Forem: John Leslie

Google Places API Now Costs $275/Month — Here's a Free Alternative

John Leslie — Mon, 11 May 2026 15:42:02 +0000

Google Places API pricing hit a breaking point in 2026. The basic "Essentials" tier starts at $275/month, and if you need contact details it's $400+. For indie developers, MVPs, and side projects, that's a non-starter.

I built BizData — a free REST API and MCP server that returns business data for any location worldwide. No API key. No signup. No rate limits (yet).

What It Does

Query any city + business category and get structured JSON back: name, address, phone, website, email, coordinates, opening hours. 37 business categories from restaurants to hospitals.

Example — find cafes in Paris:

GET https://bizdata-web.vercel.app/api/businesses?city=paris&category=cafe&limit=5

Returns:

{
  "businesses": [
    {
      "name": "Café de Flore",
      "address": "172 Boulevard Saint-Germain, Paris",
      "phone": "+33 1 45 48 55 26",
      "website": "https://cafedeflore.fr",
      "latitude": 48.8540,
      "longitude": 2.3325
    }
  ],
  "count": 2309,
  "city": "paris",
  "category": "cafe"
}

How It Works

BizData wraps the OpenStreetMap Overpass API and normalizes the messy OSM tag system into a clean REST interface. European cities have 50-70% data completeness. Paris alone has 8,300+ restaurants and 2,300+ cafes.

Also an MCP Server

If you're building with Claude, Cursor, or any MCP-compatible client:

https://bizdata-web.vercel.app/api/mcp

Add it as a remote MCP server (Streamable HTTP transport). Your AI agent can then search businesses in natural language.

When to Use BizData vs Google Places

	BizData	Google Places
Price	Free	$275-400+/month
Auth	None	API key required
Data source	OpenStreetMap (community)	Google's proprietary data
Best for	Europe, Asia, prototypes	US/UK, production apps needing 95%+ coverage
Completeness	50-70% (Europe), 20-33% (US/UK)	90%+ globally

Honest take: If you need comprehensive US data, Google is still better. If you're building an MVP, working with European data, or just don't want to pay $275/month for a side project, BizData works.

Try It

API: https://bizdata-web.vercel.app/api/businesses?city=london&category=restaurant&limit=3
Comparison of 7 alternatives: https://bizdata-web.vercel.app/compare
City explorer: https://bizdata-web.vercel.app/cities/paris

No signup. No API key. Just make a GET request.

37 Supported Categories

restaurant, cafe, bar, hotel, hospital, pharmacy, bank, supermarket, gym, school, university, library, cinema, theatre, museum, park, parking, fuel, car_repair, dentist, veterinary, post_office, police, fire_station, embassy, nightclub, fast_food, bakery, butcher, convenience, clothes, electronics, furniture, hairdresser, beauty, laundry, car_wash

Built this as a solo project. Feedback welcome — especially if you find data gaps in specific cities.

Free Google Places API Alternative: 37 Business Categories, No API Key

John Leslie — Sat, 09 May 2026 15:47:27 +0000

Google Places API removed its free tier in February 2025. The new pricing starts at ~$275/month for 100K calls.

If you need basic business/POI data and don't need Google's reviews or photos, here's a free alternative.

BizData API

A REST API wrapping OpenStreetMap's Overpass API. Free, no API key, no signup.

curl "https://bizdata-web.vercel.app/api/businesses?location=Paris&category=cafe&limit=3"

Returns structured JSON with name, address, phone, website, email, coordinates, and opening hours.

Google Places vs BizData

Feature	Google Places	BizData
Price	$275+/mo	Free
API key	Required	Not needed
Categories	100+	37
Reviews/photos	Yes	No
Coverage (EU)	Excellent	Excellent
Coverage (US)	Excellent	Good
Rate limits	Strict	Fair use

When to use this

Lead generation (bulk business listings)
Market research (how many cafes in a neighborhood?)
Location-based apps needing basic POI data
Prototyping before committing to Google's pricing
MCP tool calls from AI assistants

Endpoints

GET /api/businesses?location=X&category=Y - Find businesses
GET /api/count?location=X&category=Y - Count businesses
GET /api/categories - List all categories

Optional: radius_km (default 5), limit (max 500).

MCP Server

Also works as an MCP server for Claude, Cursor, or any MCP client:

{
  "mcpServers": {
    "bizdata": {
      "type": "streamable-http",
      "url": "https://bizdata-web.vercel.app/api/mcp"
    }
  }
}

Try it

bizdata-web.vercel.app

Data from OpenStreetMap. Built with TypeScript, hosted on Vercel.

I made a free REST API that turns 'find cafes in Paris' into structured JSON

John Leslie — Sat, 09 May 2026 11:25:44 +0000

Every few months I need local business data -- restaurants near an office, gyms in a neighborhood, dentists in a city. The usual path: learn the Overpass query language, handle raw OSM tags, geocode addresses separately, parse nested JSON.

I wanted something simpler: pass a city and category, get clean JSON back. So I built BizData API.

