Leveraged Bitcoin ETFs: Daily Reset on Real Prices

A 2x or 3x bitcoin ETF promises a multiple of bitcoin's return for one day, measured from one daily close to the next, and nothing about any longer stretch. Held for a year, the result depends on the path: replayed on Coin Metrics daily closes from 2018 to October 22, 2026, a 3x daily fund finished below its starting value in 604 of the 1,834 rolling 365-day windows in which bitcoin itself finished higher.

The replay behind that figure is a rule applied to a price series, not a fund. The method section says exactly what it leaves out, and every fund result is therefore better than a real fund's. Nothing here is a forecast or a view on whether anyone should hold one of these funds, and the first US 3x bitcoin fund had not been confirmed as trading when this was written.

What the funds say they do, in their own documents

The exchange filing for the 3x funds is the most useful primary text, because an exchange must describe a fund precisely before it can list it. Cboe BZX filed SR-CboeBZX-2026-065 to list six series of the VS Trust: 3x Gold, 3x Silver, 3x Bitcoin, 3x Ether, 3x Crude Oil and 3x Natural Gas. The filing's own words for the bitcoin fund are that it "seeks daily investment results, before fees and expenses, that correspond to three times (3x) the daily performance of bitcoin, as measured by the daily changes in the price of a specified portfolio of first- and second-month futures contracts on bitcoin that trade on the Chicago Mercantile Exchange."

Three details in that sentence decide everything that follows. The objective is daily. It is stated before fees and expenses. And the thing being tripled is a futures benchmark, not a spot price, which the filing confirms a few lines later: "The Fund does not invest in physical bitcoin, but rather seeks to achieve its investment objective primarily through investment in futures contracts on bitcoin, cash, and Cash Equivalents." The benchmark rolls from the near contract to the next over a five-day period each month, with about 20 percent of the expiring positions rolled each day of that period.

The filing also explains why an exchange rule change was needed at all. The generic listing standard for commodity-based trust shares, BZX Rule 14.11(e)(4)(F), prohibits leveraged products, and the filing says that because each fund seeks three times its benchmark's daily result, "the Funds do not meet the standard." The proposal asks for an exception to that single standard and says each fund otherwise complies with the generic rule. The filing was approved by the exchange's president under delegated authority on August 10, 2026. That is an exchange-side step and not the SEC's decision.

The 2x fund that already trades describes itself in almost the same terms. The ProShares Ultra Bitcoin ETF (BITU) page, read on October 23, 2026, says the fund "seeks daily investment results that correspond, before fees and expenses, to 2x the daily performance" of the Bloomberg Bitcoin Index. The page gives an inception date of April 1, 2024 and an expense ratio of 0.95 percent, states that the fund "invests in futures and swaps and does not directly invest in bitcoin," and lists as its largest exposures, as of October 22, 2026, swaps on the IBIT and FBTC spot funds plus a CME bitcoin futures contract. On holding periods it is blunt: "For any holding period other than a day, your return may be higher or lower than the Daily Target," and "these differences may be significant."

The trade the fund makes every afternoon

A fund that promises a fixed multiple of each day's move has to change its exposure after every move. The arithmetic is short enough to do by hand.

Suppose a 3x fund starts the day with 100 million dollars of net assets and 300 million dollars of bitcoin exposure. Bitcoin rises 5 percent. The exposure is now 315 million and the net assets are 115 million, because the fund keeps all of the gain on its 300 million of exposure (15 million) and adds it to its capital. To be 3x again tomorrow it needs 345 million of exposure, so it must add 30 million. If bitcoin had fallen 5 percent instead, exposure would be 285 million, net assets 85 million, the target 255 million, and the fund would sell 30 million.

The general rule is that the day's rebalancing trade equals L times (L minus 1) times the day's return times the starting net assets, where L is the multiple. For the 3x fund that is 3 times 2 times 5 percent times 100 million, or 30 million. For a 2x fund with the same assets and the same 5 percent day it is 2 times 1 times 5 percent times 100 million, or 10 million. The trade always has the same direction as the day's move: the fund buys after a rise and sells after a fall. That is the mechanical reason daily-reset products behave differently from the same leverage held through a margin account, where nobody resets the exposure for you. It is also the reason the funds are called path-dependent, and the next section shows the cost of that path dependence in two lines of arithmetic.

