Head-to-head vs VotePredictor Elections
The fair, apples-to-apples test: on the 11 races Leger actually polled, how its final poll's margin compared to what VotePredictor Elections predicted for those same races.
| Model | Avg miss (pts) | Called right |
|---|---|---|
| Leger | 3.02 | 73% |
| VotePredictor Elections | 1.13 | 100% |
VotePredictor Elections 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 (11)
Each race Leger 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 US President | D+0.9 | R+1.5 | 2.4 | ✗ |
| 2020 US President | D+8.0 | D+4.4 | 3.6 | ✓ |
| 2016 WA Governor | D+9.0 | D+8.8 | 0.2 | ✓ |
| 2016 AZ President | R+2.0 | R+3.5 | 1.5 | ✓ |
| 2016 CA President | D+23.0 | D+30.1 | 7.1 | ✓ |
| 2016 NV President | EVEN | D+2.4 | 2.4 | ✗ |
| 2016 US President | D+4.0 | D+2.2 | 1.8 | ✓ |
| 2016 WA President | D+15.0 | D+15.7 | 0.7 | ✓ |
| 2016 AZ Senate | R+8.0 | R+13.0 | 5.0 | ✓ |
| 2016 NV Senate | R+4.0 | D+2.4 | 6.4 | ✗ |
| 2016 WA Senate | D+16.0 | D+18.0 | 2.0 | ✓ |
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 wk | 11 | 2.55 | -1.73 | 82% |
| 3–6 wk | 4 | 4.16 | -1.58 | 100% |
| 6–9 wk | 11 | 3.92 | -2.23 | 82% |
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) |
|---|---|---|---|
| 2016 | 18 | 3.5 | R+1.4 |
| 2020 | 9 | 3.5 | 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.