GOVERNOR MODEL · v0.0.1 Experimental

Control of the governorships is a toss-up.

Updated September 23, 2026 · all 36 governor forecasts refreshed through September 22, 2026 · Texas Marist poll fielded September 17–20

Governor control simulation unavailable for this release. Race forecasts remain available; no aggregate control probability is published.

2026 GOVERNORS

Current governor estimates

Projected margin from the experimental model. Select a state in the map for race details.

Model status
Experimental

Safe DLikely DLean DTilt DToss-upTilt RLean RLikely RSafe RIndependentNo 2026 race

COYNTOSS governor projection. The bar groups seats by current forecast rating; non-up seats are shown at the ends.

CLOSEST RACES

The eight closest races

Ordered by absolute projected margin. Positive numbers favor Democrats; negative numbers favor Republicans.

Eight closest governor races in the current COYNTOSS forecast
StateRatingProjected marginDetails
Loading closest governor races…

POLLING CONTEXT

Polls used by the model

These surveys are included in the governor model’s polling blend.

Latest poll
September 20, 2026

Governor polls used by the current model
DateRacePollsterSampleResult
Aug 24ConnecticutUniversity of New Hampshire809 likely votersLamont (D) 52% · Fazio (R) 37%
Aug 22IowaSuffolk University500 likely votersSand (D) 48% · Lahn (R) 44%
Aug 19ArizonaNextGen P1,627 likely votersHobbs (D) 39% · Biggs (R) 38%
Aug 18ArkansasJ.L. Partners803 likely votersLove (D) 40% · Sanders (R) 45%
Aug 17GeorgiaInsiderAdvantage800 likely votersBottoms (D) 46% · Jackson (R) 46%
Aug 17ArizonaQuantus Insights780 likely votersHobbs (D) 49% · Biggs (R) 43%
Aug 16ArizonaKreate Strategies900 likely votersHobbs (D) 45% · Biggs (R) 40%
Aug 16KansasGlobal Strategy Group600 likely votersHolscher (D) 47% · Masterson (R) 46%
Aug 11MichiganFabrizio, Lee & Associates877 likely votersBenson (D) 49% · James (R) 44%

METHODOLOGY

How it works

  1. Baseline. State partisan lean is combined with the national generic-ballot margin: 65% of state lean and 55% of the generic ballot.
  2. Candidates and incumbents. Candidate experience is capped at ±3 points; incumbency adds ±2.5 points; challenger quality is capped at ±2.
  3. Expert ratings. Cook and Sabato ratings are translated to approximate margins (Toss-up 0, Lean ±5, Likely ±10, Safe/Solid ±18) and averaged.
  4. Polling. Polls use a 30-day half-life and receive more weight as effective polling weight accumulates, up to a 90% ceiling.

Limitations. The incumbent-history adjustment was selected using pre-2014 data and evaluated on 254 usable governor races from 2004–2022. Historical polling and expert-rating series were not consistently reconstructable, so the poll blend and uncertainty tiers remain provisional.

Read the full governor methodology