The Stat Trap II: Take the Money and Run
In 2003 baseball analytics entered the realm of the populous largely through Michael Lewis seminal work Moneyball: The Art of Winning an Unfair Game.
It told the story of the back to back 100+ win, division champion Oakland A’s and how a low payroll team had achieved such success through a unique approach to team building that focused on particular statistical attention to certain attributes in players-- specifically their ability to get on base. Now, at the same time, OBP was not a new or particularly advanced stat. And while “Moneyball” incorporates making use of objective information, the crux of the premise is actually about identifying undervalued assets in a market. Acquiring players other teams don’t value because of their skills aren’t highly regarded.
And why On Base Percentage (OBP)? Well, there is a historical link that was found to exist between high OBP and run production and, therefore, wins. Now was this some entirely unique discovery by Oakland A’s and to the exclusion of all other franchises? Not at all.
Nearly a decade earlier Dan Duquette was the architect of “The Greatest Team That Never Was”, the 1994 Montreal Expos who were cruising to the best record game in the game at 74-40 until a labor stoppage washed out the final months of the season and the entire post-season. Unbeknownst to many, Duquette at the time had employed the services of a NYC Water Department worker and self-fashioned statistical analyst named Mike Gimbel who went by the moniker “Statman”.
Gimbel was on the payroll of both the Expos and later the Boston Red Sox after Duquette had left the small market, low payroll team to relocate to Boston and shortly before the Expos were relocated themselves to Washington to become the Nationals. Ten years before the publication of Moneyball, Duquette had already figured out ‘The Art of Winning of Unfair Game’ in a little Canadian outpost. He was already playing around with, embracing and tinkering with “sabermetrics” or “SABRmetrics” (coined by Bill James and derived from the Society for American Baseball Research of which he was a member).
MLB front offices were not largely unaware of the principles being espoused. However, many if not most of them were resistant to the notion that years of observational experience and understanding could be augmented or even changed by cold hard numbers. No one wants to believe that what they see is not what it is going on. Especially guys who have been at this 20-30-40 years. Guys who have built or been a part of success. As the sub-title suggests here, it is (or was) considered to be ‘art’ not ‘science’.
However, some in the game were starting to embrace it. It’s mathematical. It’s logical. Most of all it’s historical. What has worked in the past… and any advantage is a big one in any game of incremental measurements and compounding results, especially for the teams operating on small budgets and with limited resources. And the principles in play had already began creeping into both the inward philosophy and the outward justifications of team building throughout the league. Such as when Duquette was famously chided by the rough and tumble, old school Boston media, who were already disenchanted with his businessman demeanor and accountant looks, for claiming that the signing of utility player Jose Offerman would “replace Mo Vaughn’s On Base Percentage” after the Red Sox star first basemen had departed the team in free agency for a lucrative six-year contract with the (then) Anaheim Angels after years of feuding with the Boston General Manager.
Duquette was mocked for the very notion that a player who looked average at best across the “Big 3” stat line (AVG,HR,RBI) could in any way off-set the loss of the mighty slugger Vaughn. In the end, Duquette was right in what he said. Offerman’s .391 OBP in 1998 more than adequately replaced Vaughn and, in fact, both players had nearly identical OBP over the next few years. But it wasn’t Offerman’s ability to get on-base that the media questioned. The question was: Who the hell cares about On Base Percentage? Some dweeb who actually was (and already ridiculed in the press for) employing a guy called “Statman”-- that’s who. The 92 win 1998 Red Sox, after decades of disappointment chasing the vaunted power line-ups of the Bronx Bombers who were in the midst of a dynastic World Series run at the time, had just-- not only lost-- but driven away the team’s best hitter, cheaped out once again, signed some inexpensive nobody and the doom was palpable.
The following season (and the year after that) Vaughn posted impressive Batting Average, Home Run and RBI totals-- but his peripheral numbers and overall production actually dipped dramatically. At the end of of the 2000 campaign he suffered a major injury and was out all of 2001. The following year the Angels pawned him off on the Mets were he was a shell of anything he had been formerly. And, Vaughn’s free agent deal with the Angels is now often cited as the one of the worst in baseball history.
The Red Sox? They won 94 games in 1999, two more than the previous season. Duquette, already having acquired Pedro Martinez from his old club the Expos, after having acquired him originally there from the Dodgers) went on to acquire Manny Ramirez, Johnny Damon, Derek Lowe, Jason Varitek and several other players who were instrumental to the Red Sox finally breaking through in 2004. Some of whom were still on the 2007 team and, in fact, many of the players on that team were either drafted (or were acquired using draft assets-- such as Hanley Ramirez) marshaled during the Duquette era.
