Crypto time-weighted returns link investment returns between external deposits and withdrawals, neutralizing their direct effect on measured performance. Money-weighted returns incorporate the size and timing of those flows and describe the investor's capital experience. Report both when funding changes materially, using the same account boundary, currency, costs and dates. A positive time-weighted return can coexist with an investor's dollar loss.
That apparent contradiction is an accounting question before it is a strategy question. If a small balance participates in a strong period and a much larger balance participates in a decline, the portfolio's linked return and the investor's experience will differ. Neither number should replace the other without explaining the question it answers.
Choose the question before the calculation
Time-weighted return, or TWR, asks how the invested portfolio performed across the measured intervals after removing the direct arithmetic effect of outside funding. Money-weighted return, or MWR, asks what return rate reconciles the investor's dated contributions, receipts and ending value.
CFA Institute's discussion of personal investment performance reporting distinguishes these purposes: money-weighted reporting includes contribution and withdrawal timing, while time-weighted reporting removes the effect of those external flows. This is a difference in measurement, not a reason to call one result dishonest.
For crypto research, keep the investment comparison separate. Choosing between a spot holding alternative and a cash allocation is the subject of the strategy benchmark comparison guide. This article addresses what happens when capital enters or leaves the measured portfolio.
Define what counts as external
An external flow crosses the boundary of the portfolio being measured. Buying BTC with cash already inside that portfolio is an internal allocation change. Selling it and retaining the proceeds inside is internal too. A deposit from outside increases capital without creating investment profit.
Use a boundary worksheet before importing transaction history:
| Event | Single trading-account report | Consolidated report including both accounts |
|---|---|---|
| Transfer from your spot account into your futures account | External to the receiving account | Internal transfer; reconcile both sides |
| Cash deposit from an excluded bank account | External contribution | External contribution |
| Purchase of an asset using existing account cash | Internal transaction | Internal transaction |
| Trading fee or perpetual funding debit | Performance cost under the stated policy | Performance cost under the stated policy |
| Asset transferred from an excluded wallet | External in-kind contribution, valued in reporting currency | External unless that wallet is included |
These classifications follow the selected reporting boundary. Do not label every exchange ledger credit a deposit: a realized trading gain and an outside contribution have different meanings. Preserve fees and funding as costs rather than removing them as though someone withdrew capital.
Choose a reporting currency and a consistent valuation policy for cash, assets and liabilities. A USDT balance does not automatically equal the same number of US dollars. For an in-kind transfer, record quantity, receipt time and the reporting-currency value under that policy. A mismatch between transfer valuation and account valuation can manufacture a return around the flow.
One portfolio path, two very different experiences
The following invented example uses two equal 365-day years. It is not market history, a KLINEO result or an expected return. Values are in US dollars. Assume immediate deployment of contributions, no withdrawals and no fees, funding, interest or taxes solely to isolate cash-flow timing.
The stipulated portfolio return is +50% during the first year and −20% during the second. The investor begins with $1,000 and adds $9,000 exactly between those years.
| Date and event | Account value | Investor cash flow for MWR |
|---|---|---|
| 23 October 2024: initial contribution | $1,000 | −$1,000 |
| 23 October 2025: value immediately before contribution | $1,500 | None |
| 23 October 2025: add $9,000 | $10,500 | −$9,000 |
| 23 October 2026: ending valuation | $8,400 | +$8,400 terminal value |
The positive terminal entry represents the value available to the investor at the measurement end; it does not assert an actual withdrawal occurred. Contributions use negative signs because money leaves the investor and enters the portfolio.
The time-weighted calculation splits at the contribution:
TWR = (1,500 / 1,000) × (8,400 / 10,500) − 1 = 20%
This is the cumulative return over two years. Its annualized equivalent is 1.20^(1/2) − 1 = 9.5445%.
The investor nevertheless supplied $10,000 and finishes with $8,400: a $1,600 loss. Dividing that loss by total contributions gives −16%, but ignores how long each contribution was invested. That ratio is not the money-weighted annual return.
For equal annual intervals, solve the investor's cash-flow equation:
0 = −1,000 − 9,000 / (1 + r) + 8,400 / (1 + r)^2
The economically relevant solution with 1 + r > 0 is approximately r = −14.7431% per year. The aligned annual comparison is therefore TWR +9.5445% versus MWR −14.7431%. Most of the investor's dollars encountered the losing year; only the initial $1,000 participated in the winning year.
Without the additional contribution, the same stipulated return path would end at $1,200. With only opening and ending capital and consistent annualization, TWR and MWR would agree. Real deposits can also change execution, capacity or cash drag; TWR does not reverse those economic effects.
Link subperiods; do not average their percentages
A true time-weighted calculation needs valuations around external flows so each subperiod has an appropriate capital base. The GIPS handbook's calculation discussion explains geometric linking and distinguishes valuation at every flow from approximations. Its standards apply within their own scope; citing the method does not establish GIPS compliance for an account or this article.
In the example, averaging +50% and −20% gives +15%, which is neither the cumulative 20% TWR nor its annualized 9.5445% equivalent. The log versus simple returns guide explains the aggregation units behind this distinction.
An exchange snapshot taken at midnight may not support an exact calculation for a deposit credited at 14:17 UTC. During a volatile day, subtracting the deposit from closing equity cannot reveal the return before and after its arrival. Preserve the credit timestamp and flow-time valuation, or label the estimate and its assumptions.
Use dated cash flows carefully for MWR
For irregular dates, a dated internal-rate-of-return calculation can incorporate elapsed time. Microsoft's XIRR documentation defines its cash-flow inputs, uses a 365-day year and truncates dates to whole days. A spreadsheet can reproduce the example with the three signed cash flows −1,000, −9,000 and +8,400 on their corresponding dates.
That day-level convention matters in a market trading continuously: standard XIRR does not distinguish a morning contribution from an evening one on the same date. Document any finer timestamp convention used by another calculation engine. XIRR also requires positive and negative values and can fail to converge; a software error is not a zero return. Alternating contributions and withdrawals can produce multiple IRR solutions, as the GIPS handbook cautions. Inspect the cash-flow schedule and solution rather than accepting a rate without reconciliation.
Keep the funding decision in the review record
Before publishing a return, save the account boundary, reporting currency, start and end timestamps, signed external-flow ledger, flow-time valuations, included costs and annualization convention. Reconcile ending assets and liabilities to the same ledger. Then show dollar gain or loss beside the percentage measures.
If research compares rolling and expanding backtest windows, retain the same cash-flow policy across evaluated segments. Window design changes the training history; it does not make outside deposits into trading gains. Execution and cost assumptions still need the separate checks in the backtest evidence review.
A return gap does not prove poor investor judgment. Funding may reflect income needs, a planned allocation or an operational constraint. Record the reason and authorization alongside the numbers using a decision-focused trading journal. The defensible finding here is specific: the hypothetical portfolio had a positive linked investment return, while the investor's larger later contribution produced a negative money-weighted experience.
Method and limits
This AI-assisted educational article uses the linked primary sources and an original fictional cash-flow example. The arithmetic was checked independently of market data. No live account, trading experiment or KLINEO reporting feature was tested. Reliable measurement still depends on complete flows and consistent valuations, and neither return measure establishes future performance or the suitability of a funding decision.