OK, here is the post a lot of people have been waiting for: What Geoff Thinks of The Browns Hiring Paul DePodesta.
To jump to the conclusion. “Not much.” I’m not sure this is a good
idea, because I don’t know:
- What type of analytics the Browns plan to use.
- How capably DePodesta will develop them.
- How well the Browns will use them.
- Whether they can get people to buy into them.
People seem thrilled because (a) they know various baseball teams (Boston, Oakland, the Cubs, the Mets) have been successful using the ideas Michael Lewis explained in Moneyball and (b) they remember that DePodesta was a pretty prominent character in that book. They assume I must be thrilled, because I write posts that use a lot of numbers and DePodesta works with numbers.
That’s inaccurate, for many reasons. Let’s start with the basics.
Using analytics does not guarantee success
The simplest way to begin is (fittingly enough) with a chart. One of the other prominent characters in Moneyball is J.P. Ricciardi. He was hied by Toronto and spent eight seasons as GM. Here are the Blue Jays’ records between 1998 and 2015. Identify the seasons, Ricciardi worked for the team, if you please (the yellow bar means nothing; it was just there when I did the capture):
Ricciardi was GM between 2002 and 2009. You can’t see a great deal of difference between his results and his predecessor (Gord Ash) or his successor (Alex Anthopoulos), because there aren’t any. Ricciardi had four winning seasons (86, 87, 83 and 86 wins) and four losing seasons (78, 67, 80 and 75). He never made the playoffs (admittedly difficult, with the Yankees and Red Sox around), and his career record with the Jays was slightly below .500.
DePodesta was GM of the Dodgers in 2004 and 2005. His he won 93 games and then 71, making his lifetime record 164-160. Los Angeles had won 92 and 85 games
the seasons before he arrived (177); they won 88 and 82 the season after he was fired (160).
DePodesta did win a pennant in 2004, but where you place depends on the competition. Dan Evans, his predecessor, won 92 games in 2002 (one game less than DePodesta) but his team finished third because Arizona won 98 and San Francisco won 95.
Even Le Gran Mamou of Moneyball– Billy Beane– has struggled. In his 18 seasons, they have had 11 winning seasons, six losing ones and one .500 season.
There’s a very good reason why, and I’ll get to that later.
Theo Epstein has been the most successful of the analytics guys– two championships and two ALCS losses in Boston between 2003 and 2011, before going off to Chicago. But Epstein had both an extremely high payroll (never lower than sixth– usually second), the Bill James (who understands what numbers can and cannot do better than anyone) and a manager (Terry Francona) who would go along with his idea.
Building a winning organization requires more than crunching numbers.
Analytics are not equally useful
The impact that analytics have depends on what numbers you crunch. For example, the Browns could significantly improve their chances of winning, simply by using the following analytical finding:
- Do not sign free agents at or past 30 and expect them to play well. They’re only going to get worse.
This “What the Browns are thinking” piece by Terry Pluto reads like a humor column at this point. Virtually every word is not merely wrong, but laughably so. A few one-liners:
- “They believe they know Dwayne Bowe better than most teams… The Browns believe Bowe is determined to show he is still a very good receiver.”
- “While they’ll never say it, they believe [Tramon] Williams is superior in single press coverage to Skrine. Williams is 32 and the Packers believe he is in a bit of a decline. They offered him a two-year deal in the $10 million range [The Browns offered three years and $21 million, $10 of which is guaranteed]”
As the 2015 season showed, Green Bay made a correct offer. There are kind words for Randy Starks and Brian Hartline (who was close to 30), as well.
Had the Browns had DePodesta– and. more importantly, listened to him– he presumably would have told them not to make those moves.
But you would not need DePodesta– or a huge data warehouse– to estimate the chances that signing of a 30-year-old free agent would work out. You could do it in Visicalc on an IBM PC. If you simply wrote down every multi-year free agent signing on a legal pad, looked at how well they did a year later and made checkmarks beside the good ones and “X”ed the busts, you could figure it out. If you simply have a good memory and read the papers– like probably hundreds of thousands of fans do– you could grasp it.