Quick start

curl "https://bizdata-web.vercel.app/api/businesses?location=Paris&category=cafe&limit=2"

{
  "total": 2338,
  "location_resolved": "Paris, France",
  "businesses": [
    {
      "name": "Cafe Exemple",
      "category": "cafe",
      "address": "Avenue Parmentier, 28",
      "phone": "+33 1 48 05 94 36",
      "website": "https://aunreve.fr",
      "email": "contact@aunreve.fr",
      "lat": 48.8606,
      "lon": 2.3787,
      "opening_hours": "We-Sa 08:30-19:00; Su 09:00-19:00"
    }
  ]
}

No API key. No signup. One URL.

What it does that raw Overpass doesn't

Overpass is powerful but requires learning a custom query language:

[out:json][timeout:30];
node["amenity"="cafe"](around:5000,48.8566,2.3522);
out center tags;

BizData handles the geocoding, query building, tag mapping (37 normalized categories), and result formatting. You get a standard REST endpoint with clean JSON instead of a custom QL.

Three endpoints

GET /api/businesses -- Find businesses by location and category.

Param	Required	Description
location	Yes	City or address
category	Yes	One of 37 types
radius_km	No	Search radius, default 5
limit	No	Max results, default 50, max 500

GET /api/count -- Count businesses across categories. "How many restaurants, cafes, and bars in London?" returns 6,763.

GET /api/categories -- List all 37 available categories.

Error handling

Bad category:

{"error": "Unknown category: pizza. Available: accountant, bakery, bank, ..."}

Missing params:

{"error": "Missing required parameter: location"}

Performance

/api/categories: ~50ms (cached)
/api/businesses (50 results): 1-3 seconds
/api/count (all categories): 10-30 seconds
Results cached for 1 hour on Vercel's edge

Limitations

Data quality depends on OpenStreetMap contributors
Coverage: excellent in Europe, solid in major US/Asian cities, sparse in rural areas
Shared Overpass API pool
Some businesses lack phone/email/hours

Built with

TypeScript, Vercel serverless functions, OpenStreetMap Nominatim (geocoding) + Overpass API (data).

Try it

Interactive demo: bizdata-web.vercel.app

All 37 categories: accountant, bakery, bank, bar, beauty, bookstore, cafe, car_dealer, car_repair, cinema, clothing, coworking, dentist, doctor, electronics, florist, furniture, gallery, gas_station, guest_house, gym, hairdresser, hospital, hostel, hotel, insurance, lawyer, museum, parking, pet_shop, pharmacy, real_estate, restaurant, school, supermarket, theatre, university.

How We Made 61% on Hormuz in 9 Days

John Leslie — Wed, 29 Apr 2026 11:17:44 +0000

Our Hormuz April NO position is worth $803 on a $500 bet. The market closes tomorrow at 0.4%. Nine days ago, it was at 38%.

The trade was not a gamble. It was arithmetic.

Why the Market Was Wrong

On April 20, Hormuz April YES was trading at 38%. The market gave a one-in-three chance that Strait traffic would normalize before month-end.

The ceasefire had just expired (April 19). Iran closed it with zero diplomatic progress.
Iran had fired on ships during the truce (April 18).
Only 16 ships per day were transiting, down from the normal 60+.
There was no active diplomatic track.

Even if talks started immediately, normalization requires Iran to withdraw naval assets, commercial insurers to resume coverage, and shipping lines to re-route vessels. Ten days was not enough.

Result: Hormuz April NO at $0.62 to $0.996. Return: +61% in 9 days.

The Portfolio

Fictional $1,000 portfolio (started April 20): now worth $1,637 (+63.7%).

Position	Entry	Current	Return
Hormuz April NO	$0.62	$0.996	+61%
Hormuz May NO	$0.305	$0.645	+111%
Cash	$200	$200	--

What Needs to Happen for May

Hormuz May is at 35.5% YES. For normalization by May 31, all of these must happen in 32 days:

A new ceasefire agreement (none active).
Iran withdraws its dual naval blockade.
Commercial shipping resumes normal volumes.
Traffic returns to 60+ ships/day (currently around 5).

Iran has said reopening is impossible while sanctions persist.

June

Hormuz June is at 58.5%. The uncertainty is genuinely high and our edge is not clear. We are not touching June.

Knowing when you do not have edge is as important as recognizing when you do.

Full analysis: predictionoracle.io/issue-9

Subscribe to The Market Oracle

Polymarket Wants Back Into the US. Here's What Changes If They Get In.

John Leslie — Wed, 29 Apr 2026 07:35:43 +0000

Polymarket just asked the CFTC for permission to bring its main exchange back to American traders. On the same day, it ripped out its entire trading infrastructure and rebuilt it from scratch. Kalshi posted $3.91 billion in weekly volume, its highest ever. The Senate introduced a bipartisan resolution to ban all lawmakers from prediction markets. And Finland is the betting favorite to win Eurovision.

Polymarket Wants Back Into the US

This is the biggest story in prediction markets this year, and it happened almost quietly.

On April 28, Polymarket began formal discussions with the CFTC to lift the ban that has kept its main exchange off-limits to American users since 2022. That year, the CFTC fined Polymarket $1.4 million for operating without proper registration. Polymarket settled, pulled out of the US, and spent the next four years building the largest prediction market in the world from outside America's borders.

The timing isn't accidental. Last July, Polymarket acquired QCEX, a CFTC-licensed exchange, for $112 million. That gave them a regulatory vehicle. Now they want to use it.

