Cygnal
Graded against the actual result across 56 races (from 75 polls, through 2024).
Head-to-head vs the VotePredictor model
The fair, apples-to-apples test: on the 56 races Cygnal 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 |
|---|---|---|
| Cygnal | 4.31 | 89% |
| VotePredictor | 3.75 | 88% |
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 (56)
Each race Cygnal 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 AK-1 House | R+4.1 | R+1.8 | 2.3 | ✓ |
| 2024 US House | R+0.1 | R+2.8 | 2.7 | ✓ |
| 2024 CA President | D+24.5 | D+20.2 | 4.3 | ✓ |
| 2024 FL President | R+5.1 | R+13.1 | 8.0 | ✓ |
| 2024 NJ President | D+11.5 | D+5.9 | 5.6 | ✓ |
| 2024 NY President | D+17.8 | D+12.6 | 5.2 | ✓ |
| 2024 TX President | R+8.0 | R+13.7 | 5.7 | ✓ |
| 2024 US President | D+2.0 | R+1.5 | 3.5 | ✗ |
| 2024 VA President | D+7.6 | D+5.8 | 1.8 | ✓ |
| 2024 WY President | R+42.5 | R+46.2 | 3.7 | ✓ |
| 2024 CA Senate | D+24.2 | D+17.8 | 6.4 | ✓ |
| 2024 FL Senate | R+3.9 | R+12.8 | 8.9 | ✓ |
| 2024 NJ Senate | D+16.4 | D+9.6 | 6.8 | ✓ |
| 2024 NY Senate | D+19.4 | D+18.3 | 1.1 | ✓ |
| 2024 TX Senate | R+4.0 | R+8.5 | 4.5 | ✓ |
| 2024 VA Senate | D+9.9 | D+9.0 | 0.9 | ✓ |
| 2024 WY Senate | R+47.5 | R+51.4 | 3.9 | ✓ |
| 2022 AL Governor | R+34.2 | R+37.7 | 3.5 | ✓ |
| 2022 IA Governor | R+19.0 | R+18.5 | 0.5 | ✓ |
| 2022 MI Governor | D+3.7 | D+10.5 | 6.8 | ✓ |
| 2022 MN Governor | D+3.4 | D+7.7 | 4.3 | ✓ |
| 2022 NM Governor | D+1.8 | D+6.4 | 4.6 | ✓ |
| 2022 NV Governor | R+5.0 | R+1.5 | 3.5 | ✓ |
| 2022 OH Governor | R+18.9 | R+25.0 | 6.1 | ✓ |
| 2022 FL-27 House | R+6.0 | R+14.6 | 8.6 | ✓ |
| 2022 MI-8 House | R+1.0 | D+10.1 | 11.1 | ✗ |
| 2022 OR-6 House | R+5.0 | D+2.4 | 7.4 | ✗ |
| 2022 AL Senate | R+29.6 | R+35.7 | 6.1 | ✓ |
| 2022 IA Senate | R+11.2 | R+12.2 | 1.0 | ✓ |
| 2022 NC Senate | R+3.8 | R+3.2 | 0.6 | ✓ |
| 2022 NV Senate | R+2.5 | D+0.8 | 3.3 | ✗ |
| 2022 OH Senate | R+6.2 | R+6.1 | 0.1 | ✓ |
| 2021 VA Governor | EVEN | R+1.9 | 1.9 | ✗ |
| 2020 IN Governor | R+18.3 | R+24.5 | 6.2 | ✓ |
| 2020 MO Governor | R+6.6 | R+16.4 | 9.8 | ✓ |
| 2020 US President | D+7.6 | D+4.4 | 3.2 | ✓ |
| 2020 VA President | D+9.0 | D+10.1 | 1.1 | ✓ |
| 2020 AL Senate | R+13.6 | R+20.4 | 6.8 | ✓ |
| 2020 KY Senate | R+10.3 | R+19.5 | 9.2 | ✓ |
| 2020 NE Senate | R+29.9 | R+38.3 | 8.4 | ✓ |
| 2020 TN Senate | R+20.7 | R+27.0 | 6.3 | ✓ |
| 2020 TX Senate | R+8.1 | R+9.6 | 1.5 | ✓ |
| 2020 VA Senate | D+7.1 | D+12.1 | 5.0 | ✓ |
| 2019 LA Governor | D+2.2 | D+2.7 | 0.5 | ✓ |
| 2018 FL Governor | R+0.6 | R+0.4 | 0.2 | ✓ |
| 2018 GA Governor | R+2.2 | R+1.4 | 0.8 | ✓ |
| 2018 OH Governor | R+0.2 | R+3.7 | 3.5 | ✓ |
| 2018 TN Governor | R+22.4 | R+21.0 | 1.4 | ✓ |
| 2018 FL Senate | D+1.8 | R+0.1 | 1.9 | ✗ |
| 2018 IN Senate | R+2.9 | R+5.9 | 3.0 | ✓ |
| 2018 MO Senate | R+3.0 | R+5.8 | 2.8 | ✓ |
| 2018 OH Senate | D+10.4 | D+6.8 | 3.6 | ✓ |
| 2018 TN Senate | R+5.8 | R+10.8 | 5.0 | ✓ |
| 2016 KY President | R+24.0 | R+29.8 | 5.8 | ✓ |
| 2016 KY Senate | R+9.7 | R+14.5 | 4.8 | ✓ |
| 2014 AL Governor | R+21.3 | R+27.3 | 6.0 | ✓ |
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 |
|---|---|---|---|---|
| 1–3 wk | 49 | 4.06 | -1.01 | 92% |
| 3–6 wk | 22 | 4.01 | -1.73 | 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) |
|---|---|---|---|
| 2018 | 9 | 2.5 | D+1.9 |
| 2020 | 12 | 5.0 | D+3.5 |
| 2022 | 30 | 4.1 | R+0.3 |
| 2024 | 19 | 4.2 | D+3.5 |
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.