A competent general manager would have known this. Here’s another advanced not-so-advanced analytical concept:
- Players who have alcohol or drug incidents in college– or arrests– tend to have even more incidents after you give them millions of dollars.
Warren Sapp (failed drug test) or Jabaal Sheard (an arrest) are the exception. People like Armonty Bryant, Puff Gordon and Billy Relapse are the rule.
The problem holding the Browns back has not been the absence of an advanced analytics department. It is that the people running the team have difficulty answering questions on the order of “Should we go swimming in the processing tanks at the Garrett A. Morgan Sewage Treatment Plant?”
Could DePodesta add value? Of course. Here’s a short list of questions that could be answered by basic research, which could have a huge impact on a team:
- Is the draft value chart developed by Jimmy Johnson still accurate?
Johnson based it on performance by picks in the last 20 years (the
70’s and 80’s). Has time and improvements in scouting changed the
- What are the rates of return on draft picks by position, by conference and by college coach (head coach, coordinator and position coach)?
- Can the aging curve developed by Bill James apply directly to the NFL? Assuming it needs changes (I certainly think so), what are they– and how do they vary by position?
- What is the impact of having a lefthanded quarterback? What tendencies do they have?
- What about height? Very tall quarterbacks (like tall pitchers) seem to have control problems. Is that so?
- Can you estimate growth for a football player– using only their statistics– as you can in baseball?
- How many snaps/carries/catches/tackles does it take before the stress of playing begins to wear a player down?
- Does a drop in yards per carry/catch by a back/receiver signify decline due to age? What about a drop in touchdowns?
None of those questions are terribly difficult to answer– if you have access to the data and the staff (or technology) to collect and analyze it. I assume, for example, that a center (who plays against the nose tackle in a 3-4) is more valuable if you play a lot of 3-4 teams, but that guards are more valuable if your opponents line up in a 4-3 (where they play on the tackles). Hence a center has far more value in the AFC North than the NFC North.
I have never had the data (or the time) it would take to determine that. If DePodesta can verify it, it would help the Browns immensely. They work on the “Let’s get two tackles and a center and then not worry too much about the guards” scheme.
But there are many issues where he won’t be able to add value.
It is, I think, well-known that Joe Banner spent $100,000 to commission a study of the quarterbacks available in the 2014 draft that ranked them in the following order:
- Teddy Bridgewater
- Derek Carr
- Blake Bortles
Perhaps we will look back someday and see how shrewd that was. Right now, it looks seriously tweezed. Bridgewater seems to have plateaued at the level of “Pretty Good.” As many old-fashioned scouts suggested, he seems to lack the arm strength required to succeed at the elite level. Carr has come along fastest, but Bortles showed remarkable growth in 2015, and might blow by Carr in 2016.
Lord only knows what the people who took Banner’s money did, but it appears that they simply compared their college statistics, without doing any weighting for strength of schedule or supporting players.
Which brings us to Mr. DePodesta. While he was with the A’s, DePodesta tried to develop a system to assess draft picks from their statistics in high school and college. If you merely look at the A’s draft record (here’s the data for round one; use the data entry box to see other rounds), you see what a disaster that was. Mark Teahen and Jeremy Brown– two of the players Paul DePodesta recommends on pages 29-30 of
the edition of Moneyball I have– were colossal busts.
Perhaps DePodesta refined the system and produced better results with other teams. And, yes, he did prod the A’s into taking Barry Zito (who turned out to be a lot better than Ben Sheets). But unless someone connected to a team has made a major discovery that no one has found out about, statistical evaluation of amateur football players has always been something of a rathole.
What he works on will determine the amount of value he produces. There would be much greater value in having DePodesta collect the low-hanging fruit that has gone unharvested.
Baseball does not have a clock
One of the reasons baseball has been analyzed much more than football is that baseball lends itself spectacularly well to analysis. There is a lineup, so everyone bats; they do so in a proscribed order. There are only three spots on the field where a batter can move to. Innings, outs, and ball/strike counts let you map everything.
Baseball doesn’t have unlimited substitutions. Defensive positioning can be used, but not to the degree it can be employed in football.