If the CFTC approves, the effect is simple: American retail liquidity floods back into Polymarket. The US has more prediction market traders than any other country. Polymarket has been serving them through a wink-and-nod VPN culture, but institutional capital and mainstream retail won't touch an exchange where US participation is technically prohibited. Remove that prohibition, and market depth on major events could double or triple.

The Platform Overhaul Nobody Noticed

On the same day Polymarket filed with the CFTC, it executed a complete infrastructure upgrade. New smart contracts. A rewritten order book engine. A new collateral token called Polymarket USD (pUSD), backed 1:1 by USDC on Polygon. Trading paused for roughly an hour during the cutover.

This is the kind of thing that gets zero media coverage but matters enormously. The old infrastructure was built when Polymarket processed a few hundred thousand dollars a day. Now it handles hundreds of millions.

Kalshi Just Had Its Best Week Ever

Kalshi posted $3.91 billion in weekly volume for the week of April 20-28, its highest ever. That's nearly double Polymarket's weekly volume for the same period. Sports parlays drove more than 85% of the total.

Kalshi isn't competing with Polymarket for political prediction nerds. It's competing with DraftKings and FanDuel for sports bettors. The prediction market wrapper lets Kalshi offer sports parlays with lower vig than traditional sportsbooks.

The Senate Wants Lawmakers Out

Sen. Bernie Moreno (R-OH) introduced a bipartisan resolution on April 24 that would ban all members of Congress from trading on prediction markets entirely. This is separate from and broader than the PREDICT Act. Five co-sponsors from both parties.

Paradoxically bullish for the industry. Every one of these bills implicitly accepts that prediction markets are legitimate financial instruments worth regulating. You don't ban members of Congress from doing something that's about to be shut down.

Eurovision 2026 Odds

Finland leads at 35.3% on Polymarket. Interesting discrepancy: Denmark is 13.4% on Polymarket but only 7% at bookmakers. Israel is the reverse: 5.1% on Polymarket versus 11% at bookmakers. When the same event is priced differently on two platforms, someone is leaving money on the table.

Our Portfolio: +63.7%

Starting capital $1,000 on April 20. Current value: $1,637. Hormuz April NO position resolves tomorrow at +61%. Hormuz May NO up 111%.

Full issue with charts and analysis: predictionoracle.io/issue-8.html

The Market Oracle is a weekly prediction market intelligence newsletter. Subscribe at predictionoracle.io

The Cop Left the Building: CFTC Shrinks 24% While Prediction Markets Hit $60B

John Leslie — Tue, 28 Apr 2026 07:24:50 +0000

Issue #7 · April 27, 2026 · By John Leslie

The agency responsible for policing prediction markets has lost a quarter of its workforce. Meanwhile, the industry just crossed $60 billion in trades this year. Brazil banned every prediction market platform overnight. And the economist who invented the theoretical framework behind prediction markets argues that the insider trading everyone is panicking about is actually the entire point. Welcome to the most confusing week in the industry's short history.

The Cop Left the Building

The Commodity Futures Trading Commission -- the federal agency that regulates prediction markets -- has shrunk to its smallest size in 15 years. Workforce is down 24% since January 2025. That's not attrition. That's a policy choice.

This is happening precisely as the agency's jurisdiction has become the most contested terrain in financial regulation. More than two dozen state-level lawsuits allege that Polymarket and Kalshi are unlicensed gambling operations. Three states tried to ban the platforms outright last week; the DOJ sued all three, arguing federal preemption. The CFTC is supposed to referee this fight. It now has fewer referees than at any point since 2011.

The volume numbers make the staffing collapse almost comical. Kalshi and Polymarket have processed $60 billion in trades so far in 2026. In March alone, Kalshi handled $13 billion and Polymarket roughly $10 billion. For context, the entire CFTC-regulated prediction market had less than $1 billion in annual volume as recently as 2023. We're talking about a 60x increase in three years with a 24% decrease in oversight capacity.

Our take: Underfunding the regulator is short-term bullish and long-term dangerous. Right now, it means platforms operate with minimal oversight -- good for innovation, bad for legitimacy. The risk is that a major scandal forces Congress to overcorrect.

Brazil Bans Everything

Brazil became the first major economy to ban prediction market platforms entirely. As of today, Polymarket, Kalshi, and 27 other platforms are blocked in the country. The Brazilian government classified all of them as illegal betting operations.

This isn't a small market. Brazil has the fifth-largest internet population globally and a massive sports betting culture. Polymarket had meaningful traffic from Brazilian users. That's gone overnight.

Our take: Brazil's ban is a setback for global market depth but reinforces the US as the center of gravity. If you're building in this space, your regulatory strategy is now a US-first strategy by default.

Robin Hanson: Insider Trading Is the Point

Fortune published a remarkable interview this week with Robin Hanson, the George Mason economist whose 1990s work essentially invented the theoretical framework for prediction markets. His argument: the insider trading that everyone is panicking about is exactly what makes prediction markets valuable.

The logic: prediction markets aggregate dispersed information into accurate prices. When a soldier bets on a military operation he has classified knowledge about, that information moves the price closer to truth. The market becomes more accurate, not less.

