There has been a lot of talk about how the pen has been used lately. Is Farrell using it appropriately or is he giving the highest leverage outings to the worst pitchers? Lets see.

According to FanGraphs only 10% of appearances have a leverage index of 2 or greater (ie: game on the line) while 60% of all appearances occur with leverage below 1 (ie: time for Mathis to pitch). For an example, in a 3-2 game bottom of the 9th the leverage is 3.6 - a tough save situation when you want your best in to ensure the win. When there is just 1 more out to get in a 3-2 game though the level drops to 1.9 (via Hardball Times).

So, lets split things up a bit and see who did what. + = times he increased odds of winning, - = times he dropped the odds of winning (based on WPA). Stats via Baseball-Reference. Listed by order of ERA+ before Friday the 15th's game. 5+ appearances in relief to be listed.

• 3.0 and greater (killer situation)
  • Oliver: 2, +1 -1
  • Janssen: 2, +2 -0
  • Perez: 2, +1 -1
  • Frasor: 3, +2 -1
  • Villanueva: 0
  • Crawford: 0
  • Cordero: 3, +1 -2
  • Santos: 1, +1 -0
  • Others: 0
• 2.0 - 2.99 (high pressure situation)
  • Oliver: 1, +0 -1
  • Janssen: 2, +1 -1
  • Perez: 3, +2 -1
  • Frasor: 3, +2 -1
  • Villanueva: 0
  • Crawford: 1, +1 -0
  • Cordero: 7, +4 -3
  • Santos: 3, +1 -2
  • Others: 2, +0 -2 (Beck, Igarashi)
• 1.5 - 1.99 (difficult)
  • Oliver: 4, +4 -0
  • Janssen: 3, +2 -1
  • Perez: 3, +2 -1
  • Frasor: 3, +3 -0
  • Villanueva: 3, +2 -1
  • Crawford: 0
  • Cordero: 3, +2 -1
  • Santos: 0
  • Others: 0
• 1.0 - 1.49 (above average)
  • Oliver: 6, +5 -1
  • Janssen: 1, +0 -1
  • Perez: 1, +1 -0
  • Frasor: 5, +5 -0
  • Villanueva: 2, +2 -0
  • Crawford: 0
  • Cordero: 4, +4 -0
  • Santos: 0
  • Others: 0
• 0.5 - 0.99 (ho-hum)
  • Oliver: 5, +4 -1
  • Janssen: 7, +6 -1
  • Perez: 4, +3 -1
  • Frasor: 7, +5 -2
  • Villanueva: 4, +4 -0
  • Crawford: 3, +1 -2
  • Cordero: 1, +3 -1
  • Santos: 1, +1 -0
  • Others: 2, +0 -2
• Below 0.5 (Mathis time)
  • Oliver: 7, +6 -1
  • Janssen: 8, +7 -1
  • Perez: 14, +12 -2
  • Frasor: 6, +3 -3
  • Villanueva: 11, +7 -4 (8 in sub 0.1 situations)
  • Crawford: 5, +5 -0
  • Cordero: 9, +6 -3
  • Santos: 1, +1 -0
  • Others: 9, +7 -2

So in summary...

• 3.0 and greater (killer situation): +8 -5, most appearances: Cordero/Frasor
• 2.0 - 2.99 (high pressure situation): +11 -11, most appearances: Cordero (double anyone else)
• 1.5 - 1.99 (difficult): +15 -4, most appearances: Oliver (by 1 over 5 others)
• 1.0 - 1.49 (above average): +17 -2, most appearances: Oliver
• 0.5 - 0.99 (ho-hum): +27 -10, most appearances: Janssen/Frasor
• Below 0.5 (Mathis time): +54 -16, most appearances: Perez (by 3)

FYI: the lower the leverage the harder it is to have a negative score as high leverage often only requires one or two hits allowed to switch a win into a loss while in a 10-0 game all that matters is getting outs even if you allow 2 or 3 runs in an inning.

So for the high pressure plus area (2 or greater) Cordero leads in appearances by a mile over anyone else. How did he do? 5 times he helped, 5 times he hurt the Jays odds of winning. All others combined were +14 -11, slightly better. Perez, even though he has been effective when the pressure is on (+5 -3) still is used mainly for garbage time (14 of 27 appearances in extremely low pressure situations).

It seems clear the manager is addicted to Cordero (probably due to his large number of saves in the past) but is uncomfortable with Perez. An odd situation as Perez walks a bit more (0.4 per 9), but strikes out far more (2.1 more per 9 IP) and gives up home runs a lot less (1.2 HR/9 is the spread between them) and has fewer hits against (3.7 fewer per 9). Based only on this year it isn't even close as to who to use yet Cordero keeps getting chances. AA might need to release Cordero just to stop Farrell from using him.