Most importantly, there is no clock. Unless it rains, every team gets 27 outs. Which means, as the Indians demonstrated on August 5, 2001, a team can come back from an unimaginable deficit. In that game, the score was 14-2 Seattle, going into the bottom of the seventh inning. It was 14-9 in the bottom of the ninth. With two outs in the ninth and a runner on first, the score was 14-9.
Because that sort of comeback is possible– because there is no clock– every at-bat in a game has approximately the same importance and every offensive event has roughly the same value. That makes analysis simpler.
Other than an intentional walk (which analysts violently oppose), there is never a point in a baseball game where a team would choose to let a hitter reach base. They might “pitch around” a good hitter– letting him walk if he wants to be patient, or get himself out if he doesn’t– but that would be the limit. With a big lead, teams tell
pitchers “just throw strikes”– don’t walk people and bear down if you put men on, but otherwise, a single here, or a solo homer there is OK. But no team ever willingly permits an opponent to bunch singles and walks into a big inning.
In football, because it has a clock, a team can and will let opponents score. If the lead is large enough, and enough time has elapsed, a defense will permit an opposing offense to pile up first downs– even score– while it runs the time off the clock.
There are no turnovers in baseball. The team at bat can make a baserunning error to let the defense record an out without retiring a batter. It can hit into a douleplay to consume two outs on an at-bat. But there is no play in baseball where the defensive team can simply stop pitching until after the the third out had been recorded.
Baseball’s penalties have much smaller impact than football’s– and they are much less common, Weather in football is a much more significant force.
Injuries are far more significant. To name a very small component, if Michael Brantley gets injured and the Indians send out a defensive stumblebum (like Rajai Davis) the next few hitters will not try to “pick on” Davis by deliberately hitting fly balls in his direction. If a cornerback goes down, his replacement will almost certainly be targeted on the next play.
To name a more significant impact, the existence of baseball’s minor leagues (and the ability to obtain a replacement player of some quality within 24 hours) changes the process teams use to select a roster. The Indians managed to win with Sandy Alomar Jr, who had serious injuries in 7 of his 10 seasons in Cleveland. An NFL team would never tolerate a player who missed that many games that often.
In baseball, players from other teams are virtually interchangeable. Trading, especially in-season, is a much greater factor in baseball. In the NFL, it takes weeks– often a season– to fit a new player in. It is possible for an NFL player who succeeds in one scheme to fail miserably in another.
Because there are 162 games in a baseball season– ten times as many as an NFL season– the outcome of a baseball season is rarely due to luck. In football, it often is. And we haven’t even gotten to the most important issue
Moneyball isn’t necessary in the NFL
I’m not saying that analytics can’t work in football. I’m saying that Moneyball, as the Oakland A’s practiced it, has no value in the NFl. None.
Do you remember what the subtitle of Michael Lewis’s book was? Most people don’t. It was The Art of Winning an Unfair Game.” As Lewis explained, the A’s played in
a competitive league where:
- Revenue-sharing was not implemented,
- A salary cap had not been imposed, and therefore,
- Some teams could and did outspend others by orders of magnitude.
Not one of those things is true in the NFL.
In 1998, Billy Beane’s first season as GM, the A’s had a payroll of $21,473,000. They were 28th in baseball. The #1 team in payroll, Baltimore, spent $72,525,634– more than triple the A’s. By 2002, the year Lewis wrote the book, the A’s payroll was
still 28th. It had nearly doubled, to $40,004,167… but the #1 payroll in baseball (The Yankees, with $125,928,583 ) was still more than triple.
The A’s adopted the work of Bill James because they had no alternative. They **could not **operate the way the Yankees, Red Sox and Rangers did– they didn’t have the money. They had to find a method to get $3 of value for every dollar they spent.
Because that was true, one of Billy Beane’s tasks every season is to:
- Determine which players (based on their production) are making more than he can afford to pay– or will soon reach that point, and
- Get rid of them in exchange for players who currently make less and either (a) provide more bang for the buck or (b) will become better players
In the NFL, every team receives the same amount of operating income; every team works under a ‘hard’ (no exceptions) salary cap. The Oakland Raiders haven’t been losing because they don’t have as much money to work with as the New York Giants– they lose more because they have behaved more stupidly.