Our take: Hanson is directionally right and practically wrong. In a frictionless theory, insider information makes markets more efficient. In the real world, it creates perverse incentives. The right answer is the one Kalshi is already implementing: ban the trades that create bad incentives (politicians on their own races), allow the rest.

Hormuz: "Reopening Is Impossible"

Iran's parliament speaker said this week that "reopening the Strait of Hormuz is impossible as long as the US blockade is in place." Ship traffic remains at roughly 5 vessels per day versus 140 pre-crisis. Bloomberg described it as "maritime trench warfare."

Current prices:

April normalization: 0.4% YES -- locked in
May normalization: 36.5% YES -- well below our entry at 69.5%
June normalization: 55.5% YES -- still overpriced in our view

Original Research: How Calibrated Are These Markets?

We built an interactive calibration tracker analyzing 7,661 resolved Polymarket markets. Key findings:

Excellent at the extremes. Events priced at 0-5% almost never happen (0.1% actual). Events at 95-100% always happen.
Systematic underpricing in the middle. Markets at 65-75% resolved YES 96.5% of the time.
Overall Brier score of 0.025 -- roughly 10x better than random guessing.

Explore the data yourself: polymarket-calibration.vercel.app

Portfolio Update: +62.7%

Position	Entry	Current	P&L
Hormuz April NO	YES 38%	YES 0.4%	+60.6%
Hormuz May NO	YES 69.5%	YES 36.5%	+108.1%
Cash reserve	$200	$200	--
Total	$1,000	$1,627	+62.7%

This is The Market Oracle, a weekly prediction market intelligence newsletter. Read more at https://site-two-nu-51.vercel.app

Are Prediction Markets Well-Calibrated? I Analyzed 7,661 Resolved Polymarket Markets

John Leslie — Sat, 25 Apr 2026 15:33:05 +0000

By John Leslie

Prediction markets are booming. Polymarket alone processes billions in volume. But a fundamental question remains: when a market says there's a 60% chance of something happening, does that event actually happen 60% of the time?

I pulled 7,661 resolved binary markets from Polymarket's API, grabbed price history for the top 2,000 by volume, and built an interactive calibration tracker to find out.

Try it here: Polymarket Calibration Tracker

What is calibration?

A forecaster is well-calibrated if their predicted probabilities match observed frequencies. If you say "70% chance" for 100 different events, roughly 70 of them should happen. Perfect calibration means every point sits on the diagonal line where predicted = actual.

This is the gold standard metric for forecast quality, used everywhere from weather forecasting (the NWS publishes calibration curves) to machine learning model evaluation.

The results

Markets are excellent at the extremes

Events priced at 0-5% almost never happen (0.1% actual resolution rate). Events at 95-100% always happen (100% actual). This is reassuring: when the crowd is very confident, the crowd is right.

Systematic underpricing in the middle range

This is where it gets interesting. Markets trading at 65-75% actually resolved YES 96.5% of the time (29 markets). At 45-55%, the actual rate was 60% (25 markets). The sample sizes are small in these middle bins, so interpret cautiously, but the pattern is consistent: mid-range markets seem to underprice the YES outcome.

One hypothesis: multi-outcome markets inflate the low-probability bins. If there are 10 candidates in an election, 9 of them trade at 0-5% and lose, pulling the average down. The winners cluster higher. This creates a distributional skew that looks like underpricing in the middle.

Calibration improves closer to resolution

This makes intuitive sense. The 24-hour Brier score (0.025) is significantly better than the 30-day score (0.042). As resolution approaches, uncertainty resolves, and prices converge to 0 or 1.

For context: random guessing on balanced outcomes gives a Brier score of 0.25. A Brier score of 0.025 is excellent.

Category breakdown

The tool lets you filter by category. Some highlights:

Politics (1,312 markets): Well-calibrated at extremes, biggest mid-range deviation
Sports (1,314 markets): Similar pattern to politics, slightly tighter calibration
Crypto (752 markets): More volatile, noisier calibration curve
Geopolitics (404 markets): Small sample but interesting patterns
Science/Tech (163 markets): Too few markets for reliable conclusions

Technical details

Data collection: Polymarket's Gamma API for resolved market metadata. CLOB (Central Limit Order Book) midpoint prices at 24h, 7d, and 30d before resolution for the top 2,000 markets by volume.

Methodology: Markets binned into probability ranges (0-5%, 5-15%, ..., 95-100%). For each bin, the actual resolution rate is compared to the predicted probability. Proportional dot sizes indicate sample count.

Caveats:

75% of price-tracked markets fall in the 0-5% bin, driven by multi-outcome market losers
Small sample sizes in mid-range bins (20-50 markets each) limit statistical significance
Only resolved markets are included (survivorship consideration)

Why this matters

Prediction markets are being used for increasingly high-stakes decisions. DARPA studied them for intelligence forecasting. The Federal Reserve has explored them for inflation expectations. Companies like Google and HP have used internal prediction markets for product planning.

If these markets are well-calibrated, their prices can be treated as genuine probability estimates. If they're systematically biased, traders and decision-makers need to adjust.

The data suggests Polymarket is well-calibrated overall (Brier score 0.025), with some interesting biases in the middle range worth investigating further as sample sizes grow.

