Crypto portfolio rebalancing restores chosen asset weights after prices or cash flows change them. Calendar rules act on preset dates; threshold rules act when observed allocation drift reaches a stated band. A combined rule checks the band on scheduled dates. Choose the monitoring clock, trigger, trade destination and cost limits separately; none of these rules guarantees a better return or makes an unsuitable allocation appropriate.
Your tokens' weights can change even when their quantities do not. Rebalancing decides how to respond to that drift. It does not decide whether the target holdings suit the portfolio in the first place.
What calendar and threshold rules control
Start with a target allocation, then write the trading rule separately:
| Rule | When to evaluate | What authorizes a rebalance |
|---|---|---|
| Calendar | Preset dates, such as every 28 days | The scheduled date, subject to execution checks |
| Threshold | A specified monitoring frequency | An observed weight reaches a declared drift band |
| Calendar plus threshold | Preset review dates | The date arrives and observed drift meets the stated threshold |
Vanguard's rebalancing guide describes these three approaches. Its conventional investment examples and timing recommendations do not establish a suitable crypto schedule.
A threshold policy still needs a clock. Checking once a day cannot reveal every intraday crossing. A calendar policy can also monitor balances between trading dates, but that monitoring does not authorize an extra rebalance unless the written rule permits it.
Decide whether the portfolio boundary is one crypto account or a broader set of holdings. Rebalancing token weights within a crypto sleeve does not keep that sleeve's share of a larger portfolio constant. The denominator determines which drift you are controlling.
State the band in percentage points
For an unleveraged spot portfolio, value included holdings in one reporting currency at one timestamp. An asset's weight is its market value divided by total included portfolio value. Include assigned cash consistently and value stablecoins at their reporting-currency prices rather than silently assuming they equal cash.
Percentage-point drift is current weight − target weight; its absolute value measures distance from target. A five-percentage-point band around a 60% target reaches 55% and 65%. A relative 5% band is different: 60% × 0.95 and 60% × 1.05 give 57% and 63%.
Write which convention applies, whether equality triggers a trade, and whether any holding can trigger the rule. Compare unrounded weights; a displayed 65.00% might represent a value just below or above the boundary.
Restoring weights also restores a capital allocation, not a fixed amount of economic risk. Equal token weights can have different volatility and common dependencies. A concentration check using HHI summarizes the weights, while portfolio component risk examines how exposures contribute to an estimated risk measure.
Follow the same price path with separate holdings
This original hypothetical example starts with $10,000: 60 units of token A at $100 and 40 units of token B at $100. The 60/40 target is chosen only to make the mechanics visible; it is not an allocation recommendation.
Both policies observe weekly snapshots at 00:00 UTC. The calendar policy trades on day 28 only. The threshold policy evaluates days 7, 14, 21 and 28, trading when either asset is at least five percentage points from target. Both return fully to 60/40 after a trigger.
Token B stays at $100. All prices are invented. Trades execute at the displayed marks without fees, spreads, slippage, quantity rounding or remaining cash. There are no deposits or withdrawals, and the initial purchases are excluded from subsequent turnover. Each reallocation sells one token against USD and buys the other with the proceeds.
| Snapshot | Token A price | Calendar: A weight before decision | Calendar action | Threshold: A weight before decision | Threshold action |
|---|---|---|---|---|---|
| Day 0 | $100 | 60.00% | Start | 60.00% | Start |
| Day 7 | $125 | 65.22% | Wait | 65.22% | Sell $600 A; buy $600 B |
| Day 14 | $150 | 69.23% | Wait | 64.29% | Hold |
| Day 21 | $100 | 60.00% | Wait | 54.55% | Buy $552 A; sell $552 B |
| Day 28 | $110 | 62.26% | Sell $240 A; buy $240 B | 62.26% | Hold |
On day 7, the threshold account has A worth $7,500 and B worth $4,000. Total value is $11,500. A's target value is 0.60 × $11,500 = $6,900, requiring a $600 sale. After buying B, the account owns 55.2 A units and 46 B units.
On day 21, those holdings are worth $5,520 and $4,600. A's weight is 5,520 / 10,120 = 54.5455%, breaching the lower band. Its target value is $6,072, requiring a $552 purchase. The account now holds 60.72 A units and 40.48 B units.
The calendar account still owns its original quantities until day 28. At that point A is worth $6,600 out of $10,600. Returning to 60% means selling $6,600 − 0.60 × $10,600 = $240 of A.
Each policy must carry its own holdings forward. Applying both rules to the calendar account's weight history would erase the effects of the threshold policy's earlier trades.
