Thirteen games on the board tonight, and the lowest total of the lot sits in the Pacific Northwest at a flat 7 — all nine books parked on the same number, no daylight between them. On paper it reads like a two-ace lockdown: Dylan Cease and his 36.7% strikeout rate against Luis Castillo, in the marine air of a T-Mobile Park that runs a 0.92 run factor and turns fly balls into outs.

Except only one of those arms is actually pitching to the number. Cease is the real thing — a 3.02 ERA his 3.00 xERA fully backs, a slider hitters whiff on 44.5% of the time, a changeup they miss on two of every three swings. Castillo is the crack in the total. His season xERA is 4.60, hitters are barreling him at a 9.8% clip and squaring him up 43.2% of the time — that's top-of-the-scale hard contact. The 4.93 ERA isn't bad luck hunting him down; it's the contact quality collecting its bill.

He's papered over it for five starts — a 2.33 run that has kept the reputation intact and the total suppressed. So the angle is the collision: does the ballpark keep swallowing the loud contact Castillo surrenders, or does a Blue Jays lineup that only needs to catch a hittable fastball (his four-seamer is getting hit hard 47% of the time) make a 7 look light? The park is holding this total up. Watch whether Castillo can hold up the park.

Christian Scott walks to the mound in Atlanta carrying a 3.20 ERA and a secret. The number is a mirage — his expected ERA, built from the quality of contact he actually allows, is 4.64. That 1.44-run gap is the widest on tonight's board, and it's the kind of distance between results and process that doesn't survive a summer.

The trouble for the total is that the man across from him is running the same con. Grant Holmes owns a tidy 3.96 ERA sitting on top of a 4.76 xERA — a 0.80-run overperformance of his own, propped up by a 44.7% hard-hit rate that says the soft-contact luck is on loan. Two starters, two ERAs that flatter them, one nine-run total taking both at face value.

And then there's the air. It's 97 degrees at first pitch in Cobb County — the kind of thick Georgia evening where hard contact doesn't die on the warning track. Both of these arms have survived on balls staying in the yard; tonight the yard is on the hitters' side. When two pitchers are both living on borrowed outs, the question isn't who regresses — it's whether the ballpark makes them do it on the same night.

Look at Reid Detmers' 3.88 ERA and you'd file him under "fine, nothing special." Look underneath it and the number starts to read like a clerical error. His FIP — what his strikeouts, walks and homers actually earn — is 2.96, nearly a full run better. His xERA is 3.04. The stuff has been ace-grade; the ledger just hasn't paid out.

The culprit is a 64.8% strand rate — abnormally low, and the kind of thing that reverts by force, not by choice. Detmers isn't getting lit up in trouble. He's a 27.6% strikeout arm with a slider hitters miss on a third of their swings and a curveball they whiff on 38% of the time; he's simply watched an unusual share of the runners he allows come around to score. That's noise, and noise fades.

What's shifting is perception catching up to stuff. The 3.88 keeps him quietly priced, but the underlying profile says Boston is facing a better pitcher than the line implies. He's been leaving the evidence on base all season. Eventually the strand rate stops covering — in the good direction.