The year before the 2004 season where the Red Sox won their first World Series in 86 years, and the same year Moneyball was published, the lower payroll, small market Florida Marlins defeated the vaunted New York Yankees in the World Series. In 2007, when the Red Sox won it for the 2nd time in the decade, they defeated the mid-payroll Colorado Rockies. By 2008 the Tampa Rays were in the World Series… and The Art of Winning an Unfair Game was no longer that unfair and that art was now becoming a science…
Ironically, when the Henry/ Warner group took over the Red Sox, the man they tried to replace Dan Duquette with was none other than Billy Beane. When Duquette used analytics it was geeky and misguided. When Beane did it, backed by a best-seller and later a film where he was portrayed by non-other than Brad Pitt, it was intelligent and cool. Just as it was for his ultimate successor, Theo Epstein. And over the next decade the Epstein’s and his crew, the Andrew Friedman’s and J.P. Ricchardi’s firmly took control of the driver’s seat. General Managers who weren’t on-board with the new wave were being ushered out the door and media and fans alike went from asking “Why do we have a guy on the payroll named Statman?” to asking “Why don’t we have a guy on the payroll named Statman?”
And it all came down to what?
On Base Percentage?
That’s it. No crazy Tom Tango ultra-stat that multiplies by .37 and factors x quotient over r ratio. Sure, the deeper, heavy Tom Tango stats and refinements were in the works and on the way but that’s not how this kick started. It was OBP. That was the simplicity of the beginnings of a revolution. Add the walks to hits and divide by plate appearances... It doesn’t materialize out of the ether. It is simply what happened. Right there every morning in the box score. Team score runs by getting on-base. It seems so simple and obvious it’s hard to believe in hindsight that these things were actually made to be a point of contention.
And to even break the concept down further to its most simplistic level: baseball was and is not about doing something, it is about not doing something. And that is-- making an out. No matter what happens, you succeed simply by not being out. And, really, there was no huge quantum shift here. It was the same thing-- looked at in a different way.
And the idea that a few years earlier had been bristled and scoffed at-- that Jose Offerman was there to replace Mo Vaughn’s On base Percentage (and at a fraction of the cost)-- was now new wave thinking. Wait! How can that be? Mo Vaughn hits for a good average, hits a lot of home runs, gets a ton of RBI. Those are the Triple Crown categories. The ultimate achievement a hitter can attain. How can two of the three not really be that important?
We had simply gotten too glued into the top line. The hits and runs and RBI. Too invested in action stats and clutch myths. Too enamored with the legends of Ruth and Musial, Mays and Mantle. And when the legend became fact, we just printed the legend.
Ricky Henderson was perhaps the greatest lead-off hitter in the game. I know this because Ricky told me. After breaking (basically shattering) his all-time stolen base record Ricky famously said, “Lou Brock was the greatest of all-time, today I am the greatest of all-time”. And he may be right. But not because he stole a ton of bases, because he got on-base enough to steal that many bases. Henderson, less ceremoniously, also retired as the all-time leader in walks (passing Babe Ruth) and with a career .401 OBP. We all knew the 10X All-Star and former MVP was a great player-- at the time we just didn’t have the right (or better yet, full) reason as to why.
There was a simpler reality that was sitting in plain sight for decades and being over looked by a change in thinking. Lower level lines that seemed more clean-up accounting than really pertinent. It wasn’t sexy, it wasn’t exciting-- so it was perceived as unimportant. Why would anybody want to pay attention to it? As the adage goes, “No one bought a ticket to see Ruth walk”. Of course, he did. A lot. He’s, as obviously stated before, third all-time in the category (as both he and Henderson were subsequently passed by Barry Bonds.)
Ruth is, also, 2nd All time in OBP with a .473 mark trailing only Ted Williams. Henderson, meanwhile, is a very respectable 58th. However, interestingly, the vast majority of players on the list in between played before 1970. And this suggests to me that there was much more of innate value in the origins of the game, something that far predated this Moneyball ideology, that understood the value of getting on-base. Your legendary Mantles (19th) and Musials (25th) were, in fact, all-time greats in getting on-base.
Somewhere in this process, the game got changed. Not enormously. Just enough of a shift, slowly and with little notice, and the prevailing view slightly askew from it’s origins. It was minor deviation that became a significant variation and, in turn, caused this imbalance. Billy Beane and the Oakland A’s and OBP and Moneyball wasn’t a discovery. It was a recognition. It was finding and exploiting a systematic, widespread breakdown. And simply getting back to the basics of what was already known and proven but somehow sort of got forgotten.
Now, if this true, what could have happened in the 70’s that had perhaps changed the game in this way?
Well, first there was the official formation of the MLBPA, the player’s union, in 1966 which became an increasing factor into the game as the years went by. By 1976 they had negotiated a collective bargaining agreement that for the first time included free agency on the heels of the Curt Flood case. Suddenly there is value and money consideration in sports. It is not just a directive to stockpile talent. The talent must balance to the money that can be spent on it by the organization-- without which we don’t have any need for “Moneyball”.
