One of the most scrutinized play calls in NFL history occurred in Super Bowl XLIX. The Patriots led the Seahawks, 28-24, and with 26 seconds to go, Seattle was on on the New England 1-yard line with one timeout remaining.
The Seahawks called a pass play, a quick slant. Patriots cornerback Malcolm Butler jumped the route and made the interception.
Seconds later, NBC analyst Cris Collinsworth said, “I’m sorry, but I can’t believe the call. … I cannot believe the call. You’ve got Marshawn Lynch in the backfield. You’ve got a guy that has been borderline unstoppable in this part of the field. I can’t believe the call. If I lose the Super Bowl because Marshawn Lynch can’t get it in from the 1 yard line, so be it. So be it! But there is no way… I don’t believe the call.”
Many agreed; Hall of Fame running back Emmitt Smith even called it “the worst play call in the history of football.”
The Tampa Bay Rays are a “Moneyball” baseball organization. Lacking massive local TV contracts and (even pre-coronavirus) significant gate receipts, the Rays are like bootstrapping entrepreneurs who can’t afford to throw money at problems. So they challenge conventional wisdom. They harness the power of analytics. They develop players, effectively promoting from within. Their free agent signings are judicious rather than splashy.
More so than for many teams, success for the Rays comes down to making smart decisions.
Which is why Tampa Bay Manager Kevin Cash is facing heavy criticism for pulling starting pitcher Blake Snell in the sixth inning with a 1-0 lead over the Dodgers.
Even though Snell had not pitched six full innings since July. And pitchers are statistically at a disadvantage when they face batters for the third time in a game. And the heart of the Dogers lineup was coming up.
On the other hand, Snell was pitching on a full five days of rest. He’d given up two singles, no walks, and had struck out nine. His pitch count was low. And Snell’s replacement would be Nick Anderson, a reliever had given up at least one run in each of his last six appearances.
“I was pretty happy (when Snell left the game) because he was dominating us and we just weren’t seeing him,” Dodgers Manager Dave Roberts said after the game. “Once Austin (Barnes) got the hit, Mookie (Betts) looked at me with a little smile. He was excited Snell was out of the game.”
Snell was visibly upset as he left the field. “I wanted to go that whole game,” he later said. “That was everything I wanted to do: burn the tank and see how far I could go.”
For good reason: After just six pitches, the Dodgers led the Rays 2-1 on the way to a 3-1 victory that decided the World Series.
But here’s the thing: Some decisions can’t be evaluated solely on the outcome. Sometimes decisions — especially decisions based on incomplete information and uncontrollable variables — can only be evaluated on the decision-making process itself.
Take Butler’s Super Bowl interception. In hindsight, a pass play was a terrible call.
- The Seahawks only had one timeout left. Running the ball on second down and failing to score would have likely meant burning that last timeout. Odds are the Seahawks could only have run one more play. An incomplete pass would have only cost a few seconds; the Seahawks could then have run the ball on the next play and, if necessary, call a timeout and then run a third play. Three opportunities to score are always better than two.
- Statistically, interceptions occur on just slightly over 3 percent of passing plays from the 1-yard line. (In fact, Butler’s was the only interception on the goal line all year.)
Could Jermaine Kearse have set a much better pick? Could Russell Wilson have made a slightly better throw? Yep. Did Malcolm Butler make a great play? Absolutely.
All of which are variables impossible for a coach or manager to control.
Hindsight bias makes it easy to criticize Cash. We now know the outcome. We now know Snell was upset, making it even easier to assume he would have fared better than Anderson. We know.
Just like we know that Hall of Fame pitcher Pedro Martinez successfully argued with Red Sox Manager Grady Little to stay in Game 7 of the ALCS in 2003.
And then quickly lost a 5-2 lead, setting the stage for Aaron Boone’s 10th inning walk-off home run and a Yankees win.
It’s easy to think of decisions as being right or wrong. It’s easy to decide a decision was incorrect based on the outcome. (After all, most decisions that turn out well, even if mostly due to luck or circumstance, are rarely scrutinized.)
But whether a decision worked out well or badly is often not an indication of the quality of the decision.
In the moment they are made, many decisions are no better than predictions.
Nine times out of ten, the Seattle’s decision to run a pass play might have been, if not successful, at least not spectacularly unsuccessful. Six times out of ten, Cash’s decision to pull Snell might have panned out. (I don’t know the actual probability, but I’ll bet Cash does.)
Each decision — just like many of the business decisions you make — is what Jeff Bezos would call a 70 percent decision.
As Bezos says, “Most decisions should probably be made with somewhere around 70 percent of the information you wish you had. If you wait for 90 percent, in most cases, you’re probably being slow.”
Why? While it would be great to have access to all the data and statistics and information and business intelligence you need, that’s almost never possible. All you can do is make the best decision you can with the information currently available.
Should Cash have placed more faith in what he was seeing from Snell in the moment, rather than relying on what historical data and analytics indicated? That’s a question many are asking.
That’s a question Cash and the Rays organization will certainly ask themselves — not just to second-guess themselves, but to improve their process for making decisions.
Even though after the game Cash said he regretted the result but not the thought process, and that he would make the same decision again since he “didn’t want Mookie or Seager seeing (Snell) a third time.”
Because outcomes are not always the best way to evaluate how you make decisions.
Especially when people — with all their inherently uncontrollable variables — are involved.