Examples: three real theses decomposed

Each example shows the original thesis, the atomic claims the AI extracted, the matched Polymarket markets, and the final Gap score.

The examples below were run by the founder personally. Market probabilities are snapshots from the run date and will have drifted since. All Polymarket links point to the current live markets — readers can verify whether the disagreement still exists at today's probability.

Fed cuts no more than 50bps in 2026

Run on 2026-04-10
Gap +12
Original thesis

The Fed will cut no more than 50bps cumulatively through 2026 because core PCE is sticky at 2.5-2.8%, and base effects from the oil retreat will push H2 inflation back above 3%.

Sub-claims extracted by AI
1. The Fed cuts the policy rate at most twice (25bps each) in 2026
Timeframe: 2026-12-31Importance: CriticalUser subjective probability: 65%
Matched Polymarket markets
Will the Fed cut rates 3+ times in 2026?
Contradicts24h volume: $180K
Fed funds rate target range at end of 2026
Supports24h volume: $95K
2. Core PCE YoY ≥ 2.5% in December 2026
Timeframe: 2026-12-31Importance: HighUser subjective probability: 70%
Matched Polymarket markets
Core PCE > 2.5% YoY in December 2026?
Supports24h volume: $32K
3. H2 2026 headline CPI YoY re-accelerates above 3%
Timeframe: 2026-07-01 to 2026-12-31Importance: MediumUser subjective probability: 55%
AI-only signal, not market-backed

Polymarket does not have a direct contract for "H2 CPI > 3%". AI standalone analysis: assuming Brent holds at $75-85, the base effect should kick in from Q3, making upward CPI moderately likely (~45-55%). This is pure AI inference, not money-backed — treat it as a weak signal.

Takeaway

You are more bullish than the market on the slow-cut path (+12pp Gap). The core disagreement is in the first sub-claim — the market leans toward "3+ cuts in 2026". On sticky inflation you and the market largely agree. Takeaway: if you're positioning for slow cuts (e.g. underweight short duration, long gold), the market sees 5-7 points of asymmetric upside in that view.

BTC breaks $120,000 before Q3 2026

Run on 2026-04-15
Gap +22
Original thesis

Bitcoin will touch $120,000 before September 30, 2026. Main drivers: ongoing ETF net inflows, the second institutional-accumulation wave after the halving, and potential Fed cuts releasing risk premium.

Sub-claims extracted by AI
1. BTC touches $120,000 or higher at any point before 2026-09-30
Timeframe: 2026-09-30Importance: CriticalUser subjective probability: 62%
Matched Polymarket markets
Will BTC reach $120K by August 31, 2026?
Supports24h volume: $520K
BTC 2026 peak price range
Supports24h volume: $210K
2. BTC spot ETFs net inflows ≥ $5B in H1 2026
Timeframe: 2026-06-30Importance: HighUser subjective probability: 70%
AI-only signal, not market-backed

Polymarket doesn't have a direct contract for the ETF net-inflow number. AI standalone analysis: 2025 H2 net inflows ran ~$1.5-2B/month, extrapolating to $9-12B in H1 2026. Unless macro liquidity tightens sharply, the $5B threshold likely clears (~70-80%). Weak signal.

3. The Fed cuts rates at least once before September 2026
Timeframe: 2026-09-30Importance: MediumUser subjective probability: 55%
Matched Polymarket markets
Fed cuts rates 1+ times by September 2026?
Supports24h volume: $140K
Takeaway

You're meaningfully more bullish than the market (+22pp Gap, amber). The core disagreement is on the BTC price itself — the market gives "120K by Aug 31 2026" only 41% while you put it at 62%. Your ETF-flow and Fed-cut assumptions are actually more conservative than the market's. Takeaway: this is a speculative view with a measurable gap to consensus; worth spelling out the downside if you're wrong.

OpenAI valuation exceeds $500B by year-end 2026

Run on 2026-04-18
Gap +17
Original thesis

OpenAI's private valuation will exceed $500B by year-end 2026, driven by enterprise ARR moving from ~$15B to $25B+, next-generation flagship-product API acceleration, and sovereign-AI contracts (Saudi Arabia, UAE) providing long-tail cashflow.

Sub-claims extracted by AI
1. OpenAI closes a primary or secondary round at $500B+ valuation by year-end 2026
Timeframe: 2026-12-31Importance: CriticalUser subjective probability: 55%
Matched Polymarket markets
OpenAI valuation reaches $500B by end of 2026?
Supports24h volume: $68K
2. OpenAI enterprise ARR exceeds $20B in 2026
Timeframe: 2026-12-31Importance: HighUser subjective probability: 50%
AI-only signal, not market-backed

Polymarket has no OpenAI ARR contract. AI standalone analysis: end-2025 ARR estimated at $13-16B; reaching $20B requires ~40-50% YoY growth. Given Anthropic growing ~200% in the same period and Microsoft Copilot paid-penetration acceleration, ~50% is not aggressive for OpenAI but not certain. Estimated probability 40-55%. Weak signal.

3. Anthropic valuation also reaches $300B+ by year-end 2026 (as an AI-sector valuation anchor)
Timeframe: 2026-12-31Importance: MediumUser subjective probability: 45%
Matched Polymarket markets
Anthropic valuation reaches $300B in 2026?
Supports24h volume: $22K
Takeaway

You're more bullish than the market (+17pp Gap). All three sub-claims put you 10-17pp above consensus, and the disagreement is evenly distributed rather than concentrated. The "whole-sector-more-bullish" pattern is worth scrutinizing — could be narrative bias from your own positions (CRWV / OKLO class AI infrastructure). Recommend separately verifying the precondition for a $500B OpenAI valuation (at least one trophy primary round needs to actually close).

Want to analyze your own thesis?

ThesisGap does not provide investment advice. Polymarket probabilities reflect real-money crowd consensus but are subject to liquidity, geography, and base-rate biases. AI-generated fallback analyses are labeled as model opinions, not market prices.

Markets can remain irrational longer than you can remain solvent.

Examples · Three real investment theses decomposed · ThesisGap