Explore the data yourself: polymarket-calibration.vercel.app

I write The Market Oracle, a weekly prediction market intelligence newsletter. Read more at our site.

The Regulatory Storm: 7 Bills, 3 Bans, and the Future of Prediction Markets

John Leslie — Sat, 25 Apr 2026 07:46:46 +0000

Originally published at The Market Oracle

Prediction markets just became Washington's hottest regulatory target. In a single week: seven congressional bills introduced targeting prediction market oversight, three state-level enforcement actions, Kalshi's first-ever self-enforcement against politicians betting on their own races, and the DOJ's first insider trading prosecution in the space.

The Regulatory Blitz

Seven bills. One week. That's not a regulatory drip -- that's a flood.

The headline legislation is the bipartisan PREDICT Act, which would ban politicians and their immediate family members from trading on political events.

New York Governor Hochul issued an executive order banning all state employees from prediction market insider trading. Connecticut, Arizona, and Illinois attempted outright bans. The federal government simultaneously sued all three states to block their actions -- arguing prediction markets fall under CFTC jurisdiction.

Our take: Regulation is coming, but it's bullish long-term. Clear rules mean institutional adoption.

Kalshi Becomes the Sheriff

Kalshi fined and suspended three congressional candidates for placing wagers on their own races -- the first enforcement action of its kind. Fines ranged from \$539 to \$6,229.

The fines are small. The signal is enormous. Self-policing is the cheapest insurance against existential regulatory risk.

Hormuz: The Dual Blockade Tightens

Five ships transited in the last 24 hours. The pre-crisis average was 140 per day. Oil at \$105/barrel.

April normalization at 1.7% -- effectively resolved NO, exactly as we called when it was at 38%. Our portfolio: +61.2%.

Portfolio: +61.2%

Position	Entry	Current	P&L
Hormuz April NO	YES 38%	YES 1.7%	+58.6%
Hormuz May NO	YES 69.5%	YES 37%	+106.4%
Total	\$1,000	\$1,612	+61.2%

Read the full issue | All issues

When Prediction Markets Become the News: Shoot-and-Kill Orders, Fox News Deals, and John Oliver

John Leslie — Fri, 24 Apr 2026 11:18:25 +0000

Trump just ordered the Navy to shoot Iranian mining boats in the Hormuz Strait. Kalshi is now embedded in Fox News, CNN, and CNBC. John Oliver devoted a segment to prediction markets. Forbes created a market on a children's mass shooting. And our fictional portfolio just crossed +63.6% in 4 days.

This week, prediction markets stopped being a niche curiosity and became the news itself.

"Shoot and Kill"

On April 23, President Trump ordered the US Navy to "shoot and kill" any boats caught laying sea mines in the Strait of Hormuz. The order came after Iran seized two more ships and fired RPGs at a Greek-owned cargo vessel.

The Strait remains effectively closed. The US naval blockade has turned back 31 Iran-linked vessels since April 13. Iran says reopening is impossible while the blockade continues. Oil is at $95/barrel.

Portfolio impact: When we opened our May NO position on April 20, YES was at 69.5%. It's now at 33.5%. Our position is up 117.9% on that leg alone. The "shoot and kill" order is not the language of de-escalation. We're revising our May fair value estimate down to 15-20%. The position stays open.

Prediction Markets Go Primetime

Kalshi has signed partnerships with CNBC, CNN, Fox News, and the AP. Fox Corporation announced it will weave real-time prediction data into Fox News, Fox Business, Fox Weather, and streaming. When you watch cable news, you now see prediction market odds alongside stock tickers.

Kate Knibbs at Wired established herself as a dedicated prediction markets reporter. Polymarket partnered with Substack (Feb) and Dow Jones (Jan). The infrastructure of mainstream coverage is being built in real time.

John Oliver devoted a Last Week Tonight segment to prediction markets, questioning the industry's rapid expansion and regulatory gaps. When a comedian covers your industry, you've gone mainstream.

Where's the Line?

Not everyone is celebrating.

Forbes created a prediction market that gamified a mass shooting that killed eight children. The backlash was immediate (reported by 404 Media, April 20). The dark side of "a prediction market for everything": some events should not be betting opportunities.

ProPublica added a section to its ethics code prohibiting staff from betting on news events through prediction markets. The concern: when journalists can bet on outcomes they cover, the language of forecasting begins to influence the language of reporting.

The Columbia Journalism Review published a deep analysis warning that the CFTC's pro-prediction-market posture may not survive a genuine crisis of public trust.

Portfolio: +63.6% in Four Days

Position	Entry	Current	P&L
Hormuz April NO	YES 38%	YES 2.9%	+56.5%
Hormuz May NO	YES 69.5%	YES 33.5%	+117.9%
Cash reserve	$200	$200	--
Total	$1,000	$1,636	+63.6%

We called April at 38% -- it's now 2.9%. We called May overpriced at 69.5% -- it's now 33.5%. Both calls directionally correct. The "shoot and kill" order strengthens our conviction.

Market Snapshot

Hormuz May normalization: 33.5% YES (was 69.5% four days ago)
Hormuz June normalization: 57.5% YES
2028 Presidential: JD Vance 18.4%, Gavin Newsom 17.1%, Marco Rubio 10.2%
Democrats win House 2026: 84.5% YES
Eurovision 2026: Finland 36.4%

Read the full analysis with more detail at The Market Oracle. Published by John Leslie.

