Cy Young Heat Check← Japanese
Ranking updated: 2026-05-26 01:37 JST
An independent rating model tracking the 2026 NL Cy Young race (as of May 26, 2026).
Shohei Daily Lab original model · 85% winner hit rate (2016–25, NL+AL) · as of May 26, 2026 · not affiliated with MLB
Recent Starts
Current Ranking (Independent Score)
Shohei Daily Lab Original Model. 2026 NL Cy Young rankings produced by our independent model. Backtested on the 20 NL+AL Cy Young races from 2016–2025: Winner Hit Rate 85% / Top-3 set match 93%. Updated each start by Ohtani or his main rivals. Not affiliated with MLB — an experimental lab project.
| # | Pitcher | Score | W-L | ERA | WHIP | K | IP | rWAR |
|---|---|---|---|---|---|---|---|---|
| 1— | C. SANCHEZ · PHI | 84.8 | 5-2 | 1.62 | 1.15 | 86 | 72.1 | 3.63 |
| 2— | J. MISIOROWSKI · MIL | 79.8 | 4-2 | 1.89 | 0.88 | 88 | 57.0 | 2.16 |
| 3— | C. SALE · ATL | 76.4 | 7-3 | 1.89 | 0.87 | 72 | 62.0 | 2.16 |
| 4— | C. BURNS · CIN | 75.9 | 6-1 | 1.83 | 0.95 | 64 | 59.0 | 2.87 |
| 5— | SHOHEI OHTANI · LAD | 73.8 | 4-2 | 0.73 | 0.84 | 54 | 49.0 | 2.22 |
| 6— | K. HARRISON · MIL | 69.2 | 5-1 | 1.77 | 1.07 | 59 | 45.2 | 2.02 |
| 7— | B. ELDER · ATL | 66.9 | 4-2 | 1.97 | 0.99 | 60 | 68.2 | 2.15 |
| 8— | Z. WHEELER · PHI | 63.4 | 4-0 | 1.67 | 0.82 | 36 | 37.2 | 2.02 |
| 9— | M. MEYER · MIA | 61.9 | 5-0 | 2.52 | 1.05 | 68 | 60.2 | 2.20 |
| 10▲1 | P. SKENES · PIT | 55.1 | 6-4 | 3.00 | 0.82 | 65 | 60.0 | 1.37 |
Inside the Model — Weights (out of 100)
An independent 100-point rating model. Weights were derived from how BBWAA voters actually behaved over the last decade in the AL and NL — strikeouts (the "dominance" signal) get extra weight.
| Metric | Weight | Why it’s weighted this way |
|---|---|---|
| ERA / Run prevention | 30 | The metric voters track most. 8 of 10 recent winners led their league in ERA. |
| rWAR / Cumulative value | 22 | Baseball-Reference rWAR (2-decimal). Pitchers with elite WAR consistently rise in voting. |
| Strikeouts / Dominance | 20 | Voters reward strikeouts as the marker of dominance more than is widely assumed. |
| Innings / Durability | 13 | A volume tiebreaker between pitchers with otherwise comparable rate stats (ERA/WHIP). |
| WHIP / Baserunners | 7 | Correlates with ERA, so weighted lightly to avoid double counting. |
| Wins | 6 | Limited independent predictive value — kept small as a residual "narrative" point. |
| Contender team bonus | 2 | Minor — winners have come from last-place teams. |
| Total | 100 | |
Scoring is relative within each season's actual top-5 vote pool. ERA and WHIP are scored against the league's average ERA (ERA+ style). WAR is the real rWAR (2-decimal) from Baseball-Reference.
How Reliable Is the Model?
Backtested on the 20 NL+AL Cy Young races from 2016–2025: Winner Hit Rate 85% · Top-3 set match 93% · Exact 1-2-3 order 45% (every actual winner was a starter; excluding the 2 years where a reliever finished in the top-5 — AL2016 Britton 4th, AL2024 Clase 3rd: hit 89% / top-3 93%).
- Winner Hit Rate 85% — the model's #1 matched the actual Cy Young winner in 17 of 20 races.
- Top-3 set match 93% — the average overlap between the model's top 3 and the voters' top 3.
- Exact 1-2-3 order 45% — years where the model got the precise winner→2nd→3rd ordering right. Close races where the voters themselves split make exact order naturally unstable.
Three Cases Where the Model Diverged from Voters
The 3 misses out of 20. All three were tight, contested races at the time — the kind of year even human voters find hard to call. A miss doesn't mean the model is wrong; in some cases it sides with what later analysis came to favor.
2016 AL
| Player | W-L | ERA | IP | K | Pts | 1st-pl. | Finish | |
|---|---|---|---|---|---|---|---|---|
| Actual #1 | Porcello | 22-4 | 3.15 | 223.0 | 189 | 137 | 8/30 | 1 |
| Model #1 | Verlander | 16-9 | 3.04 | 227.2 | 254 | 132 | 14/30 | 2 |
The famous "Porcello controversy." Verlander had far more first-place votes (14 vs 8), but Porcello won by 5 points on the strength of 2nd–5th place votes. The model favors the statistically dominant Verlander — the side history has tended to side with.
2021 NL
| Player | W-L | ERA | IP | K | Pts | 1st-pl. | Finish | |
|---|---|---|---|---|---|---|---|---|
| Actual #1 | Burnes | 11-5 | 2.43 | 167.0 | 234 | 151 | 12/30 | 1 |
| Model #1 | Scherzer | 15-4 | 2.46 | 179.1 | 236 | 113 | 6/30 | 3 |
The tightest NL race of the decade — Burnes 151 vs runner-up Wheeler 141 (just 10 points). Burnes had the rate-stat edge in fewer starts; Scherzer had the wins and a 0.864 WHIP. The voters themselves were genuinely split.
2019 AL
| Player | W-L | ERA | IP | K | Pts | 1st-pl. | Finish | |
|---|---|---|---|---|---|---|---|---|
| Actual #1 | Verlander | 21-6 | 2.58 | 223.0 | 300 | 171 | 17/30 | 1 |
| Model #1 | Cole | 20-5 | 2.50 | 212.1 | 326 | 159 | 13/30 | 2 |
Astros teammate showdown decided by a thin margin (1st-place votes 17-13, total margin 12 points). Cole led in ERA, strikeouts, and WHIP — the model picks Cole. The voters were split right to the end.
For context: in calmer years (deGrom’s back-to-back NL wins, the unanimous wins by Alcántara / Skubal / Cole, Snell, Bieber, and recent ones like Sale’s 2024 Triple Crown or Scherzer 2016) the model agrees with the voters completely.
How This Ranking Works
Two automated steps, every day:
- Pool selection — the top 20 NL starters by FanGraphs WAR (Ohtani is always pinned, even below the qualified-IP line).
- Relative scoring — within those 20, our independent model (real rWAR, ERA, strikeouts/K9, WHIP, innings, wins, scaled to league-average ERA) ranks them.
This is not a future prediction — it is "how the race would look if the vote were held today." Recalculated each start by Ohtani or his main rivals. Relief pitchers are outside the model's scope (low IP × tiny ERA distorts the score) — a starters' race only.