Predictbook

The $MSTR “sells any Bitcoin by May 31, 2026” market is the perfect example of how prediction markets are not just about predicting reality, but about understanding contracts, precedent, and legal‑style interpretation. $MSTR sold 32 $BTC sometime between May 26–31 and disclosed this in a June 1 Form 8‑K. The filing itself explicitly states that 32 $BTC were sold during the May 26–31 window, with proceeds of about $2.5M at an average sale price of around $77k per coin. So in the real world, the answer to “Did they sell any $BTC by May 31?” is clearly “yes.” The entire dispute is about whether that truth is allowed to matter, given the wording of the Polymarket contract and the timing of the disclosure. The rule text says the market resolves Yes “if MicroStrategy sells any of its Bitcoin by 11:59 PM ET on the date specified in the title.” Strategy’s 8‑K shows that sale activity did happen in that exact window (May 26–31), and the data is presented “as of May 31, 2026, 4:00 p.m. Eastern Time.” Under this reading, the event, a sale in that time window, controls. Disclosure is just how we learn about it later. If the oracle’s job is “resolve to truth,” then the market “should” be Yes, because the world state we asked about is, in fact, true. Basically, if you asked, “Did it rain yesterday?” you wouldn’t be wrong just because the weather report only comes out the next morning. However, Polymarket has repeatedly clarified in previous MicroStrategy markets that only information available within the market’s timeframe counts, and that confirmations arriving afterward must be ignored unless a waiting clause is explicitly written in. In the current market, Polymarket has now explicitly stated that “confirmation achieved outside of the market’s time frame does not qualify,” and that no MSTR filings, on‑chain data, or credible reporting during the window confirmed a sale. So, from this point of view, the contract isn’t “Did they ever sell in that period?”; it’s “Can the oracle confirm, using allowed sources, within the timeframe?” If the proof arrives too late, it’s legally irrelevant, even if it’s factually correct. If this market were allowed to flip to Yes purely because we later learned the sale occurred, all past markets with similar wording and similar timing problems would become unstable. If today’s market is treated as “truth at any time,” then logically you can ask: - Should old markets be reopened and re‑resolved because new filings surfaced days or weeks later? - How far do you go back? 1 day, 2 days, a month, a year? - And if you do that, are you still running a market, or are you retroactively editing a ledger of bets? Re‑opening old resolutions whenever new information emerges would be a disaster for trader confidence. Many geopolitical and corporate markets depend on later investigations, FOIA releases, or revised statistics. If any of those could retroactively flip a resolved contract, nobody could treat winnings or losses as final. Polymarket and other prediction markets are just a lawyer's game. You have to read the fine print and interpret it correctly. There is no other way.

Trump says the US will capture Kharg Island. Here's why that's unlikely to happen. Kharg Island is a tiny dot in the Persian Gulf, and it handles 90% of Iran's oil exports. Polymarket currently puts the odds of it falling by June 30 at just 7%. The problem is that the US Marines could probably storm the island. But Kharg sits just 15 miles from Iran's mainland coast. That means Iranian missiles, drones, and artillery would rain down on US troops from day one. Even if, a big if, the USA manages to capture it, they'd need 5,000 soldiers or more just to hold the thing, indefinitely under constant fire. Not to mention, Iran has been laying traps and fortifying defenses specifically in anticipation of a US landing. NO at 93c = 7.5% ROI in ~3 weeks, that's free money. TACO

Anthropic crushed Google's multi-month dominance in February and hasn't looked back since. Google was winning until people actually started using Gemini at scale. Then Google panicked about costs and nerfed its own model. OpenAI does basically the same thing that Google does. They swap expensive models for cheaper ones while charging the same price. GPT-5 feels like GPT-3.5 now. Grok gets only two things right: real-time research and Image and video generation. Everything else is super weak. So, where does this leave us? Claude. You can earn 30.38% (100c/76.7c) in just one month. It’s highly unlikely that any model will surpass Claude’s capabilities within the next 30 days. Its advancements are simply too far ahead of its competitors. If you’re skeptical, try using each of these models for yourself, and you’ll gain a better understanding.

Trump says the US will capture Kharg Island. Here's why that's unlikely to happen. Kharg Island is a tiny dot in the Persian Gulf, and it handles 90% of Iran's oil exports. Polymarket currently puts the odds of it falling by June 30 at just 7%. The problem is that the US Marines could probably storm the island. But Kharg sits just 15 miles from Iran's mainland coast. That means Iranian missiles, drones, and artillery would rain down on US troops from day one. Even if, a big if, the USA manages to capture it, they'd need 5,000 soldiers or more just to hold the thing, indefinitely under constant fire. Not to mention, Iran has been laying traps and fortifying defenses specifically in anticipation of a US landing. NO at 93c = 7.5% ROI in ~3 weeks, that's free money. TACO

BREAKING: President Trump says the US will be "hitting Iran very hard tonight" and announces that the US will be "taking Kharg Island" in the "not too distant future." President Trump also says the US will "assume total control" of Iran's oil and gas markets, "much like we have with Venezuela."

The market is pricing in a 4% chance that a goalkeeper scores at the FIFA World Cup. But across all 22 men's World Cup finals (1930–2022), in which 2,720 goals were scored in 964 matches, not a single goal was scored by a goalkeeper in match play. The closest anyone came was Paraguay's José Luis Chilavert, who in 1998 became the first goalkeeper ever to take a direct free kick at a World Cup finals (vs. Bulgaria) and nearly scored, hitting the top corner only to be denied by a fingertip save.

There is currently only about an ~80% chance that the USA and Iran will NOT sign a peace deal over the next ~2 months

Remember what we said about the election trades on Polymarket. If one option is drastically higher than the rest, that's free money. Armenia's Parliamentary Election Winner market has Civil Contract at 93%. That's a 7.52% ROI in just 2.5 weeks.

BREAKING: FTX founder Sam Bankman-Fried formally applies for a pardon from President Trump.