The Soldier Who Bet on a Secret War: First DOJ Prediction Market Prosecution

John Leslie — Fri, 24 Apr 2026 07:47:13 +0000

A US Special Forces soldier used classified intelligence about a covert military operation to make $410,000 on Polymarket. The DOJ just filed the first-ever insider trading prosecution in prediction market history.

The Big Story: A Soldier Bet on a Secret War

On April 23, the Department of Justice charged Master Sergeant Gannon Ken Van Dyke, a US Special Forces soldier stationed at Fort Bragg, with using classified military intelligence to profit on Polymarket.

What he knew: Van Dyke participated in planning "Operation Absolute Resolve," the covert US operation to capture Venezuelan president Nicolas Maduro. He had advance knowledge of the timing and details of the raid.

What he did: He placed 13 bets totaling $33,000 on Polymarket markets tied to Maduro's capture and removal from power. When the operation succeeded, he collected $410,000 in profit. A 1,142% return on classified intelligence.

The charges: Three counts of violating the Commodity Exchange Act (up to 10 years each), wire fraud (up to 20 years), and unlawful monetary transaction (up to 10 years). First-ever DOJ prosecution for prediction market insider trading.

The Integrity Week

Kalshi suspended three congressional candidates for betting on their own races
Illinois Governor barred state employees from using insider info on prediction platforms
The White House warned staff against placing prediction market bets
A Harvard/Columbia paper identified $143 million in suspicious Polymarket profits

Prediction Markets Become Derivatives Exchanges

Polymarket launched perpetual futures (10x leverage) on April 21. Kalshi launching crypto perps April 27. Both competing with Coinbase and Robinhood now.

Kalshi leads in volume ($37.5B YTD vs Polymarket $29.2B) and valuation ($22B vs $15B target).

Hormuz Portfolio: +61.7%

Metric	Value
Starting capital	$1,000
Current value	$1,617
Return	+61.7%
Days held	4

May Hormuz dropped from 69.5% to 35.5%. April at 2.9%, effectively resolved.

The Market Oracle - weekly prediction market intelligence. Read more

Kalshi at $22B, Polymarket Raises $400M, and Our Portfolio Is Up 33.6%

John Leslie — Fri, 24 Apr 2026 07:37:01 +0000

Issue #2 · April 22, 2026 · By John Leslie

This week: Kalshi is valued at $22 billion and heading to the Supreme Court. Polymarket is raising $400M with the NYSE parent company backing it. Meanwhile, IRGC gunboats are seizing ships in the Strait of Hormuz and our fictional portfolio is up 33.6% in two days. The money is flooding in and the stakes keep rising.

The Big Story: The $37 Billion Week

Three developments this week prove prediction markets have crossed from crypto curiosity to financial infrastructure:

1. Kalshi raised $1B+ at a $22 billion valuation

Coatue Management led the round. Kalshi controls 89% of the regulated U.S. market (Bank of America, April 9). At $22B, Kalshi is now valued higher than the Chicago Board Options Exchange was at its 2005 IPO. This is not a bet on prediction markets. It is a bet that event contracts become a permanent asset class.

2. Polymarket is raising $400M at $15 billion

Intercontinental Exchange (ICE), the parent company of the New York Stock Exchange, has committed up to $2 billion. When the NYSE parent company writes a $2B check into prediction markets, the institutional legitimacy question is settled. The remaining question is who wins: regulated (Kalshi at $22B) or crypto-native (Polymarket at $15B)?

3. The Supreme Court is next

Kalshi's fight with states over whether prediction markets are derivatives or gambling is heading toward the Supreme Court. The outcome determines whether the industry operates under one federal framework (CFTC) or 50 state gaming commissions. Sports betting is currently 85% of volume for both major platforms. If the Court rules prediction markets are gambling, those contracts vanish overnight. If it rules they are derivatives, the CFTC becomes the sole regulator and institutional capital pours in.

What this means: The prediction market industry is expected to hit $200B in volume this year. Bernstein projects $1 trillion by 2030. The infrastructure war between Kalshi and Polymarket will define market structure for the next decade. For traders: enjoy the promotional pricing and tight spreads. This is two platforms spending billions to buy your order flow.

The Markets

Hormuz: IRGC Seizes Two More Vessels

Our call from Issue 1: April normalization at 38% was overpriced (fair value 5-10%), May at 69% was overpriced (fair value 40-50%).

Current prices: April 9% YES, May 59% YES. Both moved sharply in our direction.

The week timeline:

April 17: Iran declares the strait completely open. Oil drops 10%. Markets surge.

April 18: Iran reverses course. IRGC gunboats fire on two Indian-flagged ships. India summons the Iranian ambassador. Iran claims communication gap.

April 19: US Navy seizes Iranian cargo ship Touska after 6-hour standoff. Marines rappel from helicopters. Tehran vows retaliation. Iran parliament drafts a law requiring hostile country ships to get approval and pay tolls.

April 20: Zero tankers cross the strait. One of the quietest days since the crisis began.

April 21: Trump says he will not extend ceasefire. Warns of lots of bombs. US forces board oil tanker M/T Tifani in the Indian Ocean.

