What did your account earn? A guide to time-weighted return
Deposits can lift a brokerage balance without producing investment gains. Learn how time-weighted return separates the two, and why the benchmark, dates and drawdowns still matter.
Research · Current account performance
A brokerage balance can rise because the investments gained value, because you deposited more money, or both. Time-weighted return (TWR) removes the effect of external deposits and withdrawals from the return calculation. It helps answer what the investments earned, rather than how much money you moved into the account.
That distinction matters when evaluating a claim that an AI-managed account grew. But TWR alone is not enough: you also need the measurement period, a comparable benchmark, and the losses along the way. Here is how to read those pieces together.
A rising balance can accompany a negative return
Consider a hypothetical account. These amounts are an illustration, not AlpAI's balance, deposits or results. The example assumes that the account can be valued immediately before and after the deposit.
| Stage | Account value | What happened | Return for this interval |
|---|---|---|---|
| Start | $10,000 | Starting value | — |
| Before the deposit | $11,000 | Investments gained $1,000 | +10% |
| Immediately after the deposit | $21,000 | Another $10,000 entered the account | The deposit is not an investment return |
| End | $18,900 | Investments lost $2,100 after the deposit | −10% |
The balance rose from $10,000 to $18,900. Yet the time-weighted return is:
(1 + 10%) × (1 − 10%) − 1 = 1.10 × 0.90 − 1 = −1%.
A performance line starting at 100 would rise to 110, then fall to 99. Adding +10% and −10% to get zero would be wrong: the second percentage applies to a different starting value.
There is another valid way to describe this account. A total of $20,000 was contributed and $18,900 remained, a loss of $1,100. That amount and the −1% TWR answer different questions. One describes the change in the money contributed; the other links investment returns across intervals without treating a deposit as profit. TWR does not make the loss after the deposit disappear.
Withdrawals follow the same principle. When the balance line and performance line diverge, check the timing of cash flows and how the account was valued around them before drawing a conclusion about investment skill.
What AlpAI's published return includes
Our published series uses daily performance records from Interactive Brokers (IBKR). The main measure is in U.S. dollars. It reflects trading fees and dividends received, and is before corporate income taxes.
We publish returns rather than the account balance, deposit amounts or share quantities. The dollar amounts in the illustration above are not disclosures of the live account.
The record describes the whole account. It is not, by itself, the return of a particular stock selection, research rule or AI model. Trading rules changed during the period, so the account should not be presented as the long-term record of one unchanged strategy.
Match the benchmark, dates and currency
An S&P 500 price index excludes dividends. A total-return series includes them. If the account receives dividends while the benchmark only measures price changes, the difference cannot simply be attributed to stock-picking skill.
Check the beginning and end dates, currency, dividend treatment and trading costs. Also distinguish a cumulative return over the stated period from an annualized return. A U.S.-dollar result can differ from the same account's yen result because exchange rates move.
The historical snapshot below used the S&P 500 price index. To compare the latest results, check the benchmark and dividend treatment on the current performance page. The date of the figures and the date the page was updated can differ.
Read the losses along the way
Drawdown measures how far the performance series has fallen from its previous peak. Current drawdown refers to the reporting date; maximum drawdown is the deepest such fall during the measurement period. Neither is the same as the return from the starting date.
Take another hypothetical performance line: 100 → 120 → 90 → 110.
| Measure | Calculation | Result |
|---|---|---|
| Return from the start to the end | 110 ÷ 100 − 1 | +10% |
| Maximum drawdown | 90 ÷ 120 − 1 | −25% |
| Drawdown at the end | 110 ÷ 120 − 1 | About −8.3% |
All three results are true at the same time. The account ended above its starting level, experienced a much larger fall on the way, and had not recovered to its earlier peak.
Use the cash-flow-adjusted performance series for this comparison. An account-balance chart can make a deposit look like recovery from an investment loss.
Keep historical figures with their original definitions
The source explanation first reported the following snapshot. These are historical figures, not the account's current results and not the figures at this English page's publication.
| Historical measure | Value and scope |
|---|---|
| Measurement period | April 14–October 1, 2026 |
| Account cumulative TWR | +17.7%, in U.S. dollars |
| S&P 500 over the same period | +10.0%, price index excluding dividends |
| Difference | +7.7 percentage points |
| Drawdown at that as-of date | −4.5% |
The account return included dividends while this historical benchmark did not. That difference matters when interpreting the comparison. The +7.7-point gap does not isolate the effect of AI or individual decisions.
The source article kept its first-published snapshot when later results became available. We retain it here for the same reason. At that original reporting date, the measurement period was less than a year and the article did not annualize it.
Three checks before trusting a performance claim
1. Identify the measure. Is it a balance, a dollar profit or loss, or a return adjusted for external cash flows? For the return, check how deposits and withdrawals were handled.
2. Align the comparison. Match the period, currency, dividends, costs and cumulative versus annualized presentation. Read the data's as-of date separately from the page's update date.
3. Look beyond the endpoint. Examine both the worst drawdown and the drawdown at the reporting date. A positive final return does not describe the path taken to reach it.
TWR provides a way to describe a past account record under stated measurement conditions. It does not establish that AI was responsible for the result, prove that a research rule works, or show that a return will repeat.
AlpAI Lab publishes research and records from investing its own capital. This explanation is not personalized investment advice, a recommendation to trade a security, or a prediction of future returns. Past research and investment performance do not guarantee future results.