Clacton: Farage held, and the model was 0.8 points off.

The count

What we said, and what happened

Reform hold, and the closest call the special engine has made. Vote-Scope's final run put Farage at a 62.6% median with a 99.6% win probability; the count gave him 63.34% — 0.78 points of error on the winner's share. Count Binface was projected at 20.8% [13.9–29.7] and took 26.93%: above the median, comfortably inside the interval, and the reason the range mattered. Where the model was wrong was below the fold — Laurence Fox was projected at 3.0% and got 0.99%, and the 31 micro-candidates were projected at 13.3% against 8.74% counted. The floor for a record 34-candidate field, calibrated against Haltemprice & Howden 2008, was too generous by about four and a half points, and that surplus went to Binface. One structural fact the projection did not carry: Farage won more raw votes than in 2024 (22,239 against 21,225) while the total collapsed 23.6% — the entire drop came from the other side, down 48%. Source: Tendring District Council declaration, 44.37% turnout on 79,785 registered, 32 lost deposits.

The Vote-Scope forecast

50,000 Monte Carlo simulations of voter flows from the 2024 result: how much of his own bloc Farage keeps under investigation, where each boycotting party's orphaned voters go (or whether they stay home), how far the protest vote consolidates behind Binface, and what the media circus pulls in. The first constituency poll has now landed and the model re-centres on it — see below for exactly how much weight it is given.

Projected vote share (10th–90th percentile, median marked)

v0.7 — the constituency poll is now discounted a second time for being the only local poll: its effective weight falls to about nine respondents after sampling, age, framing and single-poll uncertainty are combined. That moves Binface from roughly 0.03% to 0.42% — still a long shot, but without false precision.

Path to victory — how could the bin actually win?

Baseline: Vote-Scope gives Count Binface < 1% — the markets say 0.9%. For the upset, several things must break at once. Conditional win probabilities, from the same simulations:

Verdict: very unlikely — but this is the one election where the improbable is the most fun to watch. If Binface clears 25% while Farage sinks toward 40%, election night gets interesting.

The constituency poll

Unlike the markets below, this poll does feed the model. Here is what it says, what the model projects, and the gap between the two.

Survation / Mandate Research · 2026-07-16 → 2026-07-24 · n = 502

How much weight the model gives it

First and only constituency poll. The model takes it into account without reducing itself to it: it is given an effective weight of about 9 respondents — sampling precision is cut because it is the only local poll, by a 9-day field period closing 20 days before polling day, and by the framing sensitivity the poll itself reveals (Binface at 20% in the four-way question, 31% head-to-head).

Turnout, postal votes & the Reform slide

What the model can use — and what it deliberately leaves out.

Turnout is the one number the returning officer announces before the result. Our forecast assumes turnout in the range a by-election normally produces — the shaded band. The model also says which way an unusually heavy turnout would cut: towards Count Binface, because Farage's own 2024 voters are the bloc least likely to be moved by a bigger crowd, while everyone else in the constituency has to be mobilised to appear at all. How heavy is heavy enough to change the outcome is precisely what the model cannot pin down, so the chart carries no figures.

Prediction markets & bookmakers

External comparison only — market prices never feed the model. Binface opened around 500/1 with the bookies and shortened to ~5/1 within days.

Binface vote share — model vs Polymarket buckets

Polymarket also prices Farage at 98% to beat his 2024 share (46.2%) — the model agrees, at 97%.

Why this by-election exists

• Farage resigned on 7 July 2026 amid parliamentary investigations: a £5M payment from Christopher Harborne, possible gifts in kind, alleged Bank of England crypto lobbying, and property declarations — with the £5M also reported to the National Crime Agency.
• By resigning and recontesting, Farage seeks a fresh mandate at a moment of his choosing — a later suspension could still trigger a recall petition and yet another by-election.
• Labour, the Conservatives, the Lib Dems, the Greens and Restore Britain are all boycotting what Kemi Badenoch called a "fake by-election".
• The closest precedent: Haltemprice & Howden 2008 — David Davis resigned to recontest his seat, and the major parties stood aside.

National mood — context only

2024 baseline — the seat Farage vacated

July 4, 2024 general election result in Clacton.

Timeline

1. 2026-07-07 Farage resigns as MP; the £5M payment is reported to the NCA the same day.
2. 2026-07-07 Major parties vow to boycott within hours.
3. 2026-07-08 Count Binface declares. Bookmakers slash his odds from ~500/1 to ~5/1.
4. 2026-07-09 Ipsos: 33% of Britons would prefer Binface to win, vs 21% for Farage.
5. 2026-07-10 Tendring District Council confirms polling day: 13 August 2026.
6. 2026-07-17 Nominations close — final candidate list.
7. 2026-08-13 Polling day.

Sources

1. Wikipedia — 2026 Clacton by-election (date, candidats, boycott)
2. BBC News — Clacton by-election to be held 13 August
3. The Independent — Survation constituency poll (16–24 July, n=502)
4. Ipsos — British public more likely to prefer Count Binface (8–9 July 2026, n=1000)
5. Electoral Commission — how postal votes are verified and counted
6. Polymarket — Clacton by-election Winner
7. Polymarket — Clacton 2nd place
8. Polymarket — Count Binface Vote %
9. Kalshi — Clacton by-election winner
10. Oddschecker — Clacton by-election odds
11. The Independent — Binface odds slashed