Spreadsheet ball isn’t for playoffs
Hooking Snell was doomed with a weary bullpen. The Rays learned that the hard way
Contrary to the public reaction and barrage of headlines that came out of the Los Angeles Dodgers being crowned World Series champions on Tuesday night, analytics are not ruining baseball.
Advanced statistics are tools that help organizations make informed decisions. They remove biases and take the emotion out of personnel moves that happen on and off the field. Eyes can be deceiving, numbers don’t lie.
Believing in the data is what got the underfinanced Tampa Bay Rays to the World Series. The philosophy also might have cost them a shot at the title following manager Kevin Cash’s controversial decision to remove left-hander Blake Snell during the sixth inning of Game 6. The move backfired, turning a 1-0 lead into an eventual 3-1 loss.
There was a predictable amount of outrage about the decision on social media. “So, who gets to pull the manager?” asked Mets righty Noah Syndergaard. “The best pitcher in the ballpark tonight just got pulled after giving up a single!” lamented Canadian Ryan Dempster. “The Rays front office and computers may end up being the MVP for the Dodgers,” chimed in Alex Rodriguez.
The problem isn’t analytics, it’s what teams are doing with the data, often blindly following previous numbers with little regard for what’s happening in the moment. There has been an increasing trend across baseball, including the Blue Jays, of front-office executives huddling with coaches before games to map out every possible scenario hours before it happens.
Over a 162-game season, or even 60, that strategy usually pays off. When the sample size is big enough, trends form and outcomes can be reasonably predicted. A big spike often will be ignored because over time performances level out and things work out as expected. If it doesn’t work out today, it probably will tomorrow. There are decades of stats to back this up.
The post-season is a different beast. Past performance is not always indicative of how players are going to perform within a condensed period. Anyone else remember 1992 World Series MVP Pat Borders turning into Babe Ruth for about a week? Or what about R.A. Dickey turning into a pumpkin against the Kansas City Royals in 2015? There will be outliers in the post-season and while it makes sense to use the numbers as a guide, common sense needs to be applied as well.
The data suggests Snell’s performance — like a lot of pitchers — typically goes down as the game progresses. In
Snell’s career, opponents have a .592 OPS the first time they face him in a game, .711 the second and .742 the third. That track record is at least one reason why Snell did not complete six innings once this year.
During the regular season, the strategy was understandable. The Rays bullpen was arguably the best in baseball, second to the Oakland Athletics with a 2.72 ERA and only the Boston Red Sox threw more innings. Cash previously boasted about his “stable” of arms, including a slew of guys who throw with upper-90s velocity. Relief was one of Tampa’s biggest strengths.
The issue is that the bullpen Cash had for Game 6 was not the one he used throughout the year. A full month of playoff games took its toll. Nick Anderson, the pitcher who replaced Snell, allowed one earned run during the regular season. He finished the postseason by surrendering at least one in each of his final seven appearances.
Anderson conceded fatigue was a factor, but that should have been obvious to the Rays before the game began. One of MLB’s top relievers had been struggling for a while and he was not alone. Tampa Bay’s relievers, who had been so good throughout the year, were mediocre against the Dodgers with a 4.32 ERA. An adjustment needed to be made. It wasn’t.
Instead of letting Snell face the top of the lineup a third time, against batters who were 0-for-8 with seven strikeouts through five innings, Cash fell back on organizational philosophy. Tampa went with Anderson because that’s what the numbers suggested made sense.
The reaction in the Dodgers dugout said it all. A pitcher they couldn’t touch was removed in favour of a pitcher who had been struggling for weeks. There were signs of life, confidence skyrocketed and after a Mookie Betts double, a wild pitch and a fielder’s choice, the Dodgers had a lead they would not relinquish.
“I was pretty happy because he was dominating us, and we just weren’t seeing him,” Dodgers manager Dave Roberts said. “We were all kind of excited that Snell was out of the game.”
The decision to remove Snell was at least somewhat reminiscent of the Jays’ wild-card opener against the Rays when right-hander Matt Shoemaker was lifted after three scoreless innings for Robbie Ray. The veteran lefty got through another three, but gave up a run in his first inning of work.
The Jays primarily lost because their offence couldn’t score enough runs, but the early hook didn’t help. One could argue momentum doesn’t exist in baseball from one game to the next, but it does within the confines of nine innings. The Jays gave it away to the Rays and they returned the favour weeks later to the Dodgers.
Stats are obviously useful, but they can’t be the only thing teams look at when making decisions. Managers are supposed to be hired because of their experience, feel for the game, leadership and calmness under pressure.
When front offices hand their skippers a script to follow, the human element is lost and their in-game decision-making instincts become irrelevant. Over 162 games that approach might be defensible, but within the context of a seven-game series there must be more to it than that.
Analytics should be one of the main ingredients, but they can’t make up the entire meal. Spreadsheets can’t be fired, however the people who are running them can be, even the ones with fancy Ivy League diplomas.