Arbis builds its probability models from publicly available government data, financial markets, and sports statistics. No source is 100% accurate; markets are uncertain by nature. What we do is find where multiple trusted sources agree, and where Kalshi's pricing disagrees with that consensus. That gap is the edge.
Federal funds rate meeting probabilities, the most liquid interest rate market in the world.
Current Fed funds rate, historical rate decisions, economic indicators.
Weekly jobless claims (released every Thursday 8:30 AM ET), CPI inflation data, NFP payroll reports.
Official initial jobless claims PDF, parsed in real-time at 8:30 AM ET on release day for the fastest possible signal.
Live game scores, win probabilities, real-time in-play data for NBA, NFL, MLB, and more.
Consensus bookmaker odds across 6+ sportsbooks. We average these to find the true market probability.
Injury reports: OUT and DOUBTFUL listings that significantly affect game probabilities.
Bitcoin and Ethereum price history to calculate 30-day realized volatility. High volatility = less reliable models = we skip trading.
Crypto market sentiment. Extreme Fear (0-25) means we skip all crypto trades entirely.
Breaking financial news: hawkish/dovish headlines that affect Fed probability models.
Economic calendar, consensus estimates for upcoming releases, financial news sentiment.
Our own data aggregation service. It centralizes all sources above, caches results, and serves clean probability signals to every Arbis installation, so your bot never scrapes directly.
No data source is perfect. Government releases can be revised. Bookmaker odds move. Sentiment changes in minutes. Arbis does not guarantee any trade outcome.
What Arbis does guarantee: it only places trades when multiple trusted sources agree on a probability that Kalshi has priced incorrectly. The edge is mathematical. The outcome is not.
Trade responsibly. Never risk more than you can afford to lose.