Research
Between March and April 2026 we catalogued 52 papers from SSRN (15 Tier 1, 15 Tier 2, 22 Tier 3), extracted a 20-concept inventory, and tested 10 hypotheses about the prediction market ecosystem. Five hypotheses were upgraded; one was downgraded. The synthesis directly shapes product design.
Hypothesis scorecard
Section titled “Hypothesis scorecard”| Hypothesis | Before | After | Change |
|---|---|---|---|
| H1: Forecasting skill is real and persistent | 🟢 | 🟢+ | Strengthened — Reichenbach (t=5.14), Yang (96.2%) |
| H2: Skill determines profitability | 🟢 | 🟢 (reframed) | Refined — execution is a second axis (Vedova) |
| H3: Resolution risk is insurable | 🟡 | 🟢 | Upgraded — Ma ARI, Sanjabi SRS, Cong Trilemma |
| H4: Vega Wedge is real and priceable | 🟢 | 🟢+ | Strengthened — Yang λ=0.18 quantified |
| H5: Reputation can function as infrastructure | 🟡 | 🟢 | Upgraded — Agustin formal model, Atanasov selection effect |
| H6: Loss aversion is the primary adoption barrier | 🟡 | 🟡 | Refined — barrier is multi-dimensional |
| H7: A small elite outperforms the large crowd | 🟢 | 🟢+ | Strongest-supported in the corpus |
| H8: PM prices are well-calibrated probabilities | 🟢 | 🟡 | Downgraded — systematic wedge, favorite-longshot bias |
| H9: PM ecosystem is fragmenting into layers | 🟡 | 🟢 | Upgraded — 264+ projects, Gaurav lending, Aktug 7-platform hierarchy |
| H10: Institutional demand for PM risk transfer | 🟡 | 🟢 | Upgraded — Fed, ICE, Byerly AURA, Li Treasury yields |
Synthesis pages (by hypothesis cluster)
Section titled “Synthesis pages (by hypothesis cluster)”- Skill is real — H1, H2, H7
- Resolution risk — H3 (largest upgrade in the corpus)
- Pricing wedge — H4, H8
- Reputation as infrastructure — H5
- Retention & the Kelly problem — H6
- Ecosystem layers — H9
- Institutional demand — H10
- Two-score architecture — methodology note
- Reading list — all 52 papers, sortable
Headline citations
Section titled “Headline citations”- Top 1% capture 84% of gains — Akey et al. 2026 (N=1.4M users, $20B volume)
- 3.1% of wallets extract $228M net profit — Yang 2026 (N=152.8M trades)
- Skill persists out of sample: 89.4% vs 50.4% — Yang 2026
- Selection > aggregation by 21% — Atanasov et al. 2024 (IARPA ACE)
- Pricing wedge λ = 0.176 on Polymarket — Yang 2026 (291K contracts, 6 platforms)
- Oracle Trilemma is a formal impossibility — Cong 2025
- Oracle attacks cost $403M in 2022 — Sanjabi 2026 / Duley
- Adjudication Risk Index ranges 0.6353–0.9995 — Ma 2026 (N=2,738 UMA events)
- 0.1 SRS increase reduces volume 60% — Sanjabi 2026
- Kalshi beats Bloomberg on CPI, perfect FOMC record — FEDS 2026-010 (Diercks, Katz & Wright)
- PM signals explain 27–47% of Treasury yield variation — Li & Xie 2025
- Zero of 264+ projects build insurance/reputation — PredictionIndex 2025–2026
Methodology
Section titled “Methodology”All 52 papers were retrieved from SSRN, parsed, and cataloged with columns for reading order, tier, authors, hypothesis mapping, and extracted abstracts. The full catalog lives in research/academic/SSRN/SSRN-catalog.xlsx and will be published as a sortable table on the reading list page.
Each synthesis page presents: the claim, the evidence that upgraded or downgraded confidence, design implications for Chronomancy, and citations to use in external communication.