Oklahoma · U.S. Senate
Open this race on The Map ↗Kevin Hern is heavily favored to keep Oklahoma’s Senate seat in Republican hands. Predicted margin: R+25.4.
high / low: vote share (%) in the middle 80 in 100 simulations
About this field
Hern won the Republican nomination for the special election to Markwayne Mullin's former seat. Thomas leads Jim Priest into the August 25 Democratic runoff; James Lankford holds Oklahoma's other Senate seat.
Forecast share averages the simulated vote. Win chance counts how often a candidate wins. A wider range of outcomes can give a candidate a higher win chance despite a lower average share.
Forecast history
Latest reading · Sep 13 · first published Sep 9 · 80% range · margin points R+35.8 to R+15.1. Methodology
Hover the line, tap it, or use the arrow keys to inspect its history.
About this series
This comparable series begins on September 9, 2026, when the model version changed to CTMM 2.2. Earlier readings remain in the archive and are not joined as voter movement; faded points before September 9, 2026 are reconstructions, not forecasts issued on those days.
Faded: CTMM 2.2 reconstructed history: computed after the fact with each date's available polls and environment, without expert ratings; not an issued forecast. Issued forecasts begin at the launch marker.
The polls
Ledger weighted average R+25.3Used by the forecast: 1 poll (60-day half-life) · measured R+24.0 · after the fitted poll-error shift R+25.61 recorded poll · 1 eligible for the aggregate · 1 with positive weight
| Fieldwork | Added | Pollster | Result | Sample |
|---|---|---|---|---|
| from 2026-08-03 | Cole Hargrave ↗ | R+24.0D 25% · R 49% | 500 RV |
How the margin was built
Follow each contribution to the margin. Wider ribbons mean more percentage points. Tap an input for its calculation and source.
The signed totals combine at the forecast. Displayed values are rounded.
All calculations & sources
How this state voted, 2024 −34.260 points
The state's two-party presidential margin in 2024, from our own county sums — the same measurement the model's statewide terrain center is built from. Every pull below is measured from here, not from EVEN, which is why a landslide seat's listed pulls are small numbers and still land where the forecast does.
- 2024 presidential result for this state, two-party, from Call the Map's own county sums.
- Same source as Act I's 2024-president reference mark, so the origin lands on it.
Fitted model, above the terrain shown +13.505 points
What the fitted model adds on top of the starting terrain above — incumbency, past results, campaign finance and the other fitted terms — measured from that terrain rather than from EVEN. It is a comparison with the starting point already shown, not a separate input. The source detail lists the variables the fit uses.
- Fitted baseline -20.755 minus starting point -34.260 = 13.505 points.
- For orientation, the presidential results here: 2024 R+34.3, 2020 R+33.1.
- The work: the −20.8 D fundamentals push is the sum of the dynamic formula's surviving variables for this race — each one's own contribution is listed below.
- fundamentals core = 25-term dynamic formula (mass fit over the master feature matrix, held-out-cycle validated; see matrix/feature_dictionary.md + mass_fit_report.json)
- 9-term formula, held-out validated.
National environment +1.988 points
The contribution to this race’s margin, not the national polling margin itself. The race’s loading, reference level and polling blend determine how much of the national climate reaches its final forecast.
- Three-source blend, updated 2026-09-08: 52% generic-ballot aggregates (D+6.5) · 28% special-election overperformance (D+9.0 across 46 specials) · 20% approval-implied (D+8.6 from net approval -21.7).
- Latest polls — generic ballot 2026-09-08 (5 days old), approval 2026-09-10 (3 days old).
- Generic-ballot average D+6.5 · support factor 0.84 (the share of that channel's capacity the admitted polls carry) — separate inputs; their product is a capacity diagnostic, not a margin.
- Weights are the 55/25/20 capacities scaled by each channel's evidence and renormalized, so thin polling widens the interval instead of pulling the center toward zero — 16% of the weight is carried that way.
- Blended national environment: D+7.6.
- Applied at full strength to this chamber's races.
- The work: national environment × 1 chamber strength = +2.0 D of this margin.
- Blended national environment: D+7.6 (updated 2026-09-08).
- Effective poll weight 84%; fundamentals share 16%, carried by the interval (×1.08) rather than by the center.
- The support factor is the share of the generic-ballot capacity the admitted polls carry; average × support is a capacity diagnostic, not a margin, and is not what the row above multiplies.
- Applied at full strength to this chamber: +2.0 of this margin.
