Alex Baltazar

Deciphering the Astros winning formula

The Astros seem to have a formula for everything. Photo by Bob Levey/Getty Images

Jim Crane inherited a depleted roster and what was widely considered the worst baseball farm system in the majors when he bought the Houston Astros in 2011. Analytics wasn’t much of a focus for the Astros when they transitioned into the new regime. That remained the case until Jeff Luhnow interviewed for the newly available GM position. Luhnow pitched to Crane that he wouldn’t make ‘cosmetic decisions’, such as acquiring expensive players that won’t yield success in the long term. Rather, he would use data analytics to make acquisitions that were fiscally responsible, and once they had a winning team that generated profits, they would bring in higher priced players. Jim Crane bit. Crane’s previous ventures in shipping and logistics often utilized data analysis to facilitate certain aspects of his businesses. “If you have better information, faster than your competitors, you can run ‘em ragged.”, Crane explains in Ben Reiter’s book Astroball: The New Way to Win It All.

Initially, it was extremely difficult for Luhnow to introduce analytics as a guide to decision making for scouts and coaches to use. Components such as behavioral changes from lineup decisions to defensive configurations were also necessary to use in conjunction with Astros’ analytics. “That was harder, and took three or four years to get to a point that we felt good about it,” said Luhnow when reflecting upon his early portion of his tenure as GM in the Quarterly.

The human side of the Astros’ decision making is still vital to this day. The idea of implementing traditional evaluations on players from ‘old school’ coaches and using them alongside analytics was something that had never been successfully accomplished before.

Data collected from qualitative observations, or soft data, like leadership, ambition and desire are converted into quantitative numerical values. These values are then also plugged into the mathematical regression models that the Astros utilize to evaluate a player’s potential. All while trying to identify and avoid decision making psychological defectors like cognitive bias which leads to irrational decisions and unreliable heuristics or mental shortcuts.

How do you quantify a player’s propensity to get injured? A player’s motivation or drive? In ‘Astroball’, Reiter delves into how the Astros applied these soft data to their regression models. Again, how do you quantify motivation, and what does it even mean or include?

Situational dilemmas such as these pose a major challenge for econometricians, like Luhnow, who try and predict behaviors in healthcare consumption, causes of poverty in third world countries, or identify production deficiencies in big businesses. The Houston Astros, however, have seemingly mastered soft data quantification.

A challenge so major that it led Chris Correa to federal prison for 46 months because he was continuously hacking into the Astros’ databases. Correa was the scouting director for the St. Louis Cardinals when Deadspin leaked rumors that ignited an FBI investigation to look further into the allegations. Later, he was found guilty of 5 counts of unauthorized access of a protected computer. The Cardinals organization was fined two million dollars and handed over two top draft picks for the 2017 draft. The extent of the breach is still difficult to assess.

Data based on verified measurables or hard data, however, is easy to come by and interpret. Many online sources contain databases with all the quantifiable statistics and metrics baseball players are judged and ranked by. More tangible statistics, such as On Base Percentage (OBP) or a pitcher’s spin rate are also used and plugged into the Astros’ regression models.

The historic 2002 Oakland A’s ‘Moneyball’ team heavily relied on hard data and its inefficiencies. The misconception of the Houston Astros is that they follow a similar method. They don’t. Every team in modern baseball uses data analytics. The difference from that Oakland A’s team and today’s Houston Astros team is that the Astros heavily rely on their scouts’ and coaches’ gut instincts and experiences, alongside sophisticated data analytics to make decisions on acquiring the right players for the right price, via draft, free agency, or trades.

The Astros new way of winning it all is an organic method. They are learning as they go. The 2015 release of now, 2018 MVP frontrunner JD Martinez, can induce endless speculation of what could’ve been. It may also be difficult to forget when Jeff Luhnow and his then fellow St. Louis Cardinals’ scouting department passed on a generational player like Mike Trout in the 2009 draft.

Luck just as it does with anything else also plays a factor in winning. Whether it’s acquiring Justin Verlander with merely two seconds to spare from the closing of the trade window last season, or having Brady Aiken leave the Astros after negotiations fell apart with a reduced offer due to ongoing health concerns, which in turn led to the opportunity for Alex Bregman to join the good guys via the 2nd overall pick in the 2015 draft as compensation. You can argue that without either of these two, the Astros wouldn’t have gone on to win the World Series.  

The future holds an ever-growing investment in artificial intelligence to produce faster and more accurate results. That is the key to identifying new and innovative ways of concurrently capturing hard and soft data and implementing it faster than your competition. Combining analytics and technology with human experience and instinct, will give you winning results at the end of the day.

Special thanks to Mohammad-Saqib Aziz, Economics student specializing in Quantitative Analysis at the University of Houston, for your contributions to this piece. Also, I encourage all to read New York Times Best Seller, Astroball: A New Way to Win It All, by Ben Reiter. You can order a copy here.

Most Popular

SportsMap Emails
Are Awesome

Listen Live

Photo by Getty Images.

It's easy to blame Bill O'Brien for the Texans woes. He is a lousy GM, a below average play caller and his offense is boring, predictable and ineffective. Not once has he had a top 10 offense in the league. So he does not get a pass here.

But Deshaun Watson shouldn't, either.

Last year, Watson was in the MVP conversation entering the game in Baltimore. Four of the nine games he played before that, Watson had an ESPN Total QBR over 85, which is playing at an elite level.

Since that 41-7 debacle (where his QBR was 13.6), Watson has played 10 games. He has topped 85 just once (and barely - 85.6) in the win over the Patriots. While QBR is not the be all end all, it shows a trend. And before you blame the talent around him or the ridiculously stupid DeAndre Hopkins trade, eight of those games were with Hopkins in the lineup.

Over his last 10 games, Patrick Mahomes has done it five times (and just missed last week at 84.7). Lamar Jackson has done it six times in his last 10. Russell Wilson is six for his last 10. Dak Prescott? Three. Aaron Rodgers? Three. Ryan Tannehill? Three. Josh Allen? Two. Lamar Jackson led the league last year with an 83 for the season. Watson was sixth at 71.3. To be a top 10 quarterback, you had to average 64.1. In two games this season, Watson sits 20th, about where he was over the last six regular season games and two playoff games last year.

In essence, Deshaun Watson - who often gets compared to those players - is not on their level. Yes, O'Brien has a lot to do with it, but it's also time to start looking at Watson's performance and regression as an NFL quarterback.

In 2018, Watson had four such games. In 2017, four in six starts. And now ONE since that Baltimore game. In fact, he has topped 80 just once in that stretch, and 60 just three times.

What it tells us is Watson has been an average quarterback over his last 10 starts. The Texans invested heavily in an offensive line to protect him. They have added depth at WR but a net loss without Hopkins. Elite quarterbacks turn in performances like that roughly half the time. Getting more consistent has always been an issue for Watson. But since that Baltimore game, he has not been close. And he is being paid to be elite.

In the end, O'Brien is still the main culprit. He has hand picked all the players around Watson, he designed the offense, and he controls everything.

But it's time to quit giving Watson a pass. Right now, he is part of the problem.

SportsMap Emails
Are Awesome