Can a Positive Average Return Still Leave Your Portfolio Losing Money?

Summary

A positive average return does not necessarily mean a portfolio grew. In a hypothetical two-month investment, a 30% gain followed by a 25% loss produces an arithmetic average of +2.5% per month—but turns Rp10 million into Rp9.75 million, a cumulative loss of 2.5%.

For investors reading Indonesian-stock Screener studies, the first question is what the average describes: separate stock observations, successive portfolio periods, or actual invested capital. Those are different denominators, not interchangeable performance claims.

Key takeaways

• Add returns and divide by their count to describe an arithmetic average; multiply successive portfolio growth factors to calculate compounded wealth.
• An average across overlapping stock signals is not automatically a portfolio return.
• Position weights, cash, costs and the timing of capital use must be specified before a shortlist becomes an investable simulation.
• A smooth average conceals the path. Read cumulative return alongside drawdown and the underlying period-by-period record.

One portfolio, two months, three different answers

This is an original hypothetical calculation, not an IDX backtest or a Trade Inc performance result. Assume the entire account stays invested, returns include all investment proceeds, and there are no deposits, withdrawals, fees or taxes.

The arithmetic average is (30% − 25%) ÷ 2 = +2.5% per month. That correctly describes the average of the two observations, but it does not reproduce the ending account balance.

Cumulative return is 1.30 × 0.75 − 1 = −2.5% over two months. The second loss applies to Rp13 million, not the original Rp10 million.

The equivalent constant monthly growth rate is √0.975 − 1 = approximately −1.258% per month. This is the geometric mean. It reaches the same final value if applied twice; it does not suggest the account actually followed a smooth path.

Will Morrison’s methodological discussion published by CFA Institute explains why linked wealth changes use a product and geometric averaging rather than a simple sum. The distinction is mathematical, not an additional fee charged to an account. Read the original discussion.

The same observations can describe a different portfolio

Now change the capital arrangement. Invest Rp5 million in each of two stocks at the same time for the same holding period, with one returning +30% and the other −25%. Assume no rebalancing or costs.

The first position ends at Rp6.5 million; the second at Rp3.75 million. Combined wealth is Rp10.25 million: +2.5%. Here the equal-weight arithmetic average is the portfolio’s one-period return.

There is no contradiction. In the first example, all capital experiences both returns successively. In the second, each half of the capital experiences only one return. A report that omits timing and weights makes these different economic arrangements look identical.

Why a Screener’s average cannot simply be compounded

Suppose a Screener identifies five stocks each day and measures each selection over ten sessions. The observations from adjacent dates overlap. They do not represent successive ten-session returns earned by the entire account.

Compounding every observation would implicitly reuse capital without establishing whether it was available. Averaging them instead can still be useful for studying signal outcomes, provided the report labels the result as an average per selection.

To move from that study to a portfolio test, ask for:

1. Allocation: how much capital each signal receives and what happens when available cash runs out.
2. Overlap: whether repeated signals in the same stock create new positions or are ignored.
3. Execution: when orders can fill, what price is used, and how illiquid or unfilled orders are treated.
4. Accounting: how open positions, distributions and idle cash enter daily account value.
5. Costs: the applicable commissions, spread, slippage and taxes, with no double counting.

Fees reduce investor returns; the SEC’s investor bulletin explains their cumulative effect. It is general guidance, not a schedule of Indonesian brokerage charges. Investor.gov fee guide.

Ending wealth does not tell the whole risk story

The first example ends down 2.5%, but its month-end peak-to-trough decline is 25%, from Rp13 million to Rp9.75 million. With no daily observations, the largest intramonth decline is unknown.

Reversing those two returns leaves the same final balance under the stated no-cash-flow assumptions. That does not mean timing never matters: contributions, withdrawals, changing exposure or trading constraints change the problem.

Nor does a short positive compounded result establish repeatability. A useful report retains the dated account-value series, compares a benchmark over the same window and return basis, and identifies missing observations. Annualizing a short sample is a calculation, not evidence that the rate will persist.

Limitations

These examples isolate arithmetic and capital allocation. They contain no market observations, recommended securities, new backtest or tested product capability. They exclude leverage, cash flows and implementation costs; actual results require those details. The general formulas here are not a complete performance-reporting standard.

Conclusion

Before accepting “average return,” ask: average of what, over which dates, using how much capital? A selection-level average can support a research finding. A portfolio-growth claim needs a dated capital record that reconciles opening wealth, cash movements and ending wealth.

If that reconciliation is absent, retain the statistic as descriptive evidence—not as the return an investor could have earned.

Related research

• Can a strategy with a 70% win rate still lose money?
• How stock splits affect price and return comparisons

Published by Trade Inc Research, part of Trade Inc. Educational analysis, not personalized investment advice. Methodology reviewed on 27 September 2026, 08:00 WIB; no live market data used.

This article is for research and information only. It is not personalized financial advice or an instruction to trade.