Magellan Strategies
Graded against the actual result across 25 races (from 31 polls, through 2018).
Head-to-head vs VotePredictor Elections
The fair, apples-to-apples test: on the 24 races Magellan Strategies actually polled, how its final poll's margin compared to what VotePredictor Elections predicted for those same races.
| Model | Avg miss (pts) | Called right |
|---|---|---|
| Magellan Strategies | 5.92 | 88% |
| VotePredictor Elections | 4.02 | 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 (25)
Each race Magellan Strategies 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 |
|---|---|---|---|---|
| 2018 CO Governor | D+5.0 | D+10.6 | 5.6 | ✓ |
| 2018 OK Governor | R+7.0 | R+12.1 | 5.1 | ✓ |
| 2016 CO President | D+6.5 | D+4.9 | 1.6 | ✓ |
| 2016 CO Senate | D+15.0 | D+5.7 | 9.3 | ✓ |
| 2014 IA Governor | R+16.0 | R+21.7 | 5.7 | ✓ |
| 2014 ME Governor | EVEN | R+4.8 | 4.8 | ✗ |
| 2014 PA Governor | D+7.0 | D+9.9 | 2.9 | ✓ |
| 2014 IA Senate | R+9.2 | R+8.3 | 0.9 | ✓ |
| 2014 KY Senate | R+8.0 | R+15.5 | 7.5 | ✓ |
| 2014 LA Senate | R+16.0 | R+11.9 | 4.1 | ✓ |
| 2014 MI Senate | D+5.0 | D+13.3 | 8.3 | ✓ |
| 2014 NH Senate | R+1.6 | D+3.2 | 4.8 | ✗ |
| 2012 LA President | R+23.0 | R+17.2 | 5.8 | ✓ |
| 2011 NV-2 House | R+13.0 | R+21.9 | 8.9 | ✓ |
| 2011 NY-9 House | R+4.2 | R+1.7 | 2.5 | ✓ |
| 2010 CO Governor | D+35.0 | D+39.9 | 4.9 | ✓ |
| 2010 PA Governor | R+10.0 | R+9.0 | 1.0 | ✓ |
| 2010 AZ-7 House | D+2.0 | D+6.0 | 4.0 | ✓ |
| 2010 CO-2 House | D+12.0 | D+19.5 | 7.5 | ✓ |
| 2010 IL-9 House | D+18.0 | D+35.2 | 17.2 | ✓ |
| 2010 NY-22 House | EVEN | D+11.9 | 11.9 | ✗ |
| 2010 DE Senate | D+18.0 | D+16.6 | 1.4 | ✓ |
| 2010 PA Senate | R+8.8 | R+2.0 | 6.8 | ✓ |
| 2008 NV President | D+3.0 | D+12.5 | 9.5 | ✓ |
| 2000 KY President | R+15.0 | R+15.1 | 0.1 | ✓ |
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 | 9 | 4.38 | -0.69 | 78% |
| 3–6 wk | 17 | 5.40 | -0.34 | 100% |
| 6–9 wk | 5 | 5.32 | -0.83 | 80% |
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) |
|---|---|---|---|
| 2010 | 9 | 6.4 | R+6.1 |
| 2014 | 10 | 4.3 | R+0.7 |
| 2016 | 3 | 3.7 | D+3.7 |
| 2018 | 3 | 4.8 | R+1.4 |
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.