This is why the A’s have struggled so immensely since the book came out– from 2007-11, they didn’t have a single winning record. For a while, they could win because nobody knew they were doing this stuff. When other teams realized this numbers stuff helped them find players, they began doing it and it wiped out Oakland’s method of enforcing competitive balance. The money returned to being the deciding factor.
If DePodesta comes to the NFL and decides to play Moneyball the way the A’s do it, the first thing he will do is to trade Joe Thomas and Joe Haden. His reasoning will be that:
- Haden ($13.5 million) and Thomas ($11.5 million) are the highest-paid players on the team
- Thomas is 31 and plans to retire after the 2018 season
- Haden is 26 and is under contract until 2019
- The Browns are 2-3 years away from contending
- The players will not help the team when the roster is ready
- Other teams will give more in value (in the form of draft picks) than the Browns can get from Thomas or Haden
- The team can use the money saved from replacing Thomas with Austin Pasztor and Haden with K’Waun Williams or Charles Gaines to sign free agents
- It benefits them to do this
But every team has the same amount of money to spend– and NFL rules require teams to spend at least 90% of their salary pool. So there is no reason to trade Joe Thomas– he was just named first-team All-Pro; he’s the best player at his position, and left tackle is one of the most valuable positions on the team. As is cornerback.
You might think about trading Haden– but you would do it because he is grossly overpaid, given his production, and has injury problems. You wouldn’t do it because you can’t afford to pay him as much as anyone else does.
The NFL doesn’t have a development league, so DePodesta can’t suggest trading players for minor leaguers. And because NFL draft picks are so valuable– people can become stars the season they are drafted– GMs won’t trade them often.
Billy Beane would have traded Alex Mack and Tashaun Gipson in
October, knowing he couldn’t keep them.
Paul DePodesta played football in college; he also isn’t stupid. He knows things like that.And I certainly don’t want to sound like Mike Tomlin did in this excellent Tom Reed piece:
“You can take analytics to baseball and things like that, but football is always going to be football. I got a lot of respect for analytics and numbers, but I’m not going to make judgments based on those numbers. The game is the game. It’s an emotional one played by emotional and driven men. That’s an element of the game you can’t measure. Often times decisions such as that weigh heavily in the equation.”
But Tomlin isn’t entirely wrong. Building a game plan based on the opponent’s tendencies (”On third and 5, draw plays got first down 92% of the time and produced 20 yards of better 59% of the time”) has been tried since the 1960’s and not worked well.
DePodesta will need to adjust his thinking– or stay away from issues that he can’t codify. This should not be a problem, because there are plenty of subjects you can get value from– so many that you can ignore whole areas and still stay fulled booked.
But to the degree that DePodesta can’t or won’t, he’ll fail.
I’m Chevy Chase and you’re not
I think I met Paul DePodesta once, for a grand total of 45 seconds, more than 20 years ago. Everything I say in this part is second or thirdhand; I do not know people who know him intimately.
That disclaimer given, my understanding of DePodesta is that he does not play well with others. As one person put it, he thinks he is smarter than everyone else in the room. (Which is usually true, but not something that endears others to you.) He tends to assume he has thought longer and more deeply about an issue than others (which is often, but not always, true). He assumes he knows more about the issue than others (not always true) and that his insights are the only sensible viewpoint (often not true).
I work as a consultant– I have to walk into sites where people either hate or fear me on sight. An unapproachable personality is a handicap in a situation where you need to win people over.
If these rumors are true, it would explain why (as someone else told me) DePodesta works best in a behind-the-scenes role where he has autonomy over his fiefdom. When he needs to get buy-in from others (managers, coaches, players, media, fans), he runs into issues.
The reports I have might be wrong. But they’re at least as valid as the endorsements all his
friends have been giving to the media. The local media always offers a 100% positive assessment when someone arrives… and those (Bowe is serious about having a great year, Starks and Williams haven’t lost a step, Cameron Erving is NFL-ready, Justin Gilbert has a great work ethic), often as not, are wrong.
