Pitching by the Numbers: Predicting saves

After tackling wins prior to taking a vacation break last week , let me explain my season-long avoidance of the save category.

Here’s the key takaway/complaint: through Sunday, the 59-win Braves have generated 15 more saves than the 67-win A’s. That’s why for all his WHIP and K/9 brilliance, Sean Doolittle owners are feeling a little ripped off.

Since it’s human nature to think we can predict anything, we tend to seek in drafts and in in-season trades the closers on the teams that are most successful/likely to continue winning.

But clearly win totals don’t correlate that well to save totals.  It’s not just the A’s who are laggards. The 47-win Astros have only 18 saves. That’s 12 less than the Diamondbacks, who have won just two more games.

I’ve charted up team wins per save. The league average is 1.90. But the range is 1.49 to 2.79.

A complaint can be that I didn’t include blown saves. But that’s a deceptive stat because non-closers can only blow them and never accrue them.

Some believe that you can predict saves by focusing on the teams who rank better in pitching than in hitting, regardless of win totals. Let’s see how that works by looking at runs scored and runs allowed.

The A’s have the greatest differential, obviously not a recipe for generating saves. But was that projectable? Did people think the A’s run differential would be so good? I did not.

The teams with the narrowest gaps, again through Sunday, are the Rays, Royals, Cardinals, Mets, Braves and Pirates. The Mets and Rays are below average in generating saves, or save efficiency relative to wins, if you prefer. But the other teams are well above.

That’s a pretty good ratio, with 66 percent being better than expected and the Cardinals and Braves are great at it. But again, can we have projected that these teams would have such a narrow gap between runs scored and allowed?

And of course, teams can have a narrow run differential by trading blowouts on the positive side for getting blown out (though that’s unlikely, I stipulate).

What about just the good pitching teams who have bad offenses? That would be the Padres, Cardinals, Braves, Reds, Giants and Mariners. I don’t know what to make of four of these five teams being NL. Now we have all the teams being better than average in generating saves (though some just barely).

That seems like the model. But it’s only five teams. And could we have guessed them going into the season? I would have figured the Cardinals and Reds would be average-to-above-average in scoring runs, but maybe that’s just me.

But if you do want to project save efficiency, meaning more saves per win than the league average, then, yes, pick the low-scoring teams with good pitching — if you can.