Post by ericmvan on Sept 17, 2016 11:09:09 GMT -5
Bullpen numbers starting from August 27, the game after Tazawa's last really bad outing. Since Matt Barnes was awful on 8/28 and has been great since, I've given his line both ways (Barnes* is the revision).
tERA- is more accurate than ERA-. It's based on WPA/LI, which is a context-neutral stat that's based just on batters faced and hence adjusts for inherited runners. It's pretty easy to convert it to ERA-.
W/Yr- is WPA per 250 batters faced given average leverage, converted to a number where 100 is league average and designed to scale like ERA-.
Qual just averages those three metrics. Remember that the difference between xFIP- and tERA- is hardness of contact allowed, and that the difference between tERA- and W/Yr- is clutch.
UI is Usage Importance. It's simply the LI of every batter faced, totaled (BFP * pLI). (Barnes* is prorated to adjust for the two fewer games).
Excluding Ramirez and Scott ....
Even though JF hasn't done anything obviously stupid, the correlation of Quality to Usage Importance (even excluding Barnes' first game) remains pretty awful. Swapping Quality numbers to positive, it should be a strong positive correlation, but it's -.51, -.65, -.52 for the three components and -.63 for the composite.
Uehara, Ziegler, and Kelly rank 1-2-3 in Quality, but are the last three in UI.
In JF's defense, this was a period where he was figuring out if Kelly and Uehara were good ... but he hasn't really started to use either one in important situations. He also deserves some credit for Ziegler's great numbers by matching him up well, and that mysterious stretch where he didn't use him when expected might have been an undisclosed slight physical thing -- he certainly used him correctly last night.
And of course he's committed to Kimbrel as his closer, and he remains mediocre. If you exclude Kelly, Uehara, Ziegler, and Kimbrel, the Quality / UI correlation is OK, .40. It's .80 if you also exclude Abad, but now all we're saying is that he trusts Ross and Barnes more than Hembree and Tazawa. Duh.
tERA- is more accurate than ERA-. It's based on WPA/LI, which is a context-neutral stat that's based just on batters faced and hence adjusts for inherited runners. It's pretty easy to convert it to ERA-.
W/Yr- is WPA per 250 batters faced given average leverage, converted to a number where 100 is league average and designed to scale like ERA-.
Qual just averages those three metrics. Remember that the difference between xFIP- and tERA- is hardness of contact allowed, and that the difference between tERA- and W/Yr- is clutch.
UI is Usage Importance. It's simply the LI of every batter faced, totaled (BFP * pLI). (Barnes* is prorated to adjust for the two fewer games).
Excluding Ramirez and Scott ....
Even though JF hasn't done anything obviously stupid, the correlation of Quality to Usage Importance (even excluding Barnes' first game) remains pretty awful. Swapping Quality numbers to positive, it should be a strong positive correlation, but it's -.51, -.65, -.52 for the three components and -.63 for the composite.
Uehara, Ziegler, and Kelly rank 1-2-3 in Quality, but are the last three in UI.
In JF's defense, this was a period where he was figuring out if Kelly and Uehara were good ... but he hasn't really started to use either one in important situations. He also deserves some credit for Ziegler's great numbers by matching him up well, and that mysterious stretch where he didn't use him when expected might have been an undisclosed slight physical thing -- he certainly used him correctly last night.
And of course he's committed to Kimbrel as his closer, and he remains mediocre. If you exclude Kelly, Uehara, Ziegler, and Kimbrel, the Quality / UI correlation is OK, .40. It's .80 if you also exclude Abad, but now all we're saying is that he trusts Ross and Barnes more than Hembree and Tazawa. Duh.
Name BFP xFIP- tERA- W/Yr- Qual UI
Uehara 16 49 15 23 29 6.6
Ziegler 19 45 59 2 35 8.2
Barnes* 15 39 57 47 48 18.1
Kelly 29 52 44 48 48 11.0
Ross 31 87 79 80 82 16.7
Tazawa 19 57 66 158 94 11.6
Hembree 20 91 113 124 109 13.6
Kimbrel 18 105 93 137 112 18.7
Abad 21 171 103 111 128 16.6
Barnes 20 69 133 148 116 27.4
Ramirez 8 55 149 90 98 5.8
Scott 15 95 26 31 51 2.3