The model's AL #1 by a mile, and it holds up: 8-5, 2.05 ERA over 105 innings, 122 strikeouts, sub-1.00 WHIP. His 2.52 FIP says only about a half-run of luck is in that ERA, and a 24.6% strikeout-minus-walk rate is genuine front-line stuff. The model does expect some second-half cooling, but the projected full-season line — 14-8, 2.74 ERA, 219 strikeouts (80% of sims land between 208 and 229 K) over 187 innings — is the second-best projected ERA in all of baseball. That's not regression to a mirage; that's regression to an award season. It matters where it lands, too: the Yankees (51-42) hold the top wild-card spot (97.9% playoff) but trail the Rays by four in the East, so every Schlittler start is leverage in a live race. He earned his first All-Star nod for it.
The 2026 Cy Young Races at the Break — Reading the Real Leaders
Article 1 of 2 — the companion piece covers the MVP races. Stat lines are through games of July 9, 2026; model projections and Cy Young odds are from the July 7 simulator run (a two-day gap that doesn't change any ranking). All-Star designations are confirmed against the official rosters.
The 96th All-Star Game lands in Philadelphia on July 14, which makes right now the first real referendum on the 2026 Cy Young races. One note on how we rank them: BBMI grades Cy Young on earned runs saved — each pitcher's projected ERA measured against the league average of 4.22, scaled by his projected innings — so the case rewards both run prevention and the volume to deliver it. We also lean on our season simulator, which replays the remaining schedule 500 times and projects every pitcher's full-season line. That second piece is what separates the pitchers whose shiny first-half ERAs are real from the ones the model expects to regress after the break.
American League
This is not a coin flip. The model has one clear favorite — Cam Schlittler at 61% — with everyone else fighting over the scraps.
Second in the model at 12%, because he has the cleanest signal in the field: a 3.26 ERA that his 3.32 FIP nearly matches (+0.06 gap), 1.4 walks per nine, and a projected full-season 3.08 ERA with a 1.00 WHIP — the model simply doesn't see anything to give back. What caps the case isn't the rate, it's the volume: the sims project just 169 innings as Tampa Bay manages his workload, and you don't out-poll a 187-inning Schlittler from there. He anchors a 54-37 Rays club that's a lock for October (100% playoff, leading the AL East). A well-earned second selection.
Quietly the model's AL #3 at 7% despite the lost starts: a 3.06 ERA across 12 outings, and the sims project a 2.99 full-season ERA with a field-best 0.99 WHIP. The problem is the same one that kept him off the roster, having missed time — only ~150 projected innings. He's the arm nobody wants to see in a healthy second half, even as Detroit fades.
A 2.56 ERA and a 1.10 WHIP over 16 starts mark a strong return to form, and the 3.06 FIP says it's mostly real (+0.50 gap). The model gives him 4%: the strikeout rate (8.4 K/9) is good, not dominant, the projection settles at a 3.19 ERA over ~150 innings, and his own rotation-mate Rasmussen out-projects him on the same contender. He's a big reason Tampa Bay is running the East — just second in line for the trophy on his own staff. The All-Star committee read it the same way: the Rays' rotation nod went to Rasmussen, and McClanahan stayed home.
The record dazzles — 9-1, 2.69 ERA — but this is the field's loudest AL regression warning. Gray's FIP is 3.77, a full run above his ERA (+1.08), his strikeout-minus-walk rate is a middling 15%, and he's allowed 10 homers. The verdict is blunt: he doesn't crack the model's top ten in the league. Translation: the 9-1 is not coming with him. It's also happening on a 43-48 Boston team already 11 back and on the playoff bubble (23.9%), so there's little award narrative to prop it up. He surfaced in the snub conversation, but the underlying profile says the All-Star committee read it right.
The most fascinating name in the AL: a 2.76 ERA that actually understates the arm. Cease's 2.27 FIP sits nearly a half-run below his ERA (−0.49), he leads the league's contenders with 139 strikeouts and a 13.7 K/9, and he's been dinged by a bloated .316 BABIP. The sims project 262 strikeouts — the second-most in all of baseball — but the walks (4.2 per nine) push the projected ERA to 3.54, and the model holds him at 3%. The catch is also the team: Toronto is 44-49 and a long shot (17.2%), so even a monster second half may be a tree falling in an empty forest for the voters. The players, at least, heard it — they voted him in for the first All-Star selection of his career.
Also live in the AL picture: Joe Ryan (Twins — All-Star, 4%), whom the sims project for 211 strikeouts even on a fading Minnesota club; and Jacob deGrom (Rangers), pitching to a 3.33 ERA with a sharper 2.95 FIP and 118 K on a Texas team clinging to the wild-card race — though the roster committee passed, making reliever Jacob Latz the club's lone representative, and the model's top ten leaves deGrom off too, betting the innings don't hold up.
National League
The best story of the first half is also, now, the model's pick. The NL race runs through Milwaukee.
