A single Bitcoin purchase, left untouched, can grow into a portfolio's dominant risk position without the investor ever adding to it. Backtests in 21Shares' November 5 allocation research trace what that kind of growth did to portfolio losses between 2020 and 2026, and how much a simple rebalancing rule limited the damage. 21Shares issues crypto exchange-traded products, and its figures are hypothetical backtests rather than returns any investor actually earned.

The study's central findings: a rally alone can more than triple an investor's exposure; skipping rebalancing deepened the 2020–23 drawdown to 50.6%; drift paid off for as long as crypto kept rising; and sticking to a rule mattered more than how often that rule was applied.

How 5% Becomes 17% Without Another Purchase

Start with a hypothetical $100,000 portfolio holding $5,000 in Bitcoin and $95,000 in everything else. If Bitcoin quadruples while the other assets stay flat, the Bitcoin holding is worth $20,000 and the portfolio $115,000. Bitcoin now makes up 17.4% of the investor's wealth. The share stops short of 20% only because the rally enlarged the portfolio along with the position.

Portfolio managers call this drift. It requires no decision from the investor, which is what makes it easy to miss: measured by what was paid, the purchase still looks small, while the exposure has more than tripled. The mechanism is general: any holding whose return diverges from the rest of a portfolio changes its weight the same way, whether the holding is a stock, a bond or a cryptocurrency.

Now place two investors in that position and let Bitcoin halve. The first leaves the holding alone. The second sells back to the original 5% just before the fall, keeping $5,750 in Bitcoin and moving $14,250 into the rest of the portfolio.

Coindoo calculations. Other assets stay flat; no leverage, fees, taxes, contributions or withdrawals. The reset is timed just before the fall to isolate exposure, and a real review schedule could act earlier or later. The 5% figure is an arithmetic input, not an allocation suggestion.

Investor A loses more than three times as much and still ends $5,000 ahead of the original $100,000. Judged by cost basis, nothing went wrong, because the position still shows a profit. Judged by the portfolio, the investor absorbed a drawdown — a fall from its own peak — sized by a 17% bet rather than the 5% one originally chosen.

What Drift Did in 21Shares' Backtests

21Shares ran a version of this experiment on real price history. Its growth profile funds a 5% crypto sleeve from the equity side of a 60% stock, 40% bond portfolio, split 3% Bitcoin, 1.5% Ethereum and 0.5% Solana. The 60/40 mix is a conventional baseline in traditional asset allocation, which is why the study uses it as the yardstick against which each profile is measured. Bitcoin makes up most of the sleeve rather than all of it, and the firm compared the profile rebalanced every quarter with the same profile never rebalanced.

Source: 21Shares, From Theory to Allocation (Q3 2026). Hypothetical backtests.

Across September 2020 to September 2023 — a stretch that took in the 2021 rally and the 2022 crash — the two versions of the same portfolio split sharply against the plain 60/40 benchmark. Quarterly resets held the crypto sleeve's addition to the worst decline to about 2.7 percentage points. Without them, the sleeve added about 21.9 points. Annualized volatility, a measure of how widely a portfolio's returns swing around their average over a year, moved the same way: 13.99% with resets, 22.87% without.

Both portfolios began with an identical 5%. Nothing about the purchase separated them; the gap opened after the rally, when one kept its sleeve at size and the other let it grow. The moderate profile, 3% Bitcoin and 1% Ethereum, followed the same pattern over that cycle, with a 30.59% maximum drawdown when rebalanced and 36.78% when left alone.

The October 2023 to September 2026 window, when crypto mostly climbed, repeats the ordering on a smaller scale. Its Bitcoin-only conservative profile gives the most direct read on Bitcoin itself: with 2% in Bitcoin and quarterly resets, its maximum drawdown came within 0.12 percentage points of the benchmark's. Left to drift, it ended 1.30 points deeper.

Keeping the Position Small Has a Price

The same window shows what discipline cost. With crypto rising for most of those three years, portfolios left alone earned more. In the Bitcoin-only profile, the unrebalanced version returned 13.18% a year, against 11.64% with quarterly resets. The extra return came from carrying a bigger Bitcoin position than the investor had signed up for.

The mechanism runs both ways. A rebalancing rule sells part of whatever has risen, so it surrenders some of any rally that keeps going. After a fall, the same rule buys more of the asset that dropped. In the steady decline from November 2021 to the end of 2022, 21Shares found that unrebalanced portfolios lost slightly less, by under one percentage point a year, because their crypto weight shrank on its own.

None of this says whether Bitcoin is a good investment. Rebalancing controls size, not conviction, and Coindoo's analysis of why a return forecast does not settle a Bitcoin allocation deals with the separate question of how much expected return should count.

Set the Rule on the Day You Set the Size

21Shares tested each profile under five habits: weekly, monthly, quarterly and annual resets, and no resets at all. (Source: 21Shares, From Theory to Allocation (Q3 2026). Hypothetical backtest.)

The four disciplined rules ended within about a third of a percentage point of one another on worst fall, and the once-a-year reset did no worse than the weekly one. The portfolio with no rule was the outlier. 21Shares concludes that sticking to a rule counts for more than its exact frequency, and that the rule is best agreed at the same time as the allocation size.

For an individual holder, that turns rebalancing from a market call into a commitment made in advance. Trimming a winner mid-rally invites second-guessing about whether the run has further to go. A rule chosen on purchase day takes that judgment out of the moment.

The SEC's Investor.gov guide to asset allocation describes three ways to rebalance: selling, buying, or redirecting new contributions. It also separates reviews held on a calendar from reviews triggered when an allocation moves. Applied to the rally example, a calendar rule acts at the next scheduled date however far the position has run. A band — for instance, a review whenever Bitcoin leaves a 3% to 7% range — acts once the weight crosses 7%, well before it reaches 17%. Steering new money elsewhere avoids selling altogether, though modest monthly contributions would take a long time to absorb a $14,250 overweight.

What the Backtest Leaves Out

Every reset is a transaction. Commissions and bid-ask spreads apply each time, and exchange-traded products carry ongoing charges on top; the SEC's investor bulletin on fees sets out how both kinds of cost reduce returns. Selling an appreciated holding can also trigger tax, depending on the jurisdiction and the type of account, and holding coins directly involves different execution and costs from holding an exchange-traded product.

Frequency feeds straight into that bill. A weekly rule trades far more often than an annual one, and in the 2023–2026 test that extra activity showed no benefit in the reported figures.

The published report does not set out how trading costs, spreads or taxes were handled, so its figures cannot be read as after-cost outcomes. Coindoo has not independently replicated them.

For anyone who bought “just a little” Bitcoin, the number worth checking is the share of their wealth the position represents today, and whether they could absorb that share halving. If that number has grown since the purchase, so has the risk — whether or not they ever bought more.

This article first appeared on Coindoo.