AFOS-Analytics1/chile-2025-electoral-divergence

AFOS · Chile 2025 Electoral Divergence Dataset

🌐 English · Español · Português

Open dataset cross-referencing opinion polls × prediction markets for Chile's 2025 presidential election (first round 16 November 2025; runoff 14 December 2025, Jeannette Jara vs José Antonio Kast), built in the same spirit as the AFOS Brazil 2026 dataset: sources are reported side by side with explicit divergence, not blended into a single average.

Maintained by AFOS Analytics. Part of AFOS's expansion of its electoral-divergence method beyond Brazil. No personal data — only public electoral information.

---

English

Contents (start with the polls):

Path	Rows	Content
polls/chile-first-round-polls.csv	600	First-round voting intention, long format (one row per candidate × poll), all candidates, 100 polls, Jan→Nov 2025.
polls/chile-runoff-polls.csv	21	Runoff head-to-head (Jara vs Kast), Nov→Dec 2025.
polls/chile-polls.json	—	Full structured polls (first round + runoff) with pollster, fieldwork, sample.
data/chile-market-odds-timeseries.csv	2,228	Daily Polymarket win-probability per candidate (11 candidates, May→Nov 2025) from the "Chile Presidential Election" market.
data/chile-divergence-timeseries.csv	532	Market × poll divergence per candidate — each first-round poll joined to the candidate's market odds on its date.
data/chile-poly-raw.json	—	Raw Polymarket payload (event + per-candidate price histories), kept for provenance.

_{Market data fetched from Polymarket's gamma-api + clob via a US-resolving function. Divergence covers the first round; the winner market resolves to the overall election, so its probabilities already price the runoff.}

⚖️ Notable divergences (why divergence beats the average)

The point of this dataset is the gap between what the market prices (probability of winning the presidency) and what polls measure (first-round vote share) — read across the full daily series.

• José Antonio Kast — the market saw the runoff before the polls did. In late first-round polling Kast sat around 17–21% of the vote, at times behind Jara — yet the market priced his probability of winning the presidency at ~66% (a market−poll gap of +45 pp). The market was pricing the second round, where the fragmented right would consolidate behind him. He won the runoff, 58%–42%. The divergence was signal, not noise — confirmed by the result.
• Jeannette Jara — led the first round, but the market never believed she'd win. She topped almost every first-round poll (26%, and won the first round with 26.8%) while the market gave her only **16%** to win the presidency. The gap (−9 pp) captured exactly what vote share couldn't: a first-round leader with a low runoff ceiling.
• Franco Parisi — the late surge polls undersold. Polls had him around 10–14%; he finished third with ~20%, while the market priced his win probability near 1% (never a contender to win, even as his vote share climbed).

The reading: first-round vote share and win-probability are different quantities, and in a two-round system the spread between them is the signal. A blended average of "the polls" would have called Jara the favorite; the market called the eventual president.

Pollsters covered: Cadem, Criteria, Activa (Pulso Ciudadano), Data Influye, CERC-Mori, Panel Ciudadano UDD, Black & White, Feedback, and others.

Provenance & method: poll figures are compiled deterministically (rowspan/colspan-aware HTML parser) from the public Wikipedia aggregation "Opinion polling for the 2025 Chilean presidential election." Polls before the June 2025 left-wing primary measured pre-candidates (Carolina Tohá, Gonzalo Winter) separately from Jara, who then unified the left vote. Market odds come from the public Polymarket market. Nothing is imputed or smoothed; missing values are left blank.

License (dual): data → CC BY 4.0 (LICENSE-CC-BY-4.0); code/scripts → Apache 2.0 (LICENSE-APACHE-2.0), matching the repo root and the Hugging Face mirror. Underlying poll numbers are facts released by the named pollsters; the Wikipedia aggregation is CC BY-SA. Please attribute AFOS Analytics and the original pollsters.

Cite: AFOS Analytics. Chile 2025 Electoral Divergence Dataset. Hugging Face, 2026. CC BY 4.0. (see CITATION.cff)

Disclaimer: observational research. Not investment advice, not voting guidance.

---

Español

Dataset abierto que cruza encuestas × mercados de predicción para la elección presidencial de Chile 2025 (primera vuelta 16 nov 2025; segunda vuelta 14 dic 2025, Jeannette Jara vs José Antonio Kast), con divergencia explícita entre fuentes en lugar de un promedio único.

