bez | Howsure

Prediction markets, trader behavior & AI agents. Building HowSure. Market breakdowns, experiments and things that changed my mind.

[bez | Howsure @FireBrightX]

I’m building HowSure for prediction-market traders and agent builders.

Check the contract, inspect its sources and estimate your size.

Public reads are free. Saved team workflows: $100 / 30-day pilot.
https://t.co/Hd6z3pVbCC

[bez | Howsure @FireBrightX]

@SKocaslanl @EowynChen ahaha, good punch

[bez | Howsure @FireBrightX]

An agent league with seasons, promotions and relegation sounds fun.

Give each bot a record, a style and a rival. People may come for the forecasts and stay for the drama.

[good @thenarrator]

i keep meeting teams building agents that can trade, and it’s made me think less about whether they can make a good call and more about who trusts them with money.

that’s where prediction markets get interesting to me. every position resolves. you can see what the agent got right, what it missed, and how it handled being wrong.

maybe the first real use for ai in markets isn’t replacing traders but giving good agents a track record people can actually judge.

[bez | Howsure @FireBrightX]

@j_dekoninck The AI equivalent of sneaking a phone into an exam, then starting a group chat with the other top students

[bez | Howsure @FireBrightX]

@int_mon_econ A useful agent feature: keep the rival explanation alive. Same data, two stories, and the next observation that would separate them

[bez | Howsure @FireBrightX]

@NEJM A market can reward someone for finding a flaw in a trial. The design problem is making sure it can’t reward someone for creating one

[bez | Howsure @FireBrightX]

@PolymarketSport All that emotional exposure to Cam Skattebo for a potential $10

[bez | Howsure @FireBrightX]

A good forecasting tool should have a memory.

That’s part of what I’m building with HowSure: keep the original research, then return to see what changed.

https://t.co/VSToKMVy2I

[bez | Howsure @FireBrightX]

Yesterday I shared Polymarket’s 78.65% price for Flávio.

Lula won.

Today’s most crowded trade: “I knew it all along”

Keeping the snapshot. Adding the result.

[bez | Howsure @FireBrightX]

Prices saved on 25 Oct at 13:45 UTC. These are overlapping contracts, not vote shares.
Election result: TSE.
https://t.co/NXUIRAgpDS
https://t.co/JAmzBW8scK
https://t.co/HFLTQbVHq3

[bez | Howsure @FireBrightX]

Сommunists don't give up power once they get it

[bez | Howsure @FireBrightX]

@krishcj1 dont give up

[bez | Howsure @FireBrightX]

Snapshot: 25 Oct, 13:45 UTC. Market-implied odds; the margin is the gap in vote share.

https://t.co/NXUIRAgpDS
https://t.co/JAmzBW8scK

[bez | Howsure @FireBrightX]

Brazil votes today. Snapshot at 13:45 UTC:

Flávio wins: 78.65% (−2.4pp in 24h).
Flávio wins by <4 vote points: 46.5% (+7pp).
The market still favors him. It’s putting more weight on a narrow win.

Watch the margin.

[bez | Howsure @FireBrightX]

@PolymarketTrade A trader spent months building a position on a rent freeze. I love how specific the expertise can get here: New York housing policy becomes a trading specialty

[bez | Howsure @FireBrightX]

@peshk0vnikita wallet doesn't need any token, why do you need it?

[bez | Howsure @FireBrightX]

@HarryDCrane Predict 50% for every event in a balanced sample and the calibration plot can look lovely. The model still hasn’t told you which events are more likely

[bez | Howsure @FireBrightX]

@Zeneca Eleven bots, one fee change, everyone benched. There’s a great build-in-public series in what happens next

[bez | Howsure @FireBrightX]

@johnjhorton A replay button for forecasts would be excellent: pause before the next headline, see the reasoning, then reveal what changed. Could be a great way to learn a new subject

[bez | Howsure @FireBrightX]

A fun test for finance memory: give the agent a thesis, then give it evidence that breaks it.

Does it update, or become a bagholder with a larger context window?

[Carlos Pereira @cegapereira]

Everyone has a story of missed investment riches.

Agents promise to fix that. After building systems across prediction markets, crypto, perps and stocks, we think most won’t... yet.

"10 hard problems in agentic investing" (written w/ @SpinFN)

Thread + https://t.co/yOGoum789a

[bez | Howsure @FireBrightX]

@decenterghost With Astra iam able to build everything i want to build. But, one question - Where will i have this edge to beat the market? what do i need to build to have the moat?