The fee is a separate and smaller effect. At the 0.95 percent that the ProShares page lists for BITU, a year of fees on 10,000 dollars of fund value is 95 dollars, taken out gradually. The page shows a single expense ratio and does not say whether it is gross or net. Where the financing cost of swaps and the cost of rolling futures appear is in the prospectus, which this page did not read.

Two days that cancel out for bitcoin and not for the fund

The cleanest test of a daily-reset rule is a price that returns to where it started. Take a first day of plus 10 percent followed by a second day of minus 9.0909 percent, which is exactly the move that undoes a 10 percent rise (1 divided by 1.1, minus 1). Bitcoin ends at 1.0000 times its starting price. A 3x fund does not.

The first row works out as follows. After day 1 the 3x fund is at 1.30 (three times 10 percent). On day 2 it loses three times 9.0909 percent, which is 27.2727 percent, so it ends at 1.30 times 0.727273, or 0.94545. The 2x fund goes from 1.20 to 1.20 times 0.818182, or 0.98182. The second row alternates the same 10 percent move four times: bitcoin ends at 0.99 times 0.99, which is 0.9801, down 1.99 percent. The 3x fund multiplies 1.30 by 0.70 twice and ends at 0.8281, down 17.19 percent. The 2x fund multiplies 1.20 by 0.80 twice and ends at 0.9216, down 7.84 percent.

Neither sequence needs a crash. A fund that loses 5 percent while its benchmark is flat is not malfunctioning; it is doing what the filing says, and a swing of 10 percent in a day is ordinary for bitcoin. The more days of this kind a holding period contains, the larger the gap, which is what the replay measures next.

The one-day cliff for a 3x fund

There is a second property that is not about compounding at all. A fund at 3x loses 100 percent of its value on a day when its benchmark falls by one third, because three times 33.3 percent is 100 percent. A 2x fund needs a 50 percent daily fall to reach the same place. Whether a given fund would actually lose everything on such a day depends on how its futures and swaps are positioned and settled, which the prospectus governs and this page did not read. The arithmetic, though, sets a floor under how bad the daily distribution of the benchmark can be before the product's own rule breaks.

In the Coin Metrics daily series, only one close-to-close fall since 2014 was larger than one third: March 12, 2020, when the fixing went from 7,939.34 dollars on March 11 to 4,959.31 dollars, a fall of 37.53 percent. The next largest in the series is a fall of 21.6 percent on January 14, 2015. Since 2018 there are no days at or beyond a 25 percent fall other than March 12, 2020 itself, and none at or beyond 50 percent. The replay below therefore shows a 3x fund going to zero on that one day, and every later number for 2020 flows from that.

A caution belongs here. Coin Metrics' PriceUSD is a fixing as of 00:00 UTC, and the funds are benchmarked to CME futures that settle at a different time of day. The sharp fall on March 12, 2020 was a very fast intraday move, and a futures benchmark measured at a different cutoff would show a different one-day change. No US-listed 3x bitcoin fund traded in 2020.

Replaying the rule on real bitcoin prices

The method is simple enough to repeat. Take the Coin Metrics bitcoin PriceUSD daily series from the community API, pulled October 23, 2026, with the last observation dated October 22, 2026. For each day compute the close-to-close return r. A fund with multiple L starts at 1 and multiplies by 1 plus L times r each day, floored at zero. There are no fees, no financing cost, no futures roll cost and no tracking error. Results are in the series' own date labels. The sample starts on January 1, 2018 and runs through October 22, 2026, which is 3,217 daily observations. The start date is a choice, and a different one changes the counts below.

