Guide · data from September 20, 2026 to October 19, 2026
Martingale on Kalshi’s Bitcoin 15-minute market: what the real windows did
Martingale means doubling your stake after every loss so one win pays back the whole run. We ran it on every settled Kalshi Bitcoin 15-minute window from September 20, 2026 to October 19, 2026: 2,848 windows, both sides, $100 base, a $200,000 account. It won a little on most days and ran out of money on 4 of 60 runs. Summed across the days, betting up finished −$109,232.71 and betting down −$171,662.79.
How martingale works, and where it ends
Every window is a yes or no question: will Bitcoin close the 15 minutes at or above the price it started at? A contract bought near 50¢ pays $1.00 if it is right. Martingale buys one side every window. After a win it goes back to $100. After a loss it doubles, so the next win covers everything lost since the last win, plus $100.
The catch is the ladder below. Each loss doubles what the next order costs, and the account has to hold the money already lost plus the next order. From $100, the tenth loss in a row is where a $200,000 account cannot place the next bet.
| Losses in a row | Lost so far | Next order | Account needed |
|---|---|---|---|
| 0 | $0.00 | $100.00 | $100.00 |
| 1 | $100.00 | $200.00 | $300.00 |
| 2 | $300.00 | $400.00 | $700.00 |
| 3 | $700.00 | $800.00 | $1,500.00 |
| 4 | $1,500.00 | $1,600.00 | $3,100.00 |
| 5 | $3,100.00 | $3,200.00 | $6,300.00 |
| 6 | $6,300.00 | $6,400.00 | $12,700.00 |
| 7 | $12,700.00 | $12,800.00 | $25,500.00 |
| 8 | $25,500.00 | $25,600.00 | $51,100.00 |
| 9 | $51,100.00 | $51,200.00 | $102,300.00 |
| 10 | $102,300.00 | $102,400.00 | $204,700.00 |
| 11 | $204,700.00 | $204,800.00 | $409,500.00 |
| 12 | $409,500.00 | $409,600.00 | $819,100.00 |
The runs below also pay Kalshi’s taker fee on every order: 0.07 × contracts × price × (1 − price), rounded up, which is 1.75¢ a contract at 50¢ (Kalshi fee schedule).
Every day in the archive
One fresh $200,000 account a day, on each side. Bitcoin settled up in 50.1% of the 2,848 windows, so neither side had anything to double into but a coin flip.
| Day | Windows | Up | Down | Longest streak | Always up | Always down |
|---|---|---|---|---|---|---|
| Oct 19, 2026 | 96 | 45 | 51 | 6 ups | +$3,147.36 | +$1,912.39 |
| Oct 18, 2026 | 96 | 49 | 47 | 7 ups | +$2,625.25 | +$1,102.17 |
| Oct 17, 2026 | 96 | 53 | 43 | 7 ups | +$3,734.06 | −$3,276.22 |
| Oct 16, 2026 | 96 | 43 | 53 | 7 downs | +$1,868.47 | +$2,531.66 |
| Oct 15, 2026 | 88 | 47 | 41 | 9 ups | +$4,493.30 | −$8,940.17 |
| Oct 14, 2026 | 96 | 44 | 52 | 5 downs | +$4,182.12 | +$3,997.31 |
| Oct 13, 2026 | 96 | 47 | 49 | 5 downs | +$4,154.13 | +$3,925.49 |
| Oct 12, 2026 | 96 | 47 | 49 | 9 ups | +$4,202.49 | −$19,970.00 |
| Oct 11, 2026 | 96 | 50 | 46 | 4 downs | +$4,002.11 | +$3,652.88 |
| Oct 10, 2026 | 96 | 50 | 46 | 7 ups | +$2,755.78 | +$1,084.58 |
| Oct 9, 2026 | 96 | 50 | 46 | 5 downs | +$4,779.35 | +$2,701.89 |
| Oct 8, 2026 | 88 | 43 | 45 | 5 ups | +$3,387.60 | +$2,398.30 |
| Oct 7, 2026 | 96 | 41 | 55 | 5 downs | +$1,477.55 | +$4,435.38 |
| Oct 6, 2026 | 96 | 49 | 47 | 6 ups | +$3,310.05 | +$2,849.81 |
| Oct 5, 2026 | 96 | 47 | 49 | 9 downs | +$9,519.38 | +$781.54 |
| Oct 4, 2026 | 96 | 50 | 46 | 10 downs | −$98,117.84 broke 10:30 PM PT | +$2,599.21 |
| Oct 3, 2026 | 96 | 51 | 45 | 6 ups | +$4,437.00 | +$2,419.57 |
| Oct 2, 2026 | 96 | 43 | 53 | 16 downs | −$101,754.87 broke 9:30 AM PT | +$4,936.70 |
| Oct 1, 2026 | 88 | 44 | 44 | 5 downs | +$4,826.52 | +$2,352.07 |
| Sep 30, 2026 | 96 | 49 | 47 | 9 ups | +$4,906.15 | −$4,568.21 |
| Sep 29, 2026 | 96 | 43 | 53 | 5 downs | +$1,600.26 | +$3,222.18 |
| Sep 28, 2026 | 96 | 53 | 43 | 11 ups | +$4,061.05 | −$101,470.94 broke 5:30 AM PT |
| Sep 27, 2026 | 96 | 54 | 42 | 7 ups | +$4,936.51 | −$741.25 |
| Sep 26, 2026 | 96 | 54 | 42 | 10 ups | +$3,684.12 | −$102,720.32 broke 9:15 AM PT |
