There are $34.7 million sitting on whether the United States invades Iran before 2027.
There are $33.9 million on whether China invades Taiwan by December 31st.
There are $568,000 on whether a drug called Retatrutide — which could be the most effective obesity treatment in human history — gets FDA approval this year.
We live in a world where the probability of civilizational conflict is more liquid than the probability of a medical breakthrough. That's not a moral judgment. It's a data point. And it tells you something important about where the edges are.
What Prediction Markets Actually Are
Most people think prediction markets are gambling. They are not. They are epistemological infrastructure.
When $270 million accumulates in markets about US-Iran relations, you are not looking at gamblers. You are looking at hedge funds, energy traders, defense analysts, and geopolitical risk desks pricing their uncertainty. Real capital. Real stakes. Real information.
The crowd is buying and selling probability like it's a commodity — because to them, it is. A fund with Iranian energy exposure doesn't buy oil futures and hope for the best. They buy NO on "Iran closes the Strait of Hormuz" at 91¢ as a hedge. That's not gambling. That's risk management in the only venue that prices it continuously.
After analyzing 28,798 resolved Polymarket markets representing $7.9 billion in volume, we found that 74.4% resolve NO. The crowd systematically overprices YES outcomes. Markets priced between 0-10% YES resolve YES only 2.5% of the time — the market is even wrong about near-zero probabilities.
This is the NO bias. It is the most durable finding in the prediction market literature. And it is systematically exploitable.
The Taxonomy of Human Uncertainty
Last week we discovered something that changed how we think about this product.
Polymarket maintains a structured catalog of 6,512 tags — a machine-readable taxonomy of every domain where humans are willing to bet real money on uncertain outcomes. Not keywords. Not categories. Verified tags, each with a unique ID, each mapping to a curated set of live markets.
Tag 101489: FDA. Tag 78: Iran. Tag 303: China. Tag 159: Federal Reserve. Tag 793: Defense.
When you paginate through all 6,512 tags, you get something remarkable: a map of collective human uncertainty in 2026. What are we most uncertain about? What captures enough attention that people will put real money on the outcome?
By volume, here is what humanity was pricing in June 2026:
The geopolitical layer: $270M on Iran. $33M on Taiwan. $10M on Russia. These aren't casual bets. This is institutional money pricing tail risks that don't have another market.
The policy layer: $72M on Fed rate decisions. $33M on macro questions. Money that is simultaneously in interest rate swaps is also hedging in prediction markets because the swaps don't capture all the relevant risk.
The technology layer: $21M on AI company dominance. Real money on whether NVIDIA or Microsoft is the largest company by June 30th. The uncertainty about AI leadership is so high that people are paying to hedge it.
The pharmaceutical layer: $568,000 on Retatrutide. The potential successor to Ozempic, in a Phase 3 trial that could define the obesity treatment landscape for a generation — and the market is tiny. Undertraded relative to its real-world significance by orders of magnitude.
That gap — between the importance of an outcome and the volume betting on it — is where the edge lives.
The Insider Signal Layer
Here is the thing about prediction markets that most people miss.
The crowd prices uncertainty. But not everyone is equally uncertain.
When the CEO of Aurinia Pharmaceuticals files a Form 4 at 4:03am on June 3rd — disclosing that he bought $12.4 million of his own company's stock across five separate transactions — he is not uncertain. He is certain. He just can't say it publicly.
Our EDGAR monitoring system caught that filing at 4am. Score: 585 out of 600. The highest conviction rating our model generates. CEO plus 10%+ owner plus $12.4M plus five separate purchases plus repeated buying over time. Every factor aligned.
The stock opened up 2% that morning. The options market was essentially dead — $0.10 bid on August $17 calls — because the crowd hadn't figured out what the CEO already knew.
That is the compound signal. The intersection of:
1. What insiders are doing (EDGAR Form 4 data — public, filed with the SEC)
2. What Congress is doing (STOCK Act disclosures — public, filed with the House Clerk)
3. What the crowd is pricing (Prediction market data — public, priced continuously)
When all three streams align, the signal-to-noise ratio is extraordinary. And the tag discovery means we can now find the exact Polymarket market corresponding to each EDGAR signal — matching the insider's conviction against the crowd's price.
