Skip to content
Intro offer Your first month of Pro is $1

Kalshi guide

Kalshi settles a new 15-minute Bitcoin market 96 times a day.

Each market asks one question: will Bitcoin's price at the end of the 15 minutes be at or above its price at the start? This page shows how they settle, what they cost in fees, and how the last 14 days of settled windows came out, read from Kalshi's public data. You can practice the same call for free on Prediction Rush.

Read from Kalshi at Nov 18, 9:51 PM ET

15-minute windows read over 14 days
1,328
of windows ended Up
46.8%
typical move in one window, up or down
$84
15-minute windows every day
96

Each market asks whether Bitcoin ends its 15 minutes at or above where it started.

Kalshi lists these as the series KXBTC15M, titled "BTC price up in next 15 mins?", and you buy Yes (Up) or No (Down). Kalshi's rules say the market resolves to Yes if the average Bitcoin index price over the last 60 seconds is "at least" its average over the 60 seconds before the window opened.

One window

End at or above start: UpEnd below start: Down
UpDownstart priceend priceopensthe windowcloses
How one window resolves on Kalshi

A new window opens as soon as the last one closes, so Kalshi settles 96 of them a day. Each window's start price is the previous window's end price: this held for 1,325 of the 1,325 back-to-back windows we read, so the windows chain into one unbroken price line.

Kalshi settles on the CF Benchmarks Bitcoin index, and a tie counts as Yes.

The settlement source is CF Benchmarks. In the last minute before a window closes, 60 prices of its Bitcoin Real Time Index are collected, and the official value is their average, rounded to the nearest cent. That is not the price on any one exchange.

Because the rule says "at least", a window that ends exactly level with its start resolves Yes. None of the 1,328 windows we read ended exactly level.

In the last 14 days, 46.8% of Kalshi's 1,328 settled windows ended Up.

Up share by day

40%50%60%Nov 653%Nov 749%Nov 848%Nov 951%Nov 1048%Nov 1146%Nov 1348%Nov 1442%Nov 1552%Nov 1650%Nov 1747%Nov 1840%
Share of windows that ended Up on each full UTC day. The line marks 50%. Across 12 full days the daily share ranged from 40% to 53%.
Settled windows read from Kalshi, Nov 5 to Nov 19
MeasureValue
Windows read1,328
Ended Up622 (46.8%)
Ended Down706 (53.2%)
Median volume (middle half)3.9M contracts (3M to 4.6M)
Windows per day96
Settled windows read from Kalshi, Nov 5 to Nov 19

Over the same days Bitcoin fell 22%, from $80,543 to $62,581, by Kalshi's own start and end prices.

Half of the windows moved less than $84, and 246 ended within $25 of where they started.

Size of each window's move

Ended UpEnded Down
Up $500+13 (1%)Up $400 to $50010 (1%)Up $300 to $40033 (2%)Up $200 to $30051 (4%)Up $100 to $200144 (11%)Up $0 to $100371 (28%)Down $0 to $100386 (29%)Down $100 to $200181 (14%)Down $200 to $30076 (6%)Down $300 to $40034 (3%)Down $400 to $50011 (1%)Down $500+18 (1%)
How far each window's end price was from its start price, in steps of $100.

Nine in ten windows moved less than $280. The result carries no size: a window that ends $20 above its start is an Up, the same as one that ends $400 above. The 22% fall over the whole read works out to about $14 per window, against a typical move of $84 in either direction.

The longest run of one result in 15-minute windows was 7 Ups in a row.

Runs of the same result

Runs in the windowsIndependent flips
1 in a row3533322 in a row1771653 in a row90824 in a row32415 in a row19216 in a row19107 in a row258+ in a row06
How many runs of each length the 15-minute windows made, next to the count independent flips would make.

Runs of 3 or more made up 23% of the 692 runs. If every window were an independent flip with the same 46.8% Up share, about 25% of them would.

After 3 Ups in a row, the next 15-minute window ended Up 38.5% of the time (117 cases). After 3 Downs in a row, it ended Up 49.7% of the time (179 cases).

Typical volume is highest in the 6 PM hour and lowest in the 6 AM hour, New York time.

Median contracts by hour

2.1M4.8M12a3a6a9a12p3p6p9p
Median contracts traded in a window by hour of day, New York time. Kalshi counts contracts, not dollars.

A typical window in the 6 PM hour traded 4.8M contracts, against 2.1M contracts in the 6 AM hour, about 2.3 times as much.

Kalshi's taker fee on these markets is largest near 50 cents.

A taker pays 0.07 × contracts × price × (1 − price), rounded up to the next cent. Kalshi's series record lists this series with the plain quadratic fee type and a multiplier of 1, which means no separate maker fee.

Taker fee on 100 contracts at a given price
Contract priceFee on 100 contractsFee as a share of the stake
10 cents$0.636.3%
30 cents$1.474.9%
50 cents$1.753.5%
70 cents$1.472.1%
90 cents$0.630.7%
Taker fee on 100 contracts at a given price

Fees on every other Kalshi market are on Kalshi fees, explained.

You can practice the same Up or Down call for free on Prediction Rush.

Prediction Rush shows you ten real Bitcoin charts with the labels removed and gives you five seconds to call the next candle on each. Three strikes end a run. It plays without an account, and a free account lets you post a ranked run under your name. Nothing is at stake, and a call takes seconds.

These are the questions people ask about Kalshi's 15-minute Bitcoin markets.

How do Kalshi's 15-minute Bitcoin markets settle?

A market resolves to Yes if the average of the CF Benchmarks Bitcoin Real Time Index over the 60 seconds before the window closes is at least its average over the 60 seconds before the window opened, and to No otherwise. A tie therefore resolves to Yes.

Is there a strategy that works on these markets?

We cannot point to one. In the last 14 days, 46.8% of windows ended Up, half of them moved less than $84, runs of 3 or more were 23% of all runs against 25% for independent flips, and the taker fee is charged on every buy. Prediction Rush lets you practice reading short windows for free.

Do bots trade Kalshi's 15-minute Bitcoin markets?

We found no published figure for Kalshi. The bot-share study we found covers Polymarket, where Blockchain.News reported automated addresses at 55% to 62% of volume in its 5-minute and 15-minute markets. Kalshi runs a public trading API, so software can trade these markets.

Where do people discuss these markets?

We have no number to report for Kalshi, because the public market records we read carry no comment counts. Polymarket shows a comment count on each of its series.

Is there a practice mode for Kalshi's Bitcoin markets?

Kalshi offers a demo environment with mock funds, set up mainly for people testing software against its API. It uses separate credentials from the real site, and Kalshi's docs warn that demo prices and behavior may not match real markets. Prediction Rush needs no setup: ten real charts, five seconds each, and no account.

What does a trade on these markets cost?

A taker pays 0.07 × contracts × price × (1 − price), rounded up to the next cent: $1.75 on 100 contracts at 50 cents, and $0.63 at 10 cents. This series has no separate maker fee.

These pages go further on Kalshi and on the same Up or Down call.

Kalshi 15-minute Bitcoin markets explained: how a window settles, the fee on each trade, and the last 14 days of settled windows
Kalshi's 15-minute Bitcoin markets at a glance: how a window settles, what the taker fee costs, and how the last 14 days of windows came out. On BitPredict you can practice the same call for free.