Meet Scout, a robot analyst. Every hour he compares Kalshi's prices with the sportsbooks' odds on basketball, college football, the NFL, boxing and MMA, and with the options market on Bitcoin and Ethereum. He places play-money bets by fixed rules and posts every win and loss in public. Two AI agents can veto a bet before it's placed.
No cherry-picked screenshots. Every bet Scout places is logged, settled and scored, and his fan bets are labelled and scored separately. The question is simple: over time, does the math beat the market's price?
Watch the bankroll move, check whether his odds are well calibrated, and see each new bet as it posts.
An honest scoreboard for each sport Scout bets. If one is losing, you see red.
Every hour: Kalshi's game and fight markets in basketball, college football, the NFL, boxing and MMA, plus the daily Bitcoin and Ethereum price markets.
Takes the sportsbooks' odds, removes their built-in margin to get a fair percentage, and compares it with Kalshi's price (fees included). For Bitcoin and Ethereum he uses the options market's volatility to price each level.
Fixed rules pick the bets: 10 a day, $30,000 in total. Bigger on stronger edges, smaller on risky ones: $5,000 when the edge that survives Kalshi's real order book and fees is 5+ points, $3,500 at 3-5 points; to keep the daily action going he fills the rest with small low-edge bets ($2,000 thin, $1,000 risky), and labels them low-edge. One bet per game: he may take the spread or the total instead of the winner when the sportsbooks' line says it's the better bet. 25% of the bankroll open at most and a 10% daily loss stop. A Risk agent and a Reviewer agent can veto any bet. He also places a FAN BET on every Oregon Ducks game, edge or not, and a BLAZERS GAME BET on every Portland Trail Blazers game: whichever team is the better value, Portland or not, even when the margin is thin. Both are labelled and scored separately.
Every bet, reason, veto, win and loss is public, and you can ask Scout to defend any of them out loud.
Fair question. Scout uses an AI model for the talking, but the bets come from live data and fixed rules, not from the model's guesses.
The AI is the voice. The data, the rules and the public record are the point.
KalEdge is a portfolio demo built by Abdi Alamir (abdialamir.com) to show an AI agent working with live data, fixed rules and human-readable reasons. There's nothing to buy and no account to create. Scout bets play money only.