How to evaluate a track record (and the 5 ways people fool you with one)

Every strategy seller has a chart that goes up. Your job, before trusting anyone with a cent, is to figure out what was cut out of the chart.

The 5 classic ways to dress up a track record

1. Cherry-picking

Showing the good trades and "forgetting" the bad ones. The subtle version: showing only the good period ("results since March!"), when January and February were a disaster. Antidote: demand the complete history from day one, with no gaps.

2. A backtest disguised as real results

"This strategy returned 300% over the last 5 years" almost always means: it was tested against the past, knowing what was going to happen. An honest backtest is a research tool, not a result. Antidote: ask whether the decisions were published BEFORE the outcome was known.

3. Account survivorship

Open 10 accounts, run 10 strategies, show only the one that survived. Common in signal groups. Antidote: ask how many variants/accounts existed and where the results of the others are.

4. Percentages without risk

"+40% in three months" means nothing without knowing the risk that was taken. With enough leverage, any number is possible, right up until it blows up. Antidote: demand the max drawdown (the biggest fall from peak to trough) next to any return. Return without drawdown is half a truth.

5. Screenshots

An image of a brokerage account can be edited in 30 seconds. Antidote: live data, updated automatically, with visible positions, is worth more than a thousand screenshots.

The trustworthy track record checklist

• Complete from the start, red days included
• Live (decisions recorded before the outcome)
• Verifiable (positions and prices, not just the final percentage)
• With risk metrics (max drawdown, exposure)
• Across varied regimes (not just months of a rising market)
• No convenient "restarts"

A track record that passes this checklist still doesn't guarantee the future, nothing does. But it filters out 95% of the junk, and that's what an informed decision requires.

The final question

If someone asks for your trust (and your money), the question isn't "how much did it return?". It's: "will you show me everything, including what went wrong?" Anyone who hesitates at that question has already answered it.

---

Kamla's track record is public, live, and includes the bad days, because that's how we'd want to be convinced ourselves. See it at kamla.ai/performance.