clock menu more-arrow no yes mobile

Filed under:

Terry Francona learned the hard way why Batter vs. Pitcher numbers don't matter

Or: Why small sample sizes are a thing.

Bruce Kluckhohn-USA TODAY Sports

When Bryan Shaw entered the game against the Chicago White Sox yesterday in the seventh inning, it did not seem like that bad of a decision. Maybe a little odd, if you believe in bullpen "roles" and the idea that Shaw should be an eighth-inning setup man, but the Cleveland Indians had a one-run lead and Shaw was coming in fresh. Surely he could hold a one-run lead for a single inning, right?

Well, things went bad. Real bad. In a matter of a few batters, Shaw allowed five run (a career-high for any single outing) and effectively lost the game for the Cleveland Indians. He looked off from the very first batter when Austin Jackson singled to left field. But for whatever reason, Francona refused to take Shaw out until it was too late. Zach McAllister was only brought in when the damage was done following a three-run home run from Avisail Garcia.

Why did Terry Francona bring Shaw in for the seventh inning, and why didn't he take him out when it was clear the reliever did not have it that day? For starters, I have no issue with bringing in relievers in unusual situations. I would not have even been against Tito bringing in Cody Allen in that situation. With the team only up by one run, why not bring your best reliever into the game to make sure that lead stays intact for another half inning?

According to Jordan Bastian, Terry Francona said something in his post-game press conference that puts some bad light on the decision:

If that is the sole reason for how a manager chooses what reliever goes into the game, that's a problem.

When those seven batters were done beating Bryan Shaw like a drum, they were now a combined 10-for-41 against shaw, with 13 strikeouts and a home run. That's now a .244 batting average against. Using the logic of Batter vs. Pitcher matchups, it is now no longer such a great decision to use Shaw in this situation, after only one game. This is the primary issue with Batter vs. Pitcher matchup numbers-- especially with this small sample size -- they are not predictive at all.

Sure, if you want to look at a matchup and say "wow, this pitcher sure has dominated this one batter," go for it. But don't expect that to always happen, just because a pitcher had 10 good at-bats against a particular opponent.

Again, the fact that Terry Francona wanted to try Bryan Shaw in the seventh over Zach McAllister does not bother me all that much. If Shaw would have pitched well it would not have mattered and Francona would probably getting all the praise for it. Stuff happens, this is baseball. However, if the only reason he chose Shaw over McAllister was because of flawed Batter vs. Pitcher data, that is bad decision making, plain and simple. I do not question Francona's decision to use Shaw in this situation, I just question how he came to the decision.

From Francona's post-game discussions, it sounds like he had no intention of pulling Shaw throughout the inning. If that decision was based on the idea of "well, the guys coming up next have never been very good about," that is downright terrible decision making. We have to hope that is not the case, or if it was, we have to hope Tito learns quickly.