WPA Intelligence
Graded against the actual result across 33 races (from 51 polls, through 2024).
Head-to-head vs the VotePredictor model
The fair, apples-to-apples test: on the 31 races WPA Intelligence 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 |
|---|---|---|
| WPA Intelligence | 5.90 | 74% |
| VotePredictor | 4.56 | 84% |
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 (33)
Each race WPA Intelligence 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 CA-47 House | R+3.0 | D+2.8 | 5.8 | ✗ |
| 2024 OK President | R+36.0 | R+34.3 | 1.7 | ✓ |
| 2022 OK Governor | R+13.0 | R+13.7 | 0.7 | ✓ |
| 2022 NV Senate | R+2.0 | D+0.8 | 2.8 | ✗ |
| 2020 TX-21 House | R+5.0 | R+6.6 | 1.6 | ✓ |
| 2020 NV President | D+2.0 | D+2.4 | 0.4 | ✓ |
| 2018 FL-15 House | R+7.0 | R+6.0 | 1.0 | ✓ |
| 2018 IA-4 House | R+18.0 | R+3.3 | 14.7 | ✓ |
| 2018 MN-2 House | R+3.0 | D+5.5 | 8.5 | ✗ |
| 2018 TX-21 House | R+12.0 | R+2.6 | 9.4 | ✓ |
| 2018 TX Senate | R+9.0 | R+2.6 | 6.4 | ✓ |
| 2016 AZ-2 House | R+19.0 | R+13.9 | 5.1 | ✓ |
| 2016 MN-2 House | R+3.0 | R+1.8 | 1.2 | ✓ |
| 2016 WI Senate | D+3.0 | R+3.4 | 6.4 | ✗ |
| 2014 MD Governor | R+5.0 | R+3.8 | 1.2 | ✓ |
| 2014 ND-1 House | R+10.0 | R+17.1 | 7.1 | ✓ |
| 2014 LA Senate | R+24.0 | R+11.9 | 12.1 | ✓ |
| 2012 NV-3 House | R+11.0 | R+7.5 | 3.5 | ✓ |
| 2012 TX President | R+15.0 | R+15.8 | 0.8 | ✓ |
| 2012 WI Senate | D+3.0 | D+5.5 | 2.5 | ✓ |
| 2010 CA Governor | R+1.0 | D+12.9 | 13.9 | ✗ |
| 2010 DE-1 House | D+4.0 | D+15.7 | 11.7 | ✓ |
| 2010 FL-22 House | R+6.0 | R+8.7 | 2.7 | ✓ |
| 2010 MO-4 House | EVEN | R+5.3 | 5.3 | ✗ |
| 2010 OR-4 House | D+6.0 | D+10.9 | 4.9 | ✓ |
| 2010 TN-3 House | R+30.0 | R+28.8 | 1.2 | ✓ |
| 2010 CA Senate | R+3.0 | D+10.0 | 13.0 | ✗ |
| 2010 OH Senate | R+11.0 | R+17.4 | 6.4 | ✓ |
| 2006 KS-2 House | R+17.0 | D+3.5 | 20.5 | ✗ |
| 2004 OK President | R+33.0 | R+31.1 | 1.9 | ✓ |
| 2004 OK Senate | R+3.0 | R+11.5 | 8.5 | ✓ |
| 2002 LA Senate | R+8.0 | D+3.4 | 11.4 | ✗ |
| 1998 ND-1 House | D+9.0 | D+15.1 | 6.1 | ✓ |
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 | 18 | 6.75 | +1.68 | 67% |
| 3–6 wk | 24 | 7.26 | +1.52 | 71% |
| 6–9 wk | 9 | 5.53 | -0.62 | 78% |
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) |
|---|---|---|---|
| 2004 | 16 | 8.2 | D+6.8 |
| 2010 | 8 | 7.4 | R+3.8 |
| 2012 | 3 | 2.3 | R+1.8 |
| 2014 | 4 | 6.3 | R+0.4 |
| 2016 | 3 | 4.2 | D+0.0 |
| 2018 | 6 | 9.4 | R+9.4 |
| 2022 | 4 | 1.7 | D+0.3 |
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.