RMG Research
Graded against the actual result across 30 races (from 50 polls, through 2024).
Head-to-head vs the VotePredictor model
The fair, apples-to-apples test: on the 30 races RMG Research actually polled, how its final poll's margin compared to what the VotePredictor model predicted for those same races.
| Model | Avg miss (pts) | Called right |
|---|---|---|
| RMG Research | 6.03 | 77% |
| VotePredictor | 3.59 | 90% |
VotePredictor aggregates all the pollsters, so it's expected to beat any single one on margin — that's the value of averaging. The honest comparison among forecasters is on the combined board.
Every race (30)
Each race RMG Research polled, scored on its final poll — the call right before the vote — against the actual Dem−Rep result. Click a race for its full detail.
| Race | Their call | Result | Miss | Call |
|---|---|---|---|---|
| 2024 FL President | R+5.0 | R+13.1 | 8.1 | ✓ |
| 2024 US President | D+1.0 | R+1.5 | 2.5 | ✗ |
| 2024 FL Senate | R+8.0 | R+12.8 | 4.8 | ✓ |
| 2022 FL Governor | R+10.0 | R+19.4 | 9.4 | ✓ |
| 2022 FL Senate | R+5.0 | R+16.4 | 11.4 | ✓ |
| 2020 MT Governor | R+3.0 | R+12.9 | 9.9 | ✓ |
| 2020 NC Governor | D+12.0 | D+4.5 | 7.5 | ✓ |
| 2020 UT Governor | R+24.0 | R+32.6 | 8.6 | ✓ |
| 2020 MT-1 House | EVEN | R+12.8 | 12.8 | ✗ |
| 2020 UT-4 House | R+1.0 | R+1.0 | 0.0 | ✓ |
| 2020 AZ President | D+1.0 | D+0.3 | 0.7 | ✓ |
| 2020 CO President | D+8.0 | D+13.5 | 5.5 | ✓ |
| 2020 FL President | D+4.0 | R+3.4 | 7.4 | ✗ |
| 2020 IA President | EVEN | R+8.2 | 8.2 | ✗ |
| 2020 MI President | D+7.0 | D+2.8 | 4.2 | ✓ |
| 2020 MT President | R+4.0 | R+16.4 | 12.4 | ✓ |
| 2020 NC President | D+1.0 | R+1.3 | 2.3 | ✗ |
| 2020 PA President | D+6.0 | D+1.2 | 4.8 | ✓ |
| 2020 TX President | R+4.0 | R+5.6 | 1.6 | ✓ |
| 2020 US President | D+7.0 | D+4.4 | 2.6 | ✓ |
| 2020 UT President | R+12.0 | R+20.5 | 8.5 | ✓ |
| 2020 WI President | D+6.0 | D+0.6 | 5.4 | ✓ |
| 2020 AZ Senate | D+7.0 | D+2.3 | 4.7 | ✓ |
| 2020 CO Senate | D+9.0 | D+9.3 | 0.3 | ✓ |
| 2020 GA Senate | D+1.0 | D+2.1 | 1.1 | ✓ |
| 2020 IA Senate | D+3.0 | R+6.6 | 9.6 | ✗ |
| 2020 MI Senate | D+9.0 | D+1.7 | 7.3 | ✓ |
| 2020 MT Senate | R+2.0 | R+10.0 | 8.0 | ✓ |
| 2020 NC Senate | D+7.0 | R+1.7 | 8.7 | ✗ |
| 2020 TX Senate | R+6.0 | R+9.6 | 3.6 | ✓ |
Accuracy by time to election
Lower is better. Time to election runs right (election week) to left (~2 months out).
By the numbers
| Time to election | Polls | Avg miss | vs field | Called right |
|---|---|---|---|---|
| ≤1 wk | 4 | 5.37 | +1.09 | 75% |
| 1–3 wk | 21 | 6.14 | +1.07 | 71% |
| 3–6 wk | 18 | 5.04 | -0.70 | 78% |
| 6–9 wk | 7 | 1.48 | -4.67 | 86% |
vs field is this pollster's average miss minus all pollsters' at the same lead time — green beats the field, redtrails it. Our historical polls reach ~2 months out; earlier polling isn't in the record.
Track record by cycle — getting better?
| Year | Polls | Avg miss | Lean (house effect) |
|---|---|---|---|
| 2020 | 45 | 4.8 | D+4.5 |
| 2024 | 3 | 5.1 | D+5.1 |
Do we credit a pollster for fixing its bias? Each cycle, the model re-estimates every pollster's lean from all its earlier polls (walk-forward) and subtracts it before using the poll. We tested weighting recent cycles more — it doesn't help: a pollster's lean in one cycle barely predicts the next (correlation 0.28), so the swings above are mostly noise, and averaging over more history beats chasing the latest cycle. The all-time estimate we use came out within ~0.5% of the best option.