About Ant–Whale

What it does

Ant–Whale reads live Bitcoin order flow and estimates whether BTC will finish above or below Kalshi's 15-minute target, shown next to Kalshi's own odds. Each round, the model records checkpoints at T-10, T-5, T-3, and T-1 before settlement, then compares its estimate against the actual outcome so you can audit the track record. It also surfaces short-horizon "Pre-Ant" opportunities when price action looks likely to move meaningfully before the round closes. Everything is recorded before the outcome is known, so the history is free of look-ahead bias.

Who it's for

Ant–Whale is built for people who already trade Kalshi's KXBTC15M Bitcoin markets and want a second, data-driven view of each round alongside the exchange's own odds. It's a research and decision-support tool, not a signal service or automated trader. If you're comfortable reading probabilities, comparing model estimates to market prices, and managing your own risk, the dashboard gives you a structured way to watch order flow and round dynamics in real time.

Who builds it

Ant–Whale is an independent experimental research project, built and tested in public by a small team interested in short-horizon prediction markets. It's not affiliated with Kalshi or any exchange. The model is versioned (currently AW-0.2.0) and every change is tracked so the track record stays reproducible. Development is funded by voluntary community support, not by selling signals or taking a cut of trades.

How to read the Simple tab

• The big Up / Down numbers are the model's estimate for the current round.
• Green means the model leans Up; red means it leans Down.
• The "Kalshi" line shows the market's own odds for comparison.
• The Act card shows whether the round is early, middle, late, or in the settlement window.

Public beta

This is an experimental research project in public beta. Probabilities are estimates, not guarantees. Nothing here is financial advice. Only trade what you can afford to lose.

Feedback & support

Use "Report a signal" in the header to send feedback on any round, or the Contact page to reach the team. Use "Support the research" to contribute — contributions fund development and testing, not better predictions.