The Numbers That Define the Opportunity
Let's make this concrete.
Pharma CEO buys $12.4M → Score 585 → Find FDA tag market pricing approval at 15% YES → Signal: informed conviction vs. crowd skepticism = DIVERGENCE
Defense contractor insider buys $5M → Find Iran/defense market at 18% YES on US military action → Signal: smart money in defense + escalation market underpriced = DIVERGENCE
Congressional Armed Services member buys LMT → Find "Defense bill passes by X date" market at 30% YES → Signal: legislative intelligence + market underpricing = DIVERGENCE
These are not hypothetical. The EDGAR signals fire every day. The markets exist. The volume is real. The only missing piece — until June 8, 2026 — was the connection between them.
What the Tag Count Tells You About the World
6,512 tags. Let that number settle.
That is 6,512 domains of human uncertainty that Polymarket has deemed worth structuring into a market. Sports, politics, economics, science, technology, entertainment, geopolitics, culture.
But here is what the tag list actually tells you about where we are in time:
The tags with the most live volume are geopolitical. Iran. China. Russia. Taiwan. The largest concentration of prediction market capital in June 2026 is in markets pricing the probability of major power conflict. Not recession. Not climate. Not pandemic. War.
The second largest concentration is Fed policy and macro. Inflation. Rate cuts. Recession. The financial system's uncertainty about its own future.
The third is AI. Which company dominates. When AGI arrives. What regulations come.
These three clusters — geopolitical conflict, monetary policy, artificial intelligence — represent the three largest sources of systemic uncertainty in the current moment. Prediction markets, by pricing them with real money, are telling you what the sophisticated world is most afraid of.
Retatrutide at $568K is the anomaly. A drug that could reshape the entire obesity and metabolic disease landscape — affecting hundreds of millions of people — generates less prediction market volume than a single NFL player's contract situation.
That is not irrationality. That is specialization. The people who understand Retatrutide's clinical trial data don't trade prediction markets yet. But the pharma executives do trade their own stock. And Form 4 filings are public.
The gap between the prediction market price and the insider's conviction is measurable. It is real. And it is ours to exploit legally, transparently, and systematically.
The Infrastructure We're Building
Greenwood Financial Intelligence monitors:
- 23,209 historical EDGAR insider signals backtested (H2 2025) — Score 500+: 80% next-day win rate, avg +2.76%; Score 300-499: 69% win rate, avg +3.33%
- 201 live Polymarket markets across 9 GFI themes, updated hourly
- 513 congressional STOCK Act disclosures from the last 45 days
- 6,512 Polymarket tags mapped to financial signal domains
- Hourly price history accumulating since June 5, 2026
The compound signal engine fires when EDGAR insider buying + congressional disclosure + prediction market underpricing all align on the same underlying outcome. The tag discovery means we can now execute that third check with precision.
One API endpoint. $0.010 per query. Pay with USDC on Base mainnet. No subscription. No account. Machine-native.
This is what financial intelligence looks like when you remove the Bloomberg terminal and give AI agents a direct line to the data.
What Comes Next
In 30 days we will publish the first live track record: how did our signals perform against subsequent prediction market movements? Did the FDA markets move after our pharma insider signals? Did the China/Taiwan markets move after defense insider buying?
The answer will either validate or invalidate the compound signal thesis. We believe it will validate it. But we will publish the results either way, because the track record is the product.
Prediction markets are the only honest price discovery mechanism for uncertainty. EDGAR filings are the only honest record of what informed insiders are doing with their own money. Congressional disclosures are the only honest record of what legislators are doing with theirs.
We are connecting these three streams in real time, on a $6/month server, in Grand Forks, North Dakota.
The mirror already exists. We're just helping it see itself more clearly.
Data sourced from SEC EDGAR, House Clerk STOCK Act disclosures, Polymarket public API. Not investment advice.
Want to go deeper?
Get the free research note — insider signals, congressional trades, and prediction market data from the Greenwood platform.