The NL ERA leader and the model's favorite at 47%: 10-4, 1.62 ERA, 167 strikeouts, 13.5 K/9, and a 0.76 WHIP over 111 innings. Crucially, this is earned: a 2.11 FIP, only a +0.49 gap, and a field-best 33.2% strikeout-minus-walk rate. Even with the regression the sims build in, the projected full-season line leads baseball in both headline categories: a 2.57 ERA and 281 strikeouts, at 15-7 over 185 innings with a 0.97 WHIP. When the award is graded on earned runs, dominance of this order beats volume — and the model now says so out loud. His reward for the first half: a player-voted trip to Philadelphia, his second selection, on a 58-34 Milwaukee club that's a 101-win lock — though with his final first-half start falling on the Sunday before the game, he'll cede his active roster spot and be honored from the rail.
The model's NL #2 at 19%, and a case study in trusting the projection over the peripherals: his 2.82 ERA sits under a skeptical 3.64 FIP, yet the sims barely blink — a projected 2.84 full-season ERA and a 2.94 RA9 that is the lowest in the league, over a 16-7, 187-inning season on the best team in baseball. The model treats his run prevention as sustainable, full stop. The Dodgers (61-33, 100% playoff, 28% to win it all) are the class of the sport, and Yamamoto is their front-line arm for October.
Third in the model at 16%, and still the field's volume king: the sims give him the only 200-inning projection in either league (204), with 231 strikeouts and a 16-7 record. The rate case is real too — his 2.83 ERA comes with a 2.88 FIP that matches it almost exactly (+0.05) despite a .327 BABIP, meaning he's been unlucky on balls in play. But a projected 2.95 ERA can't out-poll Misiorowski's 2.57 on an earned-run basis, and innings alone no longer carry the argument. He's doing it for a 51-42 Phillies contender in a preview of his own home ballpark hosting the Midsummer Classic.
On the mound he's been otherworldly in a limited role: 1.73 ERA across 67.7 innings, the second-lowest mark in the league. The innings are still the whole story — the model gives him 4% because it projects only ~152 frames (12-5, 2.84 ERA, tied with Yamamoto for third on the projected-ERA board), and a Cy Young is an innings award as much as a rate award. Note the twist for a pitching column: Ohtani was voted into the All-Star Game as a hitter, not a pitcher — a reminder that his arm is almost a bonus on top of an MVP bat.
Still doing it at the front of Atlanta's rotation: 9-6, 2.27 ERA, 112 strikeouts, a 22.4% strikeout-minus-walk rate. The model tempers the enthusiasm (6%) — a +0.63 FIP gap and a .310 BABIP that could cut either way — but projects a strong 15-9, 211-strikeout, 3.01-ERA finish. The Braves (54-38) lead the East and sit at 96.6% to reach October, so Sale's second half is postseason-rotation setup — and the trip to Philadelphia is the 10th All-Star selection of his career.
The record screams ace — 11-1, 2.50 ERA — and he was voted in by the players. But the model is unmoved (0% across 500 sims): his 3.52 FIP is a full run over the ERA (+1.02), and the projection lands at a 3.66 ERA even while the wins keep coming (14-6). It's also stranded on a 42-49 Cincinnati team that's out of the race (0.3% playoff), the kind of record voters tend to discount. A deserving All-Star on peripheral talent; a tougher sell as the season's second half tests the ERA.
Two more worth flagging. Zack Wheeler (Phillies) has an 8-1, 2.48 ERA first half and was left off the roster entirely — a snub he called "kind of BS" — and here the model partly takes his side: despite a +0.80 FIP gap, the sims put him top-five in the league at 4%, projecting a 14-5, 3.02-ERA season with a 1.01 WHIP. The committee may have read the peripherals; the model read the pitcher. Paul Skenes (Pirates — All-Star, player vote) is the field's most unlucky arm — a 3.49 ERA sitting on a 2.93 FIP — but the sims give him effectively 0%, projecting a 12-11, 3.52-ERA season, because Pittsburgh (47-47, 8% playoff) gives him nothing to pitch for and the model doesn't see the rate separating either. Eduardo Rodriguez (Diamondbacks — All-Star, a league-office pick and the first selection of his career) rounds out the cautionary tales: a 2.10 ERA built on a 4.17 FIP (+2.07, the loudest regression siren in the field) and a career-low 7.2% strikeout-minus-walk rate. The committee rewarded the ERA; the model, which reads the peripherals, leaves him nowhere on its board.
The bottom line
At this break, the ERA leaderboard and the model mostly agree — and that's the story. Misiorowski leads the NL on both; Schlittler is a 61% favorite in the AL because his 2.05 first half is backed by the peripherals and the innings. The model's skepticism is reserved for two kinds of pretenders: the ERAs without peripherals (Gray, Rodriguez, Burns), which it expects the second half to correct, and the rates without innings (Ohtani, Skubal, McClanahan), which can't accumulate enough earned runs saved to win. Watch the arms with the clean FIPs and the full workloads. They're the ones still standing in September.
Methodology: ERA and stat lines recomputed from official box-score data through July 9. Cy Young probabilities and projections are BBMI season-simulator outputs (as of July 7): 500 simulated seasons of the remaining schedule, with each projected line the median across runs and ranges the 10th–90th percentile. Cy Young ranks on earned runs saved versus a league ERA of 4.22, on a projected-ERA basis (RA9 shown for contrast where noted). "FIP gap" = FIP minus ERA; a large positive gap flags an ERA likely to rise. Not betting advice.