The Arizona Diamondbacks made two major moves this offseason by signing Zack Greinke to the most lucrative contract in baseball history (based on average annual value) and acquiring Shelby Miller from the Braves. The rest of the baseball world immediately panned these deals. Ken Rosenthal (FoxSports) wrote:

“The contract is insane, everyone knows it’s insane, and just as with the Red Sox and David Price, the only question is when the Diamondbacks start to regret it.”

Regarding the Braves’ haul for Shelby Miller, Jayson Stark (ESPN) said he has heard the deal described as the “heist of the decade”.

When you see decisions that could cripple a team from both a financial and talent perspective, it makes you wonder, who’s running that team’s analytics department? The answer: Dr. Ed Lewis.

During the early 2000s, Dr. Lewis was a special assistant to Tony LaRussa in St. Louis and has reunited with his old partner-in-crime in the desert. Lewis runs Arizona’s analytics department. At first glance, Dr. Lewis looks like a smart investment by the Diamondbacks. He has an advanced degree, for example. The problem is that Dr. Lewis, while obviously smart, has no formal mathematical training: he’s a vet.

No, not former military. An animal doctor.

It gets better. When a local newspaper asked Dr. Lewis about his qualifications, Lewis stated that he had been a veterinarian for 18 years and has traded “pretty aggressively” on the stock market.

To be clear, Ed Lewis isn’t solely to blame for the Diamondbacks’ disastrous decision-making. The veterinarian profession is noble. But skepticism regarding analytics runs deep out in the desert, evidenced by Dave Stewart, the Diamondback’s GM, stating, “We’ll use it [analytics]. It stops before the first pitch is thrown. … It’s not that we devalue it. We value it when it’s used appropriately. We do not value its intrusion into the game.”

It’s this attitude at the top that drives the Diamondbacks’ weak investment in the maths.

What’s more, franchise-crippling contracts aren’t rare. Consider: Barry Zito (7 years, $126M; 4.4 total WAR); Ryan Howard (5 years, $125M, -2.2 WAR in the first four years); BJ Upton (5 years, $75M; 0.4 WAR), to list a few.

The argument for better math goes like this: It can help clubs avoid contracts like those while also identifying talent to for organizations to lock up before players hit arbitration. The poster child for this argument is Evan Longoria. The Tampa Bay Rays signed him to a 9-year deal only six days after his MLB debut. Longoria’s deal will pay $47.6M through the end of 2016. He has already produced roughly 42 WAR (more than $200M dollars in value) for the Rays.

Big-contract mistakes could have been avoided with investment in analytics. For example, the best research on aging curves in the post-PED era shows that players age more quickly and aging varies by position compared to the steroid/amphetamine era. Investment in forecasting models that added more accurate regression projection for players after the Joint Drugs Agreement might have led to the conclusions that clubs needed to give older players more days off and also avoid signing long term deals until the labor market stabilized. During the transition years, teams didn’t just lose one or two million dollars, but tens of millions.

Consider the following thought experiment: Say a robust baseball analytics department improves a team’s decision-making, on average, by one percent. In 2015, the average payroll for a MLB team was $113 million. What does $1.13M buy you in the analytics world? (Short answer: a lot).

The Low Cost of Good Math

The Society for Human Resource Management estimates the average starting salary for someone with a BA in Science Technology Engineering and Math in the United States is $43,000. Payscale estimates that a masters degree in a STEM field will up the starting pay to $80,000-$89,000 and hit $120,00-$171,000 mid-career.

If you make the poor life choice to go for a Ph.D., an academic statistician can expect between $80,000-$94,000 their first year and $120,000 mid-career (AMSTAT). For academics in the field of physics, pay is a bit lower.

In the private sector, Payscale estimates that a director of analytics will make between $87,000 and $187,000 in total compensation. These numbers are pretty close to the estimate at other websites. Glassdoor estimates that the average salary of a statistical consultant is about $75,000.

Remember that $1.3 million – the cost of shaving 1 percent off payroll by using advanced math? Let’s use it to hire an experienced director of analytics at $190,000, three researchers with MS degrees in statistics at $95,000 each, and six data crunchers at $50,000. Our analytics department would cost $775,000/yr. Add in software and technology and you’re looking a cost of $925,000 for a fully armed and operational battle station.That hypothetical firepower would be at the top of MLB in quality and quantity.

Back in the real world, it’s hard to get an accurate accounting of analytics department staffing and spending. I’ve heard the median department is around 3 full-time people.


The Reds are moving in the right direction. They promoted Sam Grossman from director of analytics to assistant GM. The front office added two additional positions to the Reds analytics division.

These changes, though, are incremental and limited. More aggressive spending, not only on entry-level people, but experienced analysts with advanced degrees who have worked for other baseball organizations, presents a real opportunity.

The Reds rebuilding effort is often compared in a positive way to what the Cubs and Astros have accomplished in recent years. Those clubs made decisive, all-in commitments to analytics, not half measures with one foot in old-school methods and another in modern thinking.

Like many decisions in business, how you weigh cost depends on what you think of the benefit: If the team leadership doesn’t believe analytics can improve baseball decisions, then it doesn’t matter if the Reds compare these costs to their annual revenue ($227 million), player expense ($130 million), or team value ($885 million).

We’ve heard the maxim that if you think knowledge is expensive try ignorance.

Well, if you think analytics are expensive, go ahead and sign Jason Marquis, or Willy Taveras, or Kevin Gregg, or Skip Schumaker, or Brennan Boesch or …