The Elephant in the Room
The final point– well, let me quote Reed’s piece here:
“Will the new regime demand a strict adherence to analytic thinking? Will it turn off and reduce qualified candidates? Or, will Brown and DePodesta, known for their ability to work with others, arrive at a happy medium?”
The answers, respectively, are “Yes”, “Yes” and “No.”
1. The Browns will need to insist that the coach go along with analytics if they want to get any value out of it. There’s no point to using analytics to make strategy if the people in charge of implementing it are opposed to it.
Beane realized that with Art Howe. Howe, though he wasn’t nearly as bad as the movie makes him, didn’t care for all the juggling by numbers. Like most managers, he wanted nine guys he could play every day. He lasted only five seaons, leaving the A’s before the book was published. If you don’t have everyone on the same page, the system won’t work.
2. DePodesta’s presence will make hiring much harder. A lot of coaches and GMs were going to rule out the Browns anyway. The convoluted front office structure– putting Brown in charge– eliminated still more.
Then you take the lawyer who has never worked in personnel– and have him report to a guy who crunches a bunch of numbers (which many coaches and GMs think is crapola) who has never worked in football.
Plus the owner (who might have ordered the last GM to draft the drunken quarterback) fires coaches and general managers about every 18 months.
How would you think that will go over? Who in his right mind would want to work in that environment? There are four ways this situation can turn out:
- The Browns find a talented coach who likes analytics and gets along well with Brown and DePodesta. And that coach has a younger brother who has the talent to be an All Pro defensive end, a cousin who can play quarterback. And a pony who poops hot dogs and pees soda pop that the fans can enjoy.
- They end up with an assistant linebackers coach from Tennessee– someone who might be talented, but is years away from deserving a head coaching position.
- Cleveland gets a talented assistant coach in his 60’s (who’s been passed over repeatedly) or a head coach who was fired amid embarrassing circumstances. Tom Cable went 5-11 and then 8-8 with Oakland– then he was the running game coach for Pete Carroll … but he popped off a lot, got into a fistfight with one of his
assistants and broke the guy’s jaw.
- Haslam watches 2-3 people slip away, gets desperate and tears up the structure and promises a candidate that he’ll run the show. Or, more likely, he’ll tell the coach he’ll be in charge to get him hired and then tell DePodesta and Brown that he didn’t mean they could have total control– or override the coach.
Why couldn’t it work out? Very simply.
Bad Teams Do Everything Badly
When you consult, you soon learn there is no such thing as a great client that is in trouble for reasons entirely beyond its control. If a company is losing money– if a political candidate is behind in the polls– if a sports team loses most of its games– it happens for a reason. Occasionally you can find a corporation that has been built by
acquisitions– where one of the businesses was run well before it was bought, and hasn’t gone downhill yet. But there is no such thing as a company that achieves excellence in one area– but fails miserably in others.
When companies are bad, they do everything badly. Even if they buy a successful company, they eventually drag it down. The notion that Jimmy Haslam is a smart guy– a highly skilled CEO– who has run a team for four years and gone 19-45 because he’s had some bad luck is ludicrous. Since Haslam has been lying about how long he has been in charge, let’s walk through this:
- Haslam bought the Browns after they went 4-12 in Pat Shurmur’s first year. He hired Joe Banner to run it.
- Banner immediately ran Mike Holmgren off (not a bad move). After the season, even though the team improved to 5-11, Banner fired Tom Heckert and Pat Shurmur, making it clear that their fate was sealed the day he took control (again with Haslam’s permission). All three men went to other jobs, telling people how screwed up the Browns were.
- Banner implements a system that puts him in charge of the roster, even though he has no operations experience. He hires a GM (Mike Lombardi) who’s been fired repeatedly and a coach (Rob Chudzynski) who’s been passed over repeatedly, because they’re willing to let him run the show. Haslam lets that happen.
- The team goes 4-12. Banner fires Chud– and, more importantly, Norv Turner and Ray Horton. Haslam permits Banner to do that. All three men go off to other teams, telling stories about how terribly the Browns are run.
- Banner and Haslam hire Mike Pettine. Then, after Banner has fired the coach and after it is too late to hire a competent GM (there isn’t enough time before the draft) Haslam fires Banner and Lombardi and promotes Snapchat Ray Farmer.
