Tom Skilling

Estimate. Blends the HRRR, NBM, ECMWF and GFS forecast models, then shifts toward the latest reading. Kalshi settles on the average of several airports (its “Weather Index”), and the bot reads that index straight from Kalshi every 10 seconds. If Kalshi's feed is down, it rebuilds the average from airport reports. That shift matters most in the last hour or two. The range widens with time, model disagreement and rain risk. After enough settled hours, it corrects itself using its past misses.

Odds meter on a position: the dot is the chance your side wins right now, the black line is what was paid, and the small amber marker is the market's odds. Green when the chance is above the price paid, red when below.

Odds is the forecast's chance that side wins. Market % is the midpoint of Kalshi's bid and ask. Edge is expected profit per contract after Kalshi's fee.

Kalshi settles on a decimal reading (like 84.20°), so "84° or above" means 84.00 or higher. Check each market's rules on Kalshi before trading. This is a second opinion, not financial advice.

Practice only. Nothing is bought on your account. Tom trades the market that settles at the next top of the hour, 10 contracts at a time, and wait for an edge to hold 20 seconds before buying. Every hour gets a call: if Tom hasn't bought by 15 minutes before the reading, he buys his lean anyway. If the forecast turns against a position and the other side becomes the likely winner, he sells and flips to it.

Tom (steady) only buys a side he gives 70% or better with at least 4¢ of edge, picking the best return per dollar.

Every 2 seconds Tom compares the bid (after fees) to what the forecast thinks the contract is worth, and sells if the market pays more. Otherwise it holds to Kalshi's result. Fills assume the posted price.