April 22 AM: Trump reverses, announces open-ended ceasefire extension, but the blockade stays fully in place. Iran FM Araghchi calls the blockade an act of war and a violation of the ceasefire.

April 22 PM: IRGC seizes two more vessels in the Strait of Hormuz. Negotiations deadlocked. About 800 vessels queued in the Gulf.

Updated call: April is effectively resolved as NO. May at 59% is still overpriced. Revised estimate: 30-40%. The blockade is becoming structural, not tactical. Iran is institutionalizing it with transit fees. The negotiating gap is massive (US wants 20-year enrichment pause, Iran offered 5 years). 3-8 ships per day are transiting vs. 60 needed. This does not unwind in 5 weeks.

Iran Regime Change: The Cheap Hedge

Will the Iranian regime fall by April 30 has $32M in volume despite trading at just 2% YES. Why does a 2% market have $32M?

Traders are using it as a cheap hedge. At 2 cents per share, a YES position is a lottery ticket. If Mojtaba Khamenei (installed as Supreme Leader on March 8 after his father's assassination) faces a coup, that 2-cent share becomes worth $1.

The longer-dated markets:

By May 31: 3%
By June 30: 7.5%
By end of 2026: 20%

The market says there is a 1-in-5 chance of regime change within 9 months.

2028 Presidential: Democrats Still Elevated

$548M in volume. Current top candidates:

JD Vance: 18.8%
Gavin Newsom: 17.8%
Marco Rubio: 10.2%
Alexandria Ocasio-Cortez: 5.9%
Kamala Harris: 4.1%

Democrats remain elevated, driven by backlash to Iran. Historical base rates suggest regression toward 50/50 this far from an election. I would not trade this either direction until the Hormuz situation resolves.

Eurovision: Finland Still Leads

Finland leads the overall winner market at 36%. Finland is the consensus pick but is not dominant in either the jury (13%) or televote (17%) sub-markets. Israel leads the televote at 38%. At 36%, Finland is probably fairly priced. The value, if any, is in Israel for the televote.

The Feature: Prediction Markets vs. The National Weather Service

The most underrated story in prediction markets right now.

ForecastEx has been offering daily high-temperature markets via Interactive Brokers ForecastTrader since November 2025. Patrick Brown, head of climate analytics at Interactive Brokers, found that prediction markets were more accurate than the National Weather Service.

Why would traders beat meteorologists with billion-dollar supercomputers?

They are not beating them. They are aggregating them. A prediction market on tomorrow's high temperature in Chicago does not replace the weather models. It combines them. Traders who follow the European model, the GFS model, local radar, and even look out the window all contribute to the price. The result is an ensemble of all available information, weighted by confidence.

This is the prediction market thesis in its purest form: markets as information aggregators. Not better than any individual expert, but better at combining all available expertise.

If markets can improve weather forecasting, they can improve forecasting of economic indicators, disease outbreaks, and supply chain disruptions. The category is about to get much bigger.

The Regulatory Corner

Federal wins:

CFTC enforcement advisory issued, signaling intent to regulate rather than kill the space
Third Circuit ruled prediction markets cannot be regulated by state gambling laws
New Jersey lost its appeal against Kalshi on sports contracts

State pushback:

Nevada issued a temporary ban on Kalshi
Arizona accused the platform of unlicensed gambling
Massachusetts secured preliminary injunction
40+ lawmakers led by Sen. Warren pushing for insider trading rules

The insider trading problem:

A trader made $400K correctly predicting the Maduro ouster in January
Another made about $300K on Biden pardon timing
A political candidate was caught trading on his own candidacy (fined $2,246, 5-year ban from Kalshi)

My read: federal primacy wins. The CFTC is suing states, which means they want jurisdiction. Kalshi regulatory moat deepens with each court victory. Long-term bullish for the industry. Short-term, expect volatility around rulings.

Track Record

Market	Issue	Call	Entry	Current	Status
Hormuz April	1	NO at 38%	38%	9%	Virtually resolved. +46.7%
Hormuz May	1	NO at 69%	69%	59%	Tracking right. +34.3%

Fictional $1,000 portfolio (started April 20): now worth $1,336 (+33.6% in 2 days). Two calls, both correct direction.

This is The Market Oracle, a weekly prediction market intelligence newsletter. Read more at https://site-two-nu-51.vercel.app

The Hair Dryer That Won $34,000: This Week in Prediction Markets

John Leslie — Fri, 24 Apr 2026 07:28:09 +0000

Issue #3 · April 23, 2026 · By John Leslie

Prediction markets just had their biggest integrity week ever. Kalshi suspended three congressional candidates for betting on their own races. Someone in Paris used a hair dryer to manipulate a Polymarket weather contract for $34,000. And in the Strait of Hormuz, our fictional portfolio just crossed +53.5% in three days. This issue: what these scandals mean for the industry, why they're actually bullish, and the markets that matter this week.

The Big Story: Politicians Caught Betting on Themselves

Kalshi dropped a bombshell on Tuesday: three congressional candidates were suspended for placing bets on their own political outcomes. This is the first time a U.S. prediction market has enforced "insider trading" rules against politicians.