Calendar years

The "2x of bitcoin's return" and "3x of bitcoin's return" columns are what a reader would get by multiplying the year's price change, which no daily-reset fund promises; they are shown so the gap is visible. Three things stand out. In the 2018 and 2022 declines, the funds lost less than two or three times the move, because a percentage loss cannot exceed 100 percent, yet they still lost 96.6 percent and 99.8 percent in 2018. In 2023 and 2024 the compounding worked the other way: strong trending years left the 2x fund ahead of double and the 3x fund ahead of triple in 2023. And in 2025, with bitcoin down 6.3 percent, the 3x fund was down 51.5 percent, which is nowhere near three times 6.3 percent.

The 2020 row is a special case. The 2x fund matched its naive target, but the 3x fund shows minus 100 percent in a year when bitcoin rose 304.9 percent, only because of the March 12 fixing described above. Treat that cell as a statement about the rule, not as a prediction about any product.

The same table, peak to trough

Each peak is the highest close inside a search window around it (November 2017 to January 2018, October to December 2021, calendar 2025), and each trough is the lowest close after that peak up to a stated cutoff, all from the same series. The third row shows why the products attract attention: an uninterrupted rise of 692 percent produced 7,791.6 percent for the 3x rule, more than the naive 2,076 percent. Compounding helps in a persistent trend and hurts in a choppy or falling one. It does not favor either side, and no one can know in advance which path is coming. The last row is the most recent stretch in the series: bitcoin is 34.9 percent below its October 6, 2025 close of 124,824.45 dollars, and the 3x rule is 85.2 percent below its own start.

How often a rising bitcoin still left the fund underwater

The calendar-year view depends on where the year boundaries fall. A rolling view removes that choice: take every possible start day from January 1, 2018 and hold for exactly 30, 90, 180 or 365 days. These windows overlap heavily, so they are not independent observations and the counts below describe this one price history, not odds.

Windows with a flat outcome make the same point more sharply. Among windows in which bitcoin finished within 5 percent of where it started (the set is 862 windows of 30 days, 336 of 90 days, 179 of 180 days and 78 of 365 days), the median result for the 3x rule was minus 6.6 percent, minus 18.3 percent, minus 41.7 percent and minus 82.4 percent respectively. For the 2x rule the medians in the same sets were minus 2.7, minus 7.0, minus 17.6 and minus 44.4 percent. In all 78 of the 365-day flat windows, both the 2x and 3x funds finished lower. The medians are taken over exactly those flat-outcome sets and not over all windows, and the 365-day set is small and its 78 windows overlap heavily.

Windows in which bitcoin fell mirror this. Over the 1,018 rolling 365-day windows with a lower end price, the 3x rule lost more than 90 percent in 599 of them (58.8 percent). The median bitcoin result in that set was minus 33.8 percent and the median 3x result was minus 92.0 percent.

Where the drag comes from, and a rule that checks it

The gap between a fund's result and the naive multiple is not random. For daily returns that are small, the log of a fund's growth is close to L times the log growth of bitcoin minus a drag of about L times (L minus 1) divided by 2 times the sum of the squared daily returns. For the 2x rule that is one times the sum of squares; for the 3x rule it is three times. The sum of squared daily returns over a year is close to the annualized variance of daily returns, so the 3x drag is roughly three times the annualized volatility squared, measured in natural-log points.

To turn a drag into a result, cube the bitcoin year's end-to-start ratio for the 3x fund and multiply by e raised to the minus drag. For 2019 bitcoin ended at 1.944 times its start; cubed that is 7.35, and e to the minus 1.47 is 0.230, so the 3x fund ends at about 1.69 times, a gain of 68.7 percent. The years in the table are those with no wipe-out day; 2018, 2020 and 2022 are omitted because the approximation breaks down when single-day moves are very large. The point of the table is that the cost of holding the 3x rule for a year is set by how violently bitcoin moved day to day, and it is large: a year at 70 percent volatility costs about 1.5 natural-log points, which is a factor of roughly four and a half, before any fee. This is also why the same funds look so different in a calm trending year and a choppy one.

