Home / Guides / How Kalshi's 15-minute Bitcoin contracts settle
How Kalshi's 15-minute Bitcoin contracts settle
The strike is not the price at the open, and the result is not the price at the close. Both are 60-second averages. That one fact changes how you should read a 15-minute window.
The short version
On Kalshi's 15-minute Bitcoin markets, the rules compare two averages. The reference is the simple average of CF Benchmarks' Bitcoin Real-Time Index (BRTI) over the 60 seconds before the window opens. The final value is the same average over the 60 seconds before the window closes. The market resolves Yes if the final average is at least the reference.
Rules can change and each market carries its own text, so read the rules on the market you are trading. This is how the rules read on the markets we checked.
What that means in practice
1. The strike is an average, so the open price is a rough guide
The first price you see when a window opens is not the strike. In a check of 6,465 settled windows over 90 days, the first bar's open was a median of about $10 away from Kalshi's real strike, and the 90th percentile was about $30. The average of the previous minute's open, high, low and close was much closer: a median of about $5 and a 90th percentile of about $13. That is why Carry builds its own strike from the minute before the open, and uses Kalshi's published strike when it is available.
2. A last-second spike counts for one second out of sixty
Because the final value is a 60-second average, one sudden print at the very end moves the result much less than it would move a single closing price. The last minute of a window matters, but it matters as an average.
3. The price on your screen is not what settles
If price is sitting a little past the strike with a few minutes left, the number that decides the window is the average of the final minute, not the number you are looking at. This is why Carry measures how far past the strike price is against how much movement is left, and does not just report the gap in dollars.
4. Close to the strike, even the label can flip
Our proxy result and Kalshi's real result disagreed on 407 of 6,465 windows (6.3%). The disagreements sat almost entirely near the strike. When the final price was within half a typical one-minute move of the strike, 282 of 787 were labelled differently (35.8%). Beyond two typical moves, it was 3 of 3,463. A call with a very thin margin is close to a coin flip, and no chart can fix that.
How Carry handles it
- The strike comes from Kalshi's public market data when it is available, and from the average of the minute before the open when it is not.
- The volatility Carry measures against is adjusted for 60-second average settlement.
- Carry only calls a side when similar spots held at least 90% of the time in past data, and it shows that rate.