What Makes Good Process?

Letting Russell Martin bat, but pinch-hitting for Alex Rodriguez. The idea of evaluating process vs. outcome is nuked for a while.

— Joe Sheehan (@joe_sheehan) October 11, 2012

AP Photo/Kathy Willens

One of the responses to the unexpected Alex Rodriguez – Raul Ibanez pinch hit situation was criticism of manager Joe Girardi’s “process”. The idea is, statistically speaking at least, that Alex Rodriguez carried better playoff and seasonal numbers in to the at bat, and Joe’s move to Ibanez was influenced by other factors such as the media’s criticism. As a staunch defender of the third baseman, I was a bit shaken by the move as well. Rodriguez was the hero of 2009, and he’s had other success in the playoffs that get overlooked, in fact his numbers indicate that he’s been better over his career than playoff legend Derek Jeter. With such great statistics, why did he have his at bat taken away?

I’ll have my buddy Paul DePodesta, ex-GM of the Dodgers and Padres, now VP of player development and scouting with the Mets, and more famously the sidekick in Moneyball, explain the idea of process in baseball.

“Well, I spent the rest of that weekend wandering around the casino, largely because I had lost all of my money playing blackjack, thinking about all of these different games and how they work. The fact of the matter is that all casino games have a winning process – the odds are stacked in the favor of the house. That doesn’t mean they win every single hand or every roll of the dice, but they do win more often than not. Don’t misunderstand me – the casino is absolutely concerned about outcomes. However, their approach to securing a good outcome is a laser like focus on process…right down to the ruthless pit boss.

We can view baseball through the same lens. Baseball is certainly an outcome-driven business, as we get charged with a W or an L 162 times a year (or 163 times every once in a while). Furthermore, we know we cannot possibly win every single time. In fact, winning just 60% of the time is a great season, a percentage that far exceeds house odds in most games. Like a casino, it appears as though baseball is all about outcomes, but just think about all of the processes that are in play during the course of just one game or even just one at-bat.”

The whole article is a very short and interesting read, but what it comes down to is an issue of outcome versus process. The point of advanced analytics in baseball is to create a more accurate portrayal of outcome when there is small sample size. For instance, if a player has a bad season hitting, we can use their BABIP on batted ball rates to get a better idea of instances of bad luck. Here, we have a bad outcome, however with sabermetrics we know to keep that player around for when the BABIP evens out.

Obviously the theories and numbers are far more advanced than simply BABIP, but DePodesta struggled early on to prove that using advanced numbers is good process. Often he would see bad process where the outcome would be good, which would leave his philosophy in question. In the same article, he credits Billy Beane with the patience and understanding to see past bad process and stick with what the numbers say.

The casino example is an interesting one that implies good process is based on trusting numbers, but is that really the case? I suppose I used to think so, up until Raul Ibanez hit that homerun. But is this just another instance of dumb luck that haunted poor decisions in the game for years? The more I think about, the more I believe that Joe Girardi made the right move.

Going back to the initial struggle that DePodesta faced when introducing sabermetrics into front offices, many old school scouts disagreed with the idea. In Moneyball, there were two radical ideas clearly painted out by Michael Lewis, and the ending was largely interpreted as the nerds winning, but I think this misses the point. In the end, the true innovator was Billy Beane, who listened to both scouts and data analysts to come up with the correct decisions.

The idea of good process isn’t black or white between old school scouting and number crunching, it’s a gentle balance between both. Through the decades of scouting, and the years and experience these guys cumulate, there is sincerely something of value in what only the eye can see. Perhaps there will be a day where numbers can recreate the seemingly intangible today, but for now we’re still best left trusting the old guys on certain things.

So why is this relevant? I think Joe Sheehan has the idea of good process based in numbers. Joe Girardi doesn’t make $3 million a year, and isn’t put in charge of a $200 million team, to follow numbers in a binder. Girardi has played the game, he’s watched countless others play the game, and he knows his players. Alex Rodriguez may have great post season numbers, but even to my eyes I can see that all these whiffs on fastballs are not normal.

Baseball is a game played by men, which while predictable by numbers in the long run, is prone to small sample size errors. Good process is a combination of using data for long term answers, and your experience with the short term. Girardi kept the red hot Russell Martin in the lineup and removed the scuffling Rodriguez, replacing him with an opposite handed hitter with great recent numbers. This is good process if you ask me, and it’s a necessary evil when you don’t have 162 games to wait for numbers to even out.

Follow Me On Twitter