Raines #538 baseball card
Unlike with pitching, where chasing wins and K's is a compromise to chasing saves, and sometimes at the expense of your ratios, all hitters have the opportunity to contribute to all five categories -- some more than others. What can really make a difference, however, is the position your player occupies on the field and, therefore, on your roster. Ivan "Pudge" Rodriguez is the only catcher to ever produce a 35-home run, 25-steal season while also driving in and scoring more than 100 runs, at least as far as Baseball Reference's Play Index can tell us (oh, and he hit .332 that year as well). That's why Pudge '99 showed up on 9 of our 12 rosters.
Outfielders have put up 21 such seasons just in the scope of our project, 1980 to today, and only 10 of them hit .300 or better. So finding the right combination of hitters isn't just about balancing franchises, seasons and decades, it's about whether less is more at certain positions; sacrificing bigger numbers elsewhere to maximize your output in any given roster slot.
"I tried to go for balance across all five stats on offense, although I guess with a slight emphasis on batting average and stolen bases," said Schoenfield of his approach. "It also became clear that the 1980s offered fewer great seasons to choose from, so I made sure I wasn't stuck with some relatively mediocre seasons from that decade.
"Which is why, for example, (pitcher) Mike Scott was my Astros player over many other strong possibilities (Jeff Bagwell, Craig Biggio, Gerrit Cole, Justin Verlander). Without one of those monster Rickey Henderson or Tim Raines stolen base seasons, I made sure to get enough steals, adding the likes of Hanley Ramirez, Roberto Alomar and Eric Davis (all with 50-plus steals) with my final position player slots."
That balanced approach led to an offense that finished each category with no worse than seven points (in runs and steals) and no more than 10 points (average, home runs and RBIs) for a total of 44 batting points, which, added to his 43 pitching points, set a high bar at 87 total points.
"For a little bit, I contemplated punting batting average to see if that might open up a path to winning more hitting categories," says Steve Gardner, who actually considered 2013 Chris Davis at one point. "But as I got further into the numbers, it was readily apparent that the great seasons were great across the board, and you were generally going to get a good batting average along with all the other counting numbers. Deep down though, something tells me there's still some game theory that can be applied to this to come up with the best team."
However, it wasn't game theory that ultimately decided this contest; it was good-old fashioned positional flexibility and a hobby much older than fantasy baseball: card collecting.
If you absorbed the rules of the game, you'll remember that position eligibility was determined the same way we do it today: 20 games at the position the season prior or 10 games that season. You'll also remember Schoenfield referencing Tim Raines, owner of a 90-steal season in 1983, when he also scored 133 runs and drove in a respectable 71. Finally, you'll remember what I stated about outfielders being a treasure trove of five-category mashers.
Behrens, however, remembered one more thing: Tim Raines came up as a second baseman.
"Card collecting taught me that one," Berhens said. "I can still see the '81 Donruss Raines, second base."
Raines played only seven games at second base for the 1983 Expos, but in 1982, he played 36 at the position, carrying that eligibility into '83.
Kullman also used Raines '83 on his roster but as an outfielder, opting for Jose Altuve 2016 and Ryne Sandberg 1984 as his second baseman and middle infielder, choices which contributed to his falling behind on offense and managing no more than eight points in any one category.
Behrens, however, tallied 12 points each in the runs and RBIs categories, 11 in home runs and 9 in stolen bases, and a perfectly respectable .3219 average earned him 7 points for a grand total of 94 in the overall standings.
And that's how Andy Behrens put together the greatest fantasy baseball team ever ... at the time.
Redemption
Don't feel bad for AJ Mass. While his team fell short, and someone had to finish last, I had one more trick up my sleeve. You see, while this project's shiny premise is to get to play with all these great seasons and reminisce of past rotisserie glory, it was always about the oft-overlooked and glossed-over truth about rotisserie baseball: At its core, it is a simple game of supply and demand.
Even with the restrictions imposed on our competitors, the supply was nearly limitless. The blind spot was demand. Without knowing how many saves anyone was getting, how could you know how many closers to use? Without knowledge of who was getting what from which slot, how was someone to balance power and speed, strikeouts and ratios?
We give such importance to projecting the stats we will get from our players, but it's as important to project the stats everyone else in your league is getting and whether you have the right makeup to find a path to victory. This is where trades and pickups can be honed to suit your endgame. It's not about which player is better than another one, it's about which player will provide the stats you need the most, hopefully at the expense of those you need the least.
That doesn't necessarily mean you swap the stats that you have the least of for those that you have the most of -- and that is the revelation to be found in this project.
In order to drive this point home, I proposed to Mass that he give it one more try. But this time, I gave him the answer key; the standings. No rosters, no other clues, no subtle "say, did you happen to collect baseball cards back in 1981?" But armed with his exact targets and the ability to assess each move, not in a vacuum but within the context of what his 11 opponents had already done, it didn't take too long for AJ to solve the puzzle.