The status of the first US 3x fund, and what could not be confirmed

The Cboe filing is primary text. Reports from several outlets say the SEC approved the rule change on October 2, 2026, and one cites Release No. 34-106577; this page did not retrieve the SEC order itself, so it treats those figures as secondary. The same reports say trading cannot begin until the VS Trust registration statement on Form S-1 becomes effective, and that the approval order sets no launch date. As of October 23, 2026, no source this page checked reported that shares had started trading. One outlet gave a proposed ticker for the bitcoin fund, but this page could not match it to a primary document and uses none.

An exchange approving a listing standard and the SEC reviewing an investment's suitability for any particular investor are different things. The filing itself says that exchange members who carry customer accounts follow FINRA Regulatory Notices 09-31, 09-53 and 09-65, which set sales-practice and margin expectations for leveraged and inverse products, and that the exchange will tell members about the risks before trading starts. Those notices are the place to look for what a brokerage may require of a customer before allowing a purchase.

What the sample cannot tell you

The windows are rolling and overlap, so the counts describe one history and are not probabilities, and the sample holds a small number of full market cycles, so the large payoff in the 2022 to 2025 advance is a fact about this record and not a property of the products. For margin and forced selling in the futures markets these funds use, see liquidations and cross and isolated margin; for how far bitcoin itself moves day to day, see bitcoin's volatility and drawdown history.

Frequently Asked Questions

Can a 3x bitcoin ETF lose all of its value in one day?

On the arithmetic, yes: a 3x daily fund loses 100 percent on a day its benchmark falls by one third. Whether a real fund would is a question for its prospectus and for how its futures and swaps are positioned. In the Coin Metrics series used here, only one close-to-close fall since 2014 exceeded one third, the 37.53 percent fall on March 12, 2020, and no US-listed 3x bitcoin fund traded then.

Is a leveraged bitcoin ETF the same as buying bitcoin with borrowed money?

No. With a margin loan the exposure is fixed in dollars until you change it, so a rise lets leverage fall and a fall pushes it up toward liquidation. A daily-reset fund changes its exposure every day to keep the multiple constant, buying after rises and selling after falls. That is why its long-run result is not the leveraged return of the period. See the liquidations guide for how margin calls work.

What does a daily objective mean in the prospectus language?

It means the objective applies to one trading day, measured between the same two times each day. The ProShares page says that for any holding period other than a day the return may be higher or lower than the Daily Target and that the differences may be significant. The Cboe filing for the 3x funds uses the same daily wording and adds "before fees and expenses."

Have the 3x bitcoin and ether ETFs started trading in the US?

This page could not confirm it. Secondary reports say the SEC approved the exchange rule change on October 2, 2026, that the funds must wait for their S-1 registration to become effective, and that no trades had started. Check the issuer's website and the exchange's notices for the current status before relying on any report, including this one.

Does the expense ratio capture the full cost of holding one?

Not necessarily. The ProShares page shows 0.95 percent for BITU, which is 95 dollars a year on 10,000 dollars, but it does not say whether the figure is gross or net, and the financing and roll costs of swaps and futures are separate matters that the prospectus covers.

Why does a 3x fund lose money when bitcoin goes nowhere?

Because each day's gain or loss is multiplied before it compounds. In the two-day example on this page bitcoin ends flat and the 3x fund ends 5.45 percent lower. In the 78 rolling 365-day windows in which bitcoin finished within 5 percent of where it started, the median 3x result was minus 82.4 percent.

How is this different from the spot bitcoin ETFs?

A spot fund holds bitcoin and its price tracks bitcoin's, with a fee. The funds on this page hold futures and swaps, seek a multiple of one day's return, and reset every day. The page on how spot bitcoin ETFs work covers custody and creations, and spot bitcoin ETFs compared lists the funds.

Sources

CryptosEyes publishes general educational research, not investment, legal, or tax advice. This article describes how daily-reset funds are built and how a stated rule performed on a historical price series. It does not say that any product is suitable for any person, and past results of a rule on a price series do not describe any fund. Leveraged products can lose most or all of their value over short periods.