Theory-grounded election forecasting

Forecast the election.
Without copying the crowd.

We simulate the electorate from real Census data, weight it by who actually votes, anchor it to real exit-poll results, and benchmark it on hundreds of real counted elections — instead of guessing with one model or copying the market price.

The problem

Prediction markets are smart — and reflexive. Most “models” secretly anchor to the market price, so their “edge” is circular.

election-oracle forecasts independently, and uses the market only as the baseline to beat.

Architecture

The 7-layer accuracy stack

Each layer is a discipline borrowed from the forecasting literature — stacked so errors cancel instead of compound.

01

Retrieval

Point-in-time news context (GDELT)

02

Ensemble

Median of N independent samples

03

Superforecaster scaffold

Tetlock discipline: reference class → base rate → adjust

04

Calibration

Isotonic recalibration + Brier / ECE, scored out-of-sample

05

As-of entity index

A temporal firewall against lookahead leakage

06

Causal neuro-symbolic

An LLM proposes structure, a transparent engine does the math

07 ★ Star feature

Agent-based electorate

FlockVote — a simulated electorate that votes

FlockVote · Layer 07

A simulated electorate

We don’t ask one model to guess a state. We simulate the people who decide it — Census-real voters, weighted by who actually turns out, anchored to how each group really voted.

Census-real voters

Drawn from real US Census ACS across 9 demographic dimensions, with party and religion conditioned on race and education — so each agent is internally coherent, not a random mix of traits.

Turnout-weighted & exit-poll anchored

Elections are decided by voters, not residents — so we weight by real CPS turnout and anchor each demographic cell to its real exit-poll vote share. The simulation matches how each group actually voted.

Distribution-elicited, mode-collapse resistant

We elicit a full distribution of vote intent per cell rather than one point answer — countering the LLM tendency to collapse to a single safe mode and washing out real demographic spread.

A no-LLM structural core

The measured accuracy is carried by a transparent engine — Census × exit-poll × turnout × Cook PVI, calibrated — that runs without any LLM. The structure is what scores, so the number is reproducible.

Seat aggregation → P(Senate control)

Per-state forecasts roll up through Monte Carlo into P(party controls the Senate) — so national “who controls the chamber” markets can be forecast coherently from the same simulated electorate.

A social network

Voters influence their neighbours on a small-world graph — some lead, some follow the crowd — letting late-breaking shifts propagate the way real opinion does.

9 demographic dimensions· turnout-weighted· exit-poll anchored· distribution-elicited· no-LLM structural core· seat aggregation
Engineered for accuracy

Built so the score is real.

01 / no leakage

As-of firewall

Backtests only see evidence that existed before each race resolved — no lookahead leakage, so the benchmark number is real, not inflated by hindsight.

02 / no circular edge

Market-price independence

The forecast reasons from evidence, never from the price it’s trying to beat — so its edge isn’t a circular echo of the market. The market is the baseline, scored only at the end.

03 / measured accuracy

Proven on real outcomes

The structural core — real Census demographics × exit-poll voting patterns × CPS turnout weighting × Cook PVI — earns +0.36 Brier skill over the base rate across 546 real U.S. Senate & presidential races (MIT Election Lab), scored leave-one-out, out-of-sample.

+0.36 Brier skill · 546 real races · ~75% directional · out-of-sample

The hard part holds up: the near-50/50 Senate races score as well as the easy presidential ones — accuracy where forecasts usually break.

Limitation: we’re tightening this with PVI-off / pre-2016 ablations and a pre-registered forward test.

The accuracy lineage

We didn’t invent the accuracy.
We implemented the literature.

Poststratification / MRPGelman Turnout modelingGhitza & Gelman Silicon samplingArgyle et al. Structural promptingCui & Wei

Built on open source

mesanetworkxscikit-learnGDELTUS Census ACSFREDPolymarketMIT Election Lab