"I was exploring two different paths on the hitting side, Canseco's 40/40 season vs. Rickey's 130 SB season, to see whether or not steals were worth pursuing," he explained." Because those are both A's seasons in the 80s that provided a clean pivot point for juggling. For pitching, I opted to either tank saves altogether or see if I couldn't steal a few points using Mark Eichhorn as a 14-win swing with super ratios. (In 1986, Eichorn pitched 157 innings for the Blue Jays, all in relief, striking out 166 with a WHIP of 0.955 and an ERA of 1.72). But, given Andy's totals, it made it apparent that 'all or nothing' was the more prudent route.
"After that, it was just creating SP/hitter pairs from similar teams (Verlander/Trammell, Sosa/Arrieta, Kershaw/Piazza, etc.) and trying to constantly get one half-point closer to the top at a time. Once I finally got to 80, it was then much easier to target specific goals (e.g., I can afford to lose only 2 HRs on the next swap, but I need to increase BA by 2 points, etc.) Finally, taking out Barry Bonds' 73-home run season and replacing him with Sosa so that I could slot in Jeff Kent at second base brought me to a tie, and then it was just a tinker or two -- including replacing Kent -- before I found George Bell's 1987 MVP season and slotted it into my UT spot for the win. Poor Frank Thomas never seemed to get me where I needed to go."
Let's play two
While Mass did what he was asked, there are two facts that remain. The first is that Andy Behrens will forever be known as the first winner of Project G.O.A.T. The second is that the game is never really over.
Now that Mass has changed the standings and climbed to the top, he, too, can be beat, and with every iteration, a new "meta" can emerge, as the kids say.
Right now, the meta says to blow off saves. Behrens won the first game by finding a way to fit 159 steals into his middle infield. Mass beat him with only 29 steals out of those slots. And then Mass stopped. Could there be more points to be gained by further eschewing steals? At what point does a closer, or a Mark Eichhorn or an Orel Hershiser, become a viable option to snag easy points ahead of the no-saves teams? What future seasons will be included in some future version? Let's find out.
If you'd like to take a shot at the top of the Project G.O.A.T. standings, here is your chance. It will remain a living challenge, at least for the duration of the 2020 season.
Rosters, facts and figures
No project claiming to be about the greatest fantasy team of all time could leave you hanging without revealing the rosters. You can download and peruse them all in PDF format here.
And finally, here are 10 fun facts about the teams of Project G.O.A.T. 1.0:
Three players were tied for having a single season used on the most rosters: 1985 Doc Gooden (Mets), 1999 Pudge Rodriguez (Rangers) and 2012 Mike Trout (Angels).
Alex Rodriguez was the only universally used player, appearing on all 12 rosters: Nine times as a Mariners shortstop (1996 and 1998) and three times as the Yankees' 2007 third baseman.
The next most-used player was Barry Bonds, appearing on all but one roster in Phase 1. His 2001 season was the most-used (5), followed by 1993 (4) and then 1996 as well as 1992, his final year with the Pirates. That was the Pirates' only representative in the entire project. That season was then dropped by Mass in Phase 2, shutting the Bucs out of the project and reducing Bonds' footprint to 10.
Mike Trout did appear on one more roster: Kullman opted for his 2013 season in favor of using 2012 on Aroldis Chapman's 38-save, 122K season (Reds).
The most-used pitcher was Randy Johnson, all with the Diamondbacks in 1999 (1), 2000 (1), 2001 (2) and 2002 (7).
Rickey Henderson could be the poster boy for this project, with multiple triple-digit steals seasons and excellent seasons in two different jerseys at the height of his career. However, the five squads featuring Henderson currently reside at the bottom of the standings after Phase 2. Beware chasing steals at the expense of the other categories.
There were 14 different individual closers used in the 25 spots that were dedicated to them. Only three were used more than once in a single season: Bobby Thigpen in 1990 (3), Eric Gagne in 2003 (2) and Zack Britton in 2016 (3). The most recent was Josh Hader in 2019.
Five players from last season were selected: Christian Yelich (3), Cody Bellinger, Justin Verlander, Gerrit Cole and Hader (once each). No Ronald Acuna Jr., whose .286 average and 86 RBIs may have felt a bit light on a list of all-time great seasons. Food for thought?
Six teams were universally represented in Phase 1: The Braves (Kimbrel, Maddux, Javy Lopez), Cardinals (Albert Pujols, Mark McGwire), Dodgers (Kershaw, Bellinger, Gagne, Mike Piazza, Matt Kemp), Giants (Bonds, Kent, Brian Wilson), Mets (Gooden, Piazza, Howard Johnson) and Rockies (Ellis Burks, Larry Walker).
In Phase 2, the Giants lost their representation on Mass' squad, bringing the universal franchise number down to five.
No baseball cards were harmed in the making of this project. That we know of.