Structural safeguards −3.501 points
Bounded corrections that keep the model consistent with structural facts.
- Applied only when a safe state's dynamic fundamentals point in the same direction as its strong presidential lean but are materially compressed toward zero.
- The 25% correction toward that lean improved safe-race MAE in strict expanding-window validation without changing a historical call.
- The disclosed adjustment is −3.5 D.
This race’s polls −3.153 points
What this race’s own recent polling adds beyond its prior, after each poll is weighted for recency, sample size and pollster quality.
- 1 matchup-validated, sample/recency/grade-weighted poll(s), latest 2026-08-06
- Cole Hargrave (2026-08-06, n=500): R+24.0 · weight 0.28
- Adjusted poll-bank aggregate R+25.3 · 1 recorded poll · 1 eligible for the aggregate · 1 with positive weight. The sealed model records a net polling contribution of −3.2 D. Observation weights are not the final model blend weight. Eligible polls: latest poll: 2026-08-06 (38 days old).
- Adjusted poll-bank aggregate R+25.3 · 1 recorded poll · 1 eligible for the aggregate · 1 with positive weight; recorded polling contribution −3.2 D. These observation weights must not be read as the model’s final prior-versus-polling weight. Eligible polls: latest poll: 2026-08-06 (38 days old).
The components this edition publishes for this race, in points. “Other factors” closes the sum to the two-party structural lean exactly. Here the structural lean IS the published margin, so the walk sums to the projection exactly; nothing is left over.
| Named factor | Contribution |
|---|---|
| FundamentalsThe admitted fitted baseline. Its surviving variables differ by office; the source detail above identifies those used for this race. | −20.76 |
| Structural terrain safeguardA bounded pull toward presidential terrain in heavily partisan states. | −3.50 |
| This race's pollsAdmitted head-to-head polls of this race, recency- and quality-weighted. | −3.15 |
| National environmentThe national environment’s contribution to the prior. Polling can attenuate its influence; this value is not the national polling margin. | +1.99 |
| Other factors | 0.00 |
Recent elections
| Election | Result | Votes cast |
|---|---|---|
| 2020 · U.S. Senate ↗ | R+31.5 | 1,556,361 |
| 2014 · U.S. Senate ↗ | R+40.9 | 820,733 |
| 2008 · U.S. Senate ↗ | R+18.3 | 1,346,819 |
| 2002 · U.S. Senate ↗ | R+22.4 | 1,018,424 |
Margins compare Democratic and Republican vote totals. Nonstandard ballots are labeled separately.
The money
Kevin Hern ↗
R$10.0Mraised
$2.5M cash on handFiling through 06/30/2026
N'Kiyla Jasmine Thomas ↗
D$57Kraised
$456 cash on handFiling through 08/05/2026
Bar length compares reported receipts in this race; the tick marks cash on hand. A missing filing stays blank rather than becoming $0.
One more fundraising reportCall the Map · $0
The candidates are comfortably out-fundraising the website covering them. A little rude, perhaps, but statistically unsurprising.
- Received so far
- $0
- Measured compute
- $26,188.58
Ko-fi and Buy Me a Coffee receipts are counted through Sep 1, 2026. Measured compute prices model runs at standard published API rates; it is not what I paid. Hosting, data and subscriptions, the domain and my own time are not counted. Donations buy gratitude, never a forecast adjustment.
Outside spending
Support + opposition · separate from receiptsFor and against Kevin Hern$2.4M
6 spending groups
| Group | For | Against |
|---|---|---|
| DEFENDING OUR VALUES PAC | $808K | $0 |
| DEFEND AMERICAN JOBS | $736K | $0 |
| AMERICAN MISSION | $474K | $0 |
| FELLOWSHIP PAC | $300K | $0 |
| SENATE CONSERVATIVES FUND | $89K | $0 |
| SENATE CONSERVATIVES ACTION | $22K | $0 |
Includes primary spending. Opposition is recorded against its named target; no benefit to another candidate is assumed.
No outside spending matched: N'Kiyla Jasmine Thomas.
FEC filings · through 06/30/2026 · Definitions & method
Race context
- Forecast votes cast
- 1,129,591
- Forecast turnout
- 37.9% of the voting-eligible population
- 2024 presidential margin
- R+34.9
Statewide, 38.2% of the voting-eligible population is projected to cast a ballot; the lower figure above is the share projected to vote in this race.
Population profile Select a measure for its source and denominator
U.S. Census Bureau, 2022 American Community Survey 5-year estimates ↗