| Sep 25, 2026 | 96 | 47 | 49 | 5 ups | +$2,482.80 | +$3,456.99 |
| Sep 24, 2026 | 88 | 42 | 46 | 7 ups | +$1,537.71 | +$2,494.17 |
| Sep 23, 2026 | 96 | 42 | 54 | 6 ups | +$1,519.19 | +$6,569.95 |
| Sep 22, 2026 | 96 | 49 | 47 | 4 ups | +$3,534.27 | +$4,144.82 |
| Sep 21, 2026 | 96 | 50 | 46 | 8 ups | −$8,405.81 | +$3,247.85 |
| Sep 20, 2026 | 96 | 50 | 46 | 5 ups | +$3,881.23 | +$3,207.41 |
The 4 days it broke
- October 4, 2026, always up. 10 losses in a row by 10:30 PM PT. To place the next order the account had to cover $200,517.84; it held $200,000.00, so the run stopped there.
- October 2, 2026, always up. 10 losses in a row by 9:30 AM PT. To place the next order the account had to cover $204,154.87; it held $200,000.00, so the run stopped there.
- September 28, 2026, always down. 10 losses in a row by 5:30 AM PT. To place the next order the account had to cover $203,870.94; it held $200,000.00, so the run stopped there.
- September 26, 2026, always down. 10 losses in a row by 9:15 AM PT. To place the next order the account had to cover $205,120.32; it held $200,000.00, so the run stopped there.
The longest streak in the archive was 16 windows in a row settling down on October 2, 2026, 7:00 AM to 10:45 AM Pacific. 4 of 30 days had a streak of ten or more. On a near coin flip, streaks that long are not rare events; they turn up every few days.
That is the shape of every doubling scheme: many small wins, then one run that takes all of them back and more. The small wins are capped at $100 a cycle. The loss is capped only by the account.
If you trade these markets anyway
- Use a fixed unit. The same small stake every window, 1 to 2% of a bankroll you could lose. Bankroll and units
- Write a loss limit before you start, and stop when you hit it rather than doubling to get it back. Tilt and loss limits
- Price the fee in. Near 50¢ the fee is the number you have to beat. How 15-minute markets work
- Check any day yourself. Every window, its settlement and its price path: the Bitcoin 15-minute results archive
Questions people ask
What is the martingale strategy?
Martingale doubles the stake after every loss and goes back to the base stake after a win, so the first win recovers every earlier loss plus one base stake of profit. It needs an unlimited bankroll and no limit on order size to work, and nobody has either.
Does martingale work on Kalshi Bitcoin 15-minute contracts?
Not in our archive. From September 20, 2026 to October 19, 2026, doubling from $100 in a $200,000 account ran out of money on 4 of 60 runs. Summed across the days, betting up every window finished −$109,232.71 and betting down every window finished −$171,662.79.
How many losses in a row does it take to break a $200,000 account?
Ten, starting from $100. After ten losses the account has lost $102,300 and the next order is $102,400, so it needs $204,700 to keep going. Kalshi's fee pushes it a little higher.
How long do Bitcoin 15-minute streaks get?
The longest in the archive is 16 windows in a row settling down, on October 2, 2026, from 7:00 AM to 10:45 AM Pacific. 4 of 30 days had a streak of ten or more.
Is there a safer way to size Bitcoin 15-minute trades?
A fixed unit: the same small stake every window, set as 1 to 2% of a bankroll you could lose, with a loss limit for the day written down before you start. It cannot win back a losing run in one trade, which is the point.
Education, not betting or financial advice. Nobody placed these trades: they are arithmetic on Kalshi’s settled results, entered at each window’s opening price, and a real fill would also pay the spread. Gambling problem? Call or text 1-800-MY-RESET.
Sources
- Kalshi’s public market data for the Bitcoin 15-minute series: the settled rows and one-minute candles behind every day page.
- Kalshi fee schedule
No affiliate links on this page. Found an error? See our corrections policy.