Also, in that same year, ABC launched Monday Night Baseball on the heels of their wildly popular Monday Night Football programming. The explosion of sports based media was on the way and the emphasis must as well be on entertainment value.
Two seismic shifts begin occurring in the landscape. A hindrance within the game on the ability to collect and retain talent and the marketability of certain skills in a forwardly progressing media and entertainment environment. Not long after this, you have players arriving on the scene like… Ricky Henderson in 1979. A flashy athlete electrifying stadiums and TV screens with 100+ stolen base seasons, while his larger contribution (OBP) to the team is overlooked. Nobody paid to see Henderson walk, either. It was what he did after that people cared about. One skill was exciting. The other was not. Value be damned. And just as fittingly, by the mid 80’s, we saw the small market Oakland A’s trade an increasingly expensive Henderson to the New York Yankees just as he was entering prime years. The changes to the game were causing the game to change. It’s exactly where Moneyball has now put us back to nearly 20 years later-- with the discussions and dissensions from the children of the 70’s and 80’s who pine to get back to “small ball”. They don’t want to see Ricky walk, they want to see Ricky run.
But the reality here wasn’t about Henderson’s ability to do any one particular thing. Henderson’s true greatness was his ability to do everything. His approach that made him one of the all-time leaders in walks; his speed and acumen to steal bases that also allowed him to cover ground in centerfield; and his ability to hit to hit for power (.419 SLG%, most career lead-off Home Runs).
Babe Ruth and Mickey Mantle and Stan Musial and Ted Williams smacked the hell out of the ball. The hits and Home Runs and everything else was important. It all matters. It all factors. And if any area suffers a breakdown in the perception of value, it can and should be exploited.
So, lest we not get this confused. On Base Percentage and Moneyball are not synonymous. OBP was simply the tool that was available to the A’s based on the market factors involved and they objective they set out to achieve and thus the tool that was used. The premise is not that there is only one way to play and win and the Oakland A’s or Dan Duquette or anyone else just discovered something the world was entirely oblivious to. The 2003 World Champion, $44M payroll Florida (now Miami) Marlins were exactly league average in OBP. It is not-- as most are not-- an end all, be all stat.
The premise here was to identify tools and production from players who were being undervalued in the market for other reasons-- deficiencies in their play that perhaps had more perception than value. Or where the value negated the deficiencies. It was not about one contribution above all others. It was about finding easily obtainable players that possess skills that contribute to a higher win total and are available at modest prices because the perception is skewed from the reality. And that can only be accomplished objectively, which is the methodology the A’s were bringing to the table and hence the demonstration of the value of metrics in Moneyball. And that was the true lesson here-- this is what the numbers, derived from sound methodology, said to do and, despite any misgivings, we did it and… it worked!
By employing an ‘against the grain’ ideology, they sought to discover a crack in the system and they found one by observing the functionality of all of the information involved: production, money, value. And to ask the question: what here is of the most importance? What is the interaction between input and output? How can we use that to best utilize our resources to better the team? What is being undervalued by everyone else? What is perception? What is reality?
Cut forward some 20 years later and where are we now with perception and reality in modern football analytics? The perception is quite obvious: Passing. It’s a passing league. You hear it everywhere. The emphasis is on the QB. The focus of most of the work being done-- the work being put out into the ‘public sphere’ is all about pass, pass, pass. That’s the directive coming out, the course the league is on and the focus of most analytics that are available. You pass to score, pass to win. Get the QB, evaluate the QB, trade for the QB... Which brings us to the next question of relevance from our previously referenced ESPN survey:
Which teams are among the top five most analytically inclined?
Cleveland Browns (22), Baltimore Ravens (22), Philadelphia Eagles (14), Buffalo Bills (12), Indianapolis Colts (8), Los Angeles Rams (6), Minnesota Vikings (4), San Francisco 49ers (4), Jacksonville Jaguars (3), Atlanta Falcons (2), Green Bay Packers (2), New England Patriots (2), Dallas Cowboys (1), Denver Broncos (1), Detroit Lions (1) and New York Jets (1)
There are two teams unanimously named by all (22) respondents on this list. Named as the most analytically inclined front-offices and organizations who are doing the most analytical work in the NFL and they are, of course…. run heavy football teams.
(The Ravens led the league in rush attempts from 2018-2020. The Browns were 4th last year and are so far 2nd this year).
For all the ‘pass is King slant’ we’re getting in virtually every other analytical format, the two teams doing the most work in analytics (and it’s obviously not even close according to the insiders) are designed and built to heavily run the football? Two teams near the top of the rush attempts list and near the bottom of the pass attempts list the last few years? Wait a second…. (insert Vince Lombardi-- What the hell is going on here?)
What in the numbers are they seeing that the others aren’t that are pushing them in that direction?
What is... their OBP? What are the lines that are tucked away in the box score? How (if in any way) does Moneyball factor in this? And perhaps most importantly: in football, do we even have the right box score to begin with?
We’ll discuss that further in Part III….