The Blue Jays’ Bullpen: AA And The Unloved Closer


Conventional wisdom has it that a good starting pitcher is worth far more than a good closer.  By fangraphs fWAR, the most valuable closer so far in 2015 has been Dellin Betances, with a FWAR of 1.7.  By comparison, there are 31 starting pitchers with a fWAR of 1.7 or higher.  This is presumably the basis for Alex Anthopoulos’ comments about he was prioritizing the starting rotation over the bullpen.

But sometimes fWAR does not tell the whole story.

More from Toronto Blue Jays News

When you look at WPA – win probability added – so far in 2015,

five of the top ten pitchers are relievers

.  This makes sense, as WPA is highly influenced by leverage index.  So a reliever pitching in the 9th inning in a tie game gets far higher “credit” for that inning than a starter pitching in the first inning. It is for that reason that fangraphs only gives relievers partial credit for the leverage index when they are pitching, which reduces their fWAR.

But it is equally valid that a closer can lose a game for his team more easily than a starter can.  To see the effect, consider the following thought experiment.

Suppose last year the Jays had a closer robot.  Let’s call him “Greg”.  Suppose Greg could only be used to finish games,  in a save situation.  And suppose that every time Greg pitched, his team had a 46/48 (96%) chance of winning.

In 2014, the Jays had a total of 63 save opportunities, of which they converted 45.  Of the 18 blown saves, 10 were before the 9th inning.  Say that Greg was not involved in any of those 10, but that he pitched in the remaining six.  So that would give Greg 53 save opportunities.

Of those 53 GSOs (Greg save opportunities), the 2014 Jays successfully converted 45.  Of the eight ninth-inning blown saves, the team came back to win two.  So the team record in GSOs was 47-6.

Greg, in our example, would have converted 96% of the GSOs, or 51 saves.  The team would accordingly have had a record of 51-2 in those games.

So the difference in games won would have been four.  Other things equal, that would have raised the Jays’ 2014 record from 83-79 to 87-75.

Now suppose we relax our assumptions to assume that Greg pitched, say, 12 additional innings (over and above the 53 one-inning GSOs).  These would be in high-leverage situations other than save opportunities.  And suppose that those 11-12 holds translated into one additional win, raising the Jays’ 2014 record to 88-74.

88-74 would have put the Jays into a tie with Oakland for the second wild card spot.

The bottom line:  the analysis above is clearly hypothetical.  It is by no means certain that Greg the robot would perform at the same level in Toronto that he would elsewhere.  But, just as flags fly forever, blown save losses are gone forever.  And in 2015, a handful of wins could make the difference between baseballs and golf balls in October.

Next: Hoffman for Cueto and Chapman: Would The Blue Jays Do It?