The perils of managing by formula

I long railed on Power Line against the formulaic use of relief pitchers that was followed pretty much universally when we started our blog in 2002, and for many years thereafter. The formula was that the team’s “closer” would be used only in the ninth inning of games in which his team had a lead of one to three runs. The closer would always start the ninth inning; he would never be asked to enter the game with runners on base. A “set up” man would handle the eighth inning under similar terms.

This rigid scheme defied common sense. However, it wasn’t abandoned until “analytics” proved that it also defied the numbers.

In this year’s MLB playoffs, and especially the World Series, we have seen the wholesale abandonment of the traditional formula for using relievers. Putative closers enter games in the early middle innings. They often enter with runners on base. The notion — myth, I would say — that relievers need highly defined roles to prosper has been overturned.

This doesn’t mean, though, that managers aren’t using formulas when it comes to relief pitchers. They may be using new, more complicated formulas (or algorithms) that I’m not smart enough to discern.

One thing I have discerned is a bias against having a starting pitcher face the opponent’s lineup three times and against a reliever facing it twice. The idea isn’t that the pitcher will be tired the third (or second time) through the lineup. Rather, it’s that the pitcher will be out of “tricks” and that the batters will have adjusted to his approach by then.

This view is plausible in terms of common sense and is backed by numbers. But that doesn’t mean it should rise to the level of religion.

Unfortunately, it did last night in Game Six of the World Series — to the detriment of the Tampa Bay Rays.

The Rays were down three games to two, so this was a must-win contest for them. Fortunately, one of their ace pitchers, Blake Snell, was pitching brilliantly. Through five innings, he had allowed no runs and only one hit. Nine of the 15 Dodgers he faced had struck out. He hadn’t allowed a walk. His pitch count was under 70. The Rays held a 1-0 lead.

Snell retired the first batter he faced in the sixth inning. The second batter singled to center field, thus becoming the second Dodger baserunner of the game.

At this point, to the amazement of most (and certainly to Snell’s), Tampa Bay manager Kevin Cash pulled his ace starter and brought in star reliever Nate Anderson.

Why? Because Snell was about to begin his third trip through the opposing lineup.

But Snell had twice struck out Mookie Betts, who was due to bat. And Betts, as good as he is, struggled against left-handed pitchers this season. Snell is left-handed. Anderson is a righty.

Furthermore, Anderson, although great during the regular season, has not pitched well in the playoffs. If I heard the announcer correctly, he had given up runs in six consecutive outings. In all, he had allowed eight runs in 14.1 post-season innings this year before last night’s game.

In sum, pulling Snell for Anderson with Betts coming to the plate was malpractice.

Sometimes managers get away with awful moves, but Cash did not. Betts promptly doubled. A wild pitch and a fielder’s choice later, the Dodgers had a 2-1 lead.

Anderson did manage to retire Justin Turner, albeit on a fly ball that looked, off the bat, like it might be a home run. After that, having faced the mandatory minimum of three batters, Anderson was pulled.

Cash’s malpractice very likely cost the Rays the lead, which the Dodgers never relinquished. It probably did not cost the Rays, who scored only one run, the game. As well as Snell was pitching, it’s doubtful that he and the relievers who presumably would have come on late in the game could have shut Los Angeles out.

Still, the lesson from last night is clear. Managing purely by formula is a bad idea, even if, as a general matter, a formula has the imprimatur of “analytics.”