Compare turnover and drift together
Here, gross traded value counts both non-cash legs: a $600 sale followed by a $600 purchase creates $1,200 of activity. Half-gross turnover divides total buys plus sells by twice the specified capital base. The portfolio turnover guide explains why other reporting conventions can produce different percentages.
For this comparison, the denominator is the fixed initial $10,000, rather than average equity:
| Measure across the 28-day example | Calendar | Threshold |
|---|---|---|
| Rebalancing events | 1 | 2 |
| Buys plus sells | $480 | $2,304 |
| Half-gross turnover on initial equity | 2.40% | 11.52% |
| Largest observed pretrade absolute drift | 9.23 percentage points | 5.45 percentage points |
| Ending value before costs | $10,600.00 | $10,727.20 |
The threshold policy has the smaller maximum sampled drift and trades more. It is not closer at every snapshot: on day 21, the calendar account is exactly on target. Threshold turnover is 2,304 / (2 × 10,000) = 11.52%. It also finishes with a higher gross value on this particular rise-and-reversal path. That difference is an arithmetic consequence of the selected path, not evidence of a repeatable return advantage.
The five-point band did not cap observed drift at five points: the day-21 observation was already 5.45 points below target. The maximum between weekly snapshots is unknown. A gap, stale feed or delayed execution can increase the overshoot.
For a separate fee sensitivity, applying an invented 0.10% charge to the unchanged gross fills gives $0.48 for calendar trades and $2.304, displayed as $2.30, for threshold trades. This is a fee estimate, not an after-fee simulation; deducting fees from the account would change later balances and possibly order sizes. Spreads and execution differences require a separate transaction cost analysis.
A calendar-plus-threshold policy checking only on day 28 would make no trade here: its untraded A weight is 62.26%, inside the band. The monitoring schedule changes the policy even when the target and threshold stay identical.
Separate the trigger from the destination
Crossing a band need not mean returning all the way to target. A policy can specify a destination inside the band, reducing the immediate adjustment while leaving more residual drift. The destination must be declared before the trade.
One documented implementation illustrates the distinction: Vanguard's target-date fund methodology describes a trigger beyond 200 basis points from target and a destination 175 basis points away. Those are that implementation's settings, not crypto parameters to copy. The worked example above uses a different rule: full restoration to target.
Also distinguish restoring fixed weights from changing the targets themselves. A volatility-targeting policy can change exposure as estimated volatility changes; it requires its own inputs and validation. Do not relabel a discretionary target change as routine rebalancing.
Write a policy that can survive execution
Use this worksheet before turning a weight calculation into orders:
| Policy field | What to record |
|---|---|
| Portfolio boundary | Accounts, included holdings, cash and reporting currency |
| Target | Weights summing to 100%, with a version and effective date |
| Observation | Timestamp, frequency, price source and maximum data age |
| Trigger | Calendar dates or drift formula, band, equality and any-asset rule |
| Destination | Full target or an explicit point inside the band |
| Cash flows | How contributions and withdrawals enter values and decisions |
| Execution limits | Fee budget, available balance, minimum order and rounding treatment |
| Completion | Fill reconciliation, residual weights and conditions requiring review |
Incoming funds can reduce the need for sales if directed to underweight holdings. The SEC's asset-allocation and rebalancing guide describes contributions as one rebalancing route and flags transaction costs and possible tax consequences. The guide is general U.S. investor education; the tax treatment of a particular crypto trade depends on the applicable rules and account circumstances.
Recalculate after a contribution arrives because both available funds and the portfolio denominator have changed. A planned deposit is not a confirmed balance. Decide whether its arrival authorizes an immediate review or waits for the normal clock.
Order rules can prevent exact restoration. For example, Binance Spot's filters define quantity steps and minimum or permitted notional amounts. Read the actual venue's current rules rather than assuming a dollar adjustment is executable. Record any residual drift after rounding instead of reporting the theoretical target as achieved.
Validate the decisions before trading
Replay a saved price-and-balance sequence using only information available at each observation. Check decisions, holdings and costs, not just ending value. Include these failure cases:
- A weight exactly at the boundary: confirm that the chosen equality rule is applied to unrounded inputs.
- A missing or stale mark: defer the decision; dropping that holding and renormalizing the rest invents a different portfolio.
- A completed sale followed by a failed purchase: reconcile actual fills and remaining cash before computing another order. A submitted order is not a restored allocation.
- A rapid reversal or persistent trend: inspect repeated trading and the consequences of repeatedly reducing the stronger holding.
Compare candidate policies on turnover, sampled drift, time spent outside bands where the data can support it, unresolved orders and execution costs. Use several distinct paths; one selected reversal cannot validate a schedule or threshold.
Before the first live rebalance, save one valuation snapshot and show exactly which rule authorizes each order, where its size comes from, and how actual fills will update the next decision.