• polls/chile-first-round-polls.csv — intención de voto en primera vuelta, formato largo, todos los candidatos, 100 encuestas (ene→nov 2025).
• polls/chile-runoff-polls.csv — segunda vuelta cara a cara (Jara vs Kast), 21 encuestas.
• data/chile-market-odds-timeseries.csv / data/chile-divergence-timeseries.csv — probabilidad de Polymarket por candidato y divergencia mercado × encuesta.

⚖️ Divergencias destacadas (por qué la divergencia supera al promedio)

• José Antonio Kast — el mercado vio el balotaje antes que las encuestas. En la primera vuelta rondaba el 17–21% del voto, a veces detrás de Jara — pero el mercado valoraba su probabilidad de ganar la presidencia en ~66% (brecha de +45 pp). Estaba precificando la segunda vuelta, donde la derecha fragmentada se uniría tras él. Ganó el balotaje, 58%–42%. La divergencia fue señal, no ruido.
• Jeannette Jara — lideró la primera vuelta, pero el mercado nunca creyó que ganaría. Encabezó casi todas las encuestas de primera vuelta (26%, y la ganó con 26,8%) mientras el mercado le daba solo **16%** de ganar la presidencia (brecha −9 pp): una líder de primera vuelta con techo bajo en balotaje.
• Franco Parisi — la sorpresa tardía que las encuestas subestimaron. Las encuestas lo daban ~10–14%; terminó tercero con ~20%, con el mercado valorando su probabilidad de ganar cerca del 1%.

La lectura: voto de primera vuelta y probabilidad de ganar son cantidades distintas; en un sistema de dos vueltas, la brecha es la señal. Un promedio de "las encuestas" habría coronado a Jara; el mercado nombró al presidente electo.

Encuestadoras: Cadem, Criteria, Activa, Data Influye, CERC-Mori, Panel Ciudadano UDD, Black & White, Feedback, entre otras. Fuente: agregación pública de Wikipedia; cada cifra remite a una encuestadora con nombre. Licencia: CC BY 4.0 (atribuir a AFOS Analytics y a las encuestadoras). Investigación observacional; no es asesoría de inversión ni orientación de voto.

---

Português

Dataset aberto cruzando pesquisas × mercados de previsão para a eleição presidencial do Chile 2025 (1º turno 16/nov; 2º turno 14/dez, Jeannette Jara × José Antonio Kast), com divergência explícita entre fontes. Pesquisas (todos os candidatos, 100 do 1º turno + 21 do 2º) compiladas deterministicamente da agregação pública da Wikipedia; odds do Polymarket. Licença CC BY 4.0 (atribuir AFOS Analytics + institutos originais). Pesquisa observacional; não é recomendação de investimento nem orientação de voto.

⚖️ Divergências em destaque (por que a divergência supera a média)

O ponto é a diferença entre o que o mercado precifica (probabilidade de vencer a presidência) e o que as pesquisas medem (voto de 1º turno) — lida na série diária inteira.

• José Antonio Kast — o mercado viu o 2º turno antes das pesquisas. No fim do 1º turno tinha 17–21% do voto, às vezes atrás de Jara — mas o mercado precificava a chance dele vencer a presidência em **66%** (diferença de +45pp), já precificando o 2º turno, onde a direita fragmentada se uniria a ele. Ele venceu o runoff, 58%–42%. A divergência foi sinal, não ruído — confirmada pelo resultado.
• Jeannette Jara — liderou o 1º turno, mas o mercado nunca acreditou que venceria. Liderou quase toda pesquisa de 1º turno (26%, e venceu o 1º turno com 26,8%) enquanto o mercado lhe dava só **16%** de vencer a presidência (diferença −9pp): líder de 1º turno com teto baixo no 2º turno.
• Franco Parisi — a surpresa tardia que as pesquisas subestimaram. Pesquisas davam ~10–14%; terminou em 3º com ~20%, com o mercado precificando a chance de vencer perto de 1%.

A leitura: voto de 1º turno e probabilidade de vencer são quantidades diferentes; num sistema de dois turnos, a diferença é o sinal. Uma média das "pesquisas" teria coroado Jara; o mercado nomeou o presidente eleito.

---

Sources / Fuentes / Fontes: Pollsters (Cadem, Criteria, Activa, Data Influye, CERC-Mori, …) · Wikipedia aggregation · Polymarket. Column definitions in DATA_DICTIONARY.md.