As we all agree on the idea that relief pitching stats are volatile in part because of small sample sizes, why focus only on 2018 regular season stats for Workman? Suppose you combine Workman's Regular season and playoff stats, plus his MLE for the innings he pitched in AAA. How do they look? Those 3G, 1 IP combined, 7 H, 3 BB, 5 R, 2HR skews things....or was that the regression to the mean?
Yes, Workman certainly had a rough postseason -- but if you include it in his 2018 MLB line, it's
still better than Plutko's. And of course we don't know how Plutko would have done against playoff competition, because the Indians didn't deem him worthy of the opportunity.
Computing MLEs seems like too big a lift for a question like this, but projections might be worth looking at. I took the average of three projection systems -- Steamer, ZiPS, and THE BAT -- for six rate stats for each pitcher. Here's what that produced:
ERA: Workman 4.52, Plutko 4.62
WHIP: Workman 1.38, Plutko 1.27
FIP: Workman 4.44, Plutko 4.70
K/9: Workman 8.55, Plutko 7.99
BB/9: Workman 3.48, Plutko 2.87
HR/9: Workman 1.26, Plutko 1.58
If anything, Workman comes out looking better here, but the projection systems see both as mediocre, which seems plausible enough. (FWIW, ZiPS liked Workman best and Plutko worst of the three systems, while THE BAT had that reversed, and Steamer saw them as nearly equivalent, with a slight edge to Workman.)