Here’s another angle…
What is the reason to keep Jorge Lopez over Pagan?
So far this season, Pagan has a positive WPA in 17 games and a negative in 7, a “winning percentage” of .708. Lopez has had a positive in 19 and a negative in 9, a “winning percentage” of .679. Advantage Pagan.
In Pagan’s negative WPA games, the values were .680, .611, .299, .041, .024, .023 and .002. In Lopez’s, the values were .396, .331, .276, .181, .156, .143, .070, .050, .035.
Looking a little deeper, Pagan had two disasters — in LA and yesterday. The .299 was when he came in for an injured Maeda, didn’t pitch well in his first inning and then took one for the team and kept pitching. His other negatives include a game where he came in down two in the seventh and gave up one run; another where he came in in the 8th, down three, and gave up a run while pitching the final 1.1; the game where he came in down two, hit a batter and was pulled because of injury; and a game where he came in on the road, down six with a runner on base, and allowed that runner to score while pitching the final two innings. In those four, he had no real effect on the outcome, but did help the team by covering an inning (or more) that someone else would have needed to cover.
Given that he came in already behind, I think you can only say that Pagan played a significant role in cost the team two wins. Perhaps a third (the Maeda game), but the win probability was already down at .35 when he came in, so it was already a long shot.
Contrast to Lopez, who has had a greater effect on the number of losses. In three of the negatives, he has a blown save and in another he got the loss. He has eight games where his negative is bigger than Pagan’s fourth-worst. He has six games where he either entered with a lead and left either tied or down or entered tied and left down. That’s somewhere from 4-6 losses at his hand.
On the other side, Pagan’s positive games include a .309, .120 and .105. His remaining 14 are between .056 and .001, largely reflective of doing his job, most often in the 7th. The most common scenario is his coming in in the 7th, down one or two, and pitching a scoreless inning. In two of those, he vultured himself a win. The .309 was his scoreless 11th inning in the 12-inning win over Chicago, when he also got the win.
Lopez’s positive games include a .138, .135 and .128. None are as high as the Pagan extra inning game, but his remaining 16 are more evenly spread between .093 and .006, largely the result of pitching an inning later than Pagan.
I’m not sure how you compare the positive-side games, but I’d suggest that Pagan played a major role in one win, a moderately major role in a couple others, and a positive effect in another 14 games. Lopez hasn’t had the major effect on any games, but has had a modest effect on three or so and and a positive effect in another 16 games.
If you look at the two side-by-side, I’d argue that the team is 1-2 in games where Pagan has had a major effect on the outcome, but somewhere from 0-4 to 0-6 in games where Lopez has had a major effect.
If the argument that Pagan should be gone is that he’s blown two games, I think you should be saying the same about Lopez.
But Lopez has had past success, you might say.
Not that much. He’s been truly atrocious for most of his career. Any sense of success is driven by the first part of last season when he was with Baltimore. From 2018 to 2021, his ERAs were 6.35, 6.33, 6.69 and 6.07. During that time, Pagan’s were 4.35, 2.31, 4.50 and 4.83. I know that ERA isn’t a great measurement for relievers, but that’s a stark contrast in Pagan’s favor. During the two months they were teammates last year, they were virtually identical in WPA. Pagan was -.225 in 21 games and Lopez was -.273 in 23.