The three cases:

Matthew Klein (D-MN), a state senator running for an open U.S. House seat, bet $50 that he'd win his own nomination back in October. He cooperated with Kalshi's investigation and was fined $530 plus a five-year ban.

Mark Moran (I-VA), running for U.S. Senate, bet that he would announce a political run, then publicly launched his campaign in January. He refused to cooperate and was fined $6,200 plus a five-year ban.

A Texas Republican who finished with 1% in a March primary bet on their own House race. Fined $780, five-year ban.

Why this matters: The bets were tiny ($50-$200 range). The penalties were modest. But the signal is enormous: Kalshi is building an enforcement infrastructure before regulators force one on them. Every prediction market exchange is watching. Polymarket has no comparable system.

The bigger picture: Congress has been scrutinizing prediction markets since the 2024 election. Multiple lawmakers have called for stricter regulation. By policing itself now, Kalshi is attempting to preempt regulation by demonstrating the industry can self-govern. Whether regulators buy that argument will shape the next decade of prediction markets.

What this means: This is exactly the kind of industry maturation that happens before mainstream adoption. Credit card companies went through it. Crypto exchanges went through it. The fact that prediction markets are having their "insider trading" moment means they're being taken seriously enough to need rules. Long-term, this is bullish for the industry.

The Hair Dryer That Won $34,000

This is the wildest prediction market story of the year.

Polymarket runs temperature contracts on Paris weather, settled using an automated sensor at Charles de Gaulle Airport. On April 6, someone walked up to the sensor and apparently aimed a hair dryer at it.

The result: The sensor recorded a 4C spike in 12 minutes, briefly hitting 22.5C before returning to normal. Neighboring stations showed nothing unusual. Polymarket paid out $14,000 on temperature contracts that resolved because of the spike.

It happened again. On April 15, the same sensor recorded another anomalous rise to 22C under calm, cloudy conditions. Another $20,000 in payouts.

Total haul: ~$34,000. Meteo France has filed criminal charges. Polymarket has since switched its Paris temperature source from CDG to the Paris-Le Bourget station.

Why this matters for bettors: Physical oracle manipulation is the prediction market equivalent of match-fixing. If a contract settles based on a single data point, someone will eventually try to manipulate that data point. Weather contracts are uniquely vulnerable because the sensors are physically accessible and automated. Before placing large positions: always check the resolution source. Multi-station averaging or satellite data would eliminate this attack vector, but neither platform has announced plans to switch.

Hormuz Tracker: Our Best Week Yet

Our flagship call continues to pay off. The fictional portfolio crossed +53.5% in three days.

Current prices (April 23, 09:20 CEST):

April normalization: 4.5% YES (down from 38% when we first called it). Effectively resolved as NO.
May normalization: 42.5% YES (down from 69.5% at our entry). Crashed 16.5 percentage points overnight.
June normalization: 66.5% YES (down from 88.5% at our first mention).

What happened this week: Trump extended the ceasefire indefinitely but the blockade stays. The IRGC seized two more container ships in the strait on April 22, hours after the ceasefire extension. Iran is institutionalizing control of the waterway, charging tolls exceeding $1 million per ship and seizing vessels that attempt unauthorized passage.

Updated call: May at 42.5% is now much closer to fair value. The structural thesis: the IRGC has transformed the blockade from a wartime tactic into an ongoing revenue and leverage operation. Reopening the strait requires either (a) a comprehensive deal that gives Iran something it values more than strait control, or (b) military action to clear IRGC naval assets. Neither appears imminent. Revised estimate for May: 25-35%.

Portfolio Update

Metric	Value
Starting capital (Apr 20)	$1,000
Current value	$1,535
Return	+53.5%
Best position	May NO: +88.4%
Days held	3

Markets Worth Watching

AI Model Race: Anthropic 94%, OpenAI 6%

Monthly Polymarket contract on best AI model. Claude Opus 4.7 launched April 16 and topped the Artificial Analysis index. At 94%, this feels like it could only surprise to the downside. OpenAI hasn't shipped a frontier model update in months. The 6% is priced as a long-shot lottery ticket, and honestly that's about right.

FIFA World Cup: $668M in Volume

The largest sports prediction market ever on Polymarket. Tournament starts June 11. Spain (16%), France (15.8%), England (10.9%). Wide open. No team has repeated since Brazil in 1962.

Supreme Court Watch

Circuit split forming between the Third Circuit (pro-Kalshi) and a pending Ninth Circuit case (potentially anti-Kalshi). If the Ninth Circuit rules against prediction markets, Supreme Court review becomes almost certain by 2027. This is the long-term regulatory risk that could reshape the entire industry.

Track Record

Call	Issue	Direction	Entry	Current	Move	Status
Hormuz Apr is a NO	#1	NO at 38%	62c	95.5c	+33.5c	Effectively won
Hormuz May overpriced	#1	NO at 69.5%	30.5c	57.5c	+27c	Strongly winning
Hormuz Jun overpriced	#1	Observation	88.5% YES	66.5% YES	-22pp	Tracking right

Three calls, three correct directions. Fictional portfolio +53.5% in 3 days. Small sample size, but the methodology is working: identify structural mispricing in geopolitical markets where the crowd underestimates inertia.

This is The Market Oracle, a weekly prediction market intelligence newsletter. Read more at https://site-two-nu-51.vercel.app