- On draft day, the team trades a #1 and #3 to move up and draft a guy who throws like Tim Tebow and drinks like Tim Tebow’s evil twin. Haslam tells a story about the draft room that suggests that Haslam made the decision to make the pick. He only stops after he realizes that it doesn’t reflect well on himself or the team.
- The team goes 7-9, but the offensive coordinator– a respected veteran– bolts, complaining that the GM was interfering with him. The Browns hire one of Pettine’s buddies to replace him, making two coordinators who have never run their units. Haslam permits this.
- The Browns fall to 3-13. Haslam fires everyone, but tells four assistants he wants to keep them.
- He devises an organizational scheme that no one in the NFL uses– or has ever used. He puts two people who have never run a football team– never been coaches, scouts or worked in football operations– at the controls.
- He puts the coach outsider the hierarchy, reporting directly to himself.
If you think DePodesta will succeed– that the system will work– let me ask you a question: What basis in fact do you have for making that silly assumption? What logical assumption can you even offer?
The Browns are a bad team in pretty much every respect. They draft badly. They sign free agents who flop; the people they let go succeed with other teams. They negotiate contracts that are not favorable to the team. They used the franchise player and transition tags wrongly. They’ve hired coaches who got arrested, GMs who got suspended and inept scouts. It hired terrible, boring announcers for the pre-season
games; their TV shows are good only for curing insomnia. The redesigned uniforms are hideous. People who go to the redesigned stadium are unimpressed by it; it has bad, overpriced food, parking problems, insufficient public transportation to and from games and plays obnoxious music very loudly.
So suddenly this team will design a new organizational paradigm that functions well, hire the right people to implement it and win big?
On Sashi Brown and Alex Mack
In writing that litany of errors, something just occurred to me. Something that ought to take the bloom off the Sashi Brown rose for you.
When Brown was put in charge of the roster, the Cleveland media– which always has to make every hiring seem like “greatest move ever”– pointed out that he had been in charge of the salary cap– and all contract negotiations since he joined the team in 2013. He’s contributed behind the scenes and made incredibly great decisions. He is a Wile E. Coyote-style Supergenius…
Maybe so. But let me ask a question: If Brown is so smart, why did Cleveland use the transition tag– not the franchise tag– on Alex Mack in 2014?
The Browns are about to lose Mack in free agency. In 2014, he signed an offer sheet with Jacksonville, which gave him:
- $10 million, fully guaranteed, in 2014
- $8 million, also fully guaranteed, in 2015
- The right to void the final three years of the deal after the 2015 season
- $8 million, also fully guaranteed, in 2016
- $6 million, not guaranteed, and a roster bonus of $2 million in 2017
- $6 million, not guaranteed, and a roster bonus of $2 million in 2018
From the point of view of the Browns, it is a terrible deal. Mack has made $18 million over the last two years (the highest-paid center in the NFL), now he can sign a new contract. Since he was just named to the Pro Bowl a third time, his new deal will be much higher. It could be with Cleveland, but more likely will not.
Jacksonville was allowed to make Mack an offer because the Browns used the transition tag on him. Had they used the franchise tag, they couldn’t have.
The franchise tag requires the team to pay the player an annual salary that matches the average of the five highest-paid players at his position,. Because they average offensive linemen, Mac would have been paid like a left tackle– about $24.5 million, instead of $18.
But the Jaguars couldn’t have made the offer. And, after Mack broke his leg, Cleveland might have been able to negotiate a long-term deal (players are always more receptive to a long-term deal when they’re recovering from a season-ending injury).
Instead he is most likely leaving. I mean, he said winning would be important to him, and he has no chance to do that here.
Perhaps Brown recommended the franchise tag, and Ray Farmer overruled Brown’s wise counsel. But if he did, how was that possible? How could Brown have the authority if he couldn’t decide when to franchise a player and when to use the transition tag?
I’ve always assumed that Snapchat made the call– it’s the sort of rookie mistake a rookie GM makes. But the novice might have been Brown… in which case, we all might owe Snapchat an apology on that front.
So if you’re assuming this might go well, we already have one reason to say it won’t.