Mason Miller's ERA Spike Traces to Two Homers, Not Stuff Decline

Padres closer Mason Miller's ERA has risen recently, but a dig into his last 20 outings shows the damage is concentrated in two home runs while strikeout rate and WHIP remain elite.

Published: August 31, 2026 · Figures as of August 31, 2026

Asked: Should the Padres be worried about Mason Miller?

Miller's ERA blip traces to two home runs, not eroding stuff

Mason Miller's earned run average has risen from 0.86 to 2.30 over his last 15 outings (July 24–Aug 30), but that whole increase is two home runs allowed in his last two appearances — everything else in the underlying process (command, strikeout rate) is still elite. His season ERA sits at 0.96 with 32 saves, and nothing here points to a broader decline.

What actually happened recently

Miller's last three outings, all against playoff-relevant opponents, included two rough ones sandwiching a clean one:

• Aug 30 at Tampa Bay: 1 IP, 2 ER, 1 HR allowed, blown save (Padres lost 4-5)
• Aug 25 vs Pittsburgh: 1 IP, 1 ER, 1 HR allowed, took the loss (Padres lost 0-1)
• Aug 24 vs Pittsburgh: 1 IP, 0 ER, 1 hit, 1 walk, 2 K, no decision

Those two homers account for 3 of the 4 earned runs he's allowed across his last 20 appearances — in the other 18 outings over that stretch he allowed one earned run total. The chart below shows how concentrated the damage is:

Decomposing the change

• WHIP actually improved: 0.77 over the last 15 outings vs. 0.86 before that (season: 0.84) — he isn't walking or being hit more often, he's well below the league mean of 1.27 for pitchers over the same stretch.
• Strikeout rate cooled slightly but is still dominant: 14.94 K/9 in the window vs. 16.85 before and a season mark of 16.33 — still more than 5 strikeouts per nine above the league's 9.04 average over those same dates (2.05 standard deviations above the field).
• Pitch counts per outing ticked up modestly (18.0 vs. a season rate of 17.11), consistent with a couple of longer, harder-fought at-bats rather than any workload change.

So the diagnostic picture is a pitcher who is missing bats and limiting baserunners at essentially the same rate as before, but who has served up two home runs in a six-appearance span where he'd allowed zero all year before that.

What the sustainability read can and can't tell us

DataBaller's process-vs-results model for pitchers needs at least 80 balls in play in a trailing window to score whether a hot or cold stretch is "real" — it works off how far a pitcher's batting-average-on-balls-in-play sits from his own established level. Miller doesn't have a stored read this season, and the reason is structural rather than a data gap: as an elite strikeout reliever (roughly 73 trailing-year innings with the bulk of his outs coming via strikeout rather than balls in play), he simply doesn't accumulate enough contact events to clear that floor. That cuts both ways — it means we can't lean on the model to say the two homers are "owed back," but it's also why a stretch like this has limited room to compound: there's very little contact for luck to run against him on in the first place.

Team context

San Diego is on a three-game losing streak, having been swept by Tampa Bay to close out August, and sits 72-65, third in the NL West, 10 games back in that race. Miller's blown save was part of that sweep, which is likely what's driving the "should we worry" question, but the losing streak as a whole is broader than one reliever — the same series saw the offense manage just 4-6 runs.

Confidence: Medium. The command indicators (WHIP, walk rate, strikeout rate) show no decline, which argues against a mechanical problem. But the sample is small — two home runs in six outings — and without the sustainability model's regression read on him, there isn't a stored basis for calling this fully random variance either. The specific thing worth watching over his next several outings is whether the home-run rate continues or reverts, since that's the one indicator that has actually moved.