Open almost any trading bot sales page and you will find the same picture: a green line that starts in the bottom left and ends in the top right, smooth, steep and reassuring. It is the most persuasive image in the business, and on its own it is worth nothing. A track record is a claim, and most that you will see are built to survive a glance and nothing more.
Here is how to look past the glance. These are the specific tells of a doctored or cherry-picked record, and what a record you can actually check looks like instead. No single tell is proof of fraud, but the more of them a fake record carries, the less it is telling you and the more it is selling you.
The one thing to remember
A track record you cannot reproduce is not evidence, it is a marketing asset. If there is no account size, no contract count and no way to run the same test yourself, a beautiful curve tells you exactly nothing about what the system does.
Tell 1: an equity curve with no pain in it
Every real strategy has drawdowns, the stretches where the equity curve falls from a high and stays down until it recovers. They are not a defect, they are the cost of the edge. A curve that rises in a near-straight line with no dips is not a sign of a great system but a sign that the drawdown has been hidden, smoothed or cropped out of the frame.
Ask one question of any curve: where is the worst drawdown, and how big was it in dollars? A record that cannot answer, or answers with a percentage too small to survive a bad week, has removed the most important number on the page. Why it matters more than the profit is in what drawdown is.
Tell 2: a number with no account behind it
"Up $180,000." On what? A profit figure means nothing without an account size and a contract count beside it. $180,000 on a $50,000 account trading one contract is one claim; the same figure on a $2,000,000 account, or ten contracts, is a completely different and far less impressive one. Leaving out the denominator is not an oversight, it is the trick.
The same goes for contracts. Futures results scale almost linearly with position size, so any result can be inflated by trading more contracts, along with the risk. If a record does not state the account size and the contract count, you are looking at half a fraction, and half a fraction rounds to nothing.
Tell 3: "live results" that are quietly a backtest
There is a world of difference between a simulation over historical data and a real account with real money and slippage, and both are legitimate when labeled honestly. The problem is the record that blurs them: the vague word "results," a backtest dressed up with a live-looking dashboard, or a live account that quietly began the moment after a bad month ended.
A clean record says plainly what it is. If you cannot tell within a few seconds whether a claim is hypothetical or real, treat it as hypothetical and discount it, then read what a backtest is so the distinction becomes second nature.
Tell 4: a track record that starts on a good day
Give anyone a long enough history and they can find a seven-month window that looks extraordinary, even inside a mediocre system. Cherry-picking a flattering start date is the oldest trick in the book, invisible unless the full history is shown. A record that begins in a suspiciously strong month, or quotes "since inception" numbers without letting you see the month-by-month path, is choosing the frame for you.
The tell within the tell is a missing losing month. Real systems have them. A record showing twenty-four green months in a row has either found a genuine miracle, which almost never happens, or quietly dropped the red ones. Look for the losses; their absence is louder than their size.
See what a record you can check looks like
Account size, contract count, the full period, the worst drawdown and the exact steps to run the same test yourself. That is the whole point of the page.
Tell 5: a win rate and a profit factor that cannot both be true
Two numbers get advertised more than any others, and they constrain each other. Win rate is the share of trades that make money; profit factor is gross wins divided by gross losses. A high win rate paired with a high profit factor and a tiny drawdown essentially never survives contact with a live market, because it usually means one of two things: the losers are hidden by a martingale that doubles down until a win erases them, or the numbers are invented.
A believable record often looks unimpressive. The system on this site, in backtest, wins only 46.2% of its trades and posts a profit factor of 1.58: it loses more often than it wins and makes money because the average winner is larger than the average loser. That is what an honest edge tends to look like, modest and lopsided, nowhere near the 90%-win-rate fantasy. The relationship between these numbers is unpacked in risk, reward and win rate.
Hypothetical performance. The Rentabilio figures used here as an example come from a backtest over historical data, not a live account. Simulated results are prepared with hindsight, carry no financial risk, and cannot fully reflect real execution or slippage. Past performance, real or simulated, does not guarantee future results.
Tell 6: a record you are not allowed to reproduce
This is the one that settles it. A real backtest can be handed to you as a strategy you load into your own NinjaTrader 8, run over the same dates, and check against the published numbers. A fake one cannot, because there is nothing behind the picture. If a seller will not name the platform, instrument, period or settings, and gives you no way to run the test, the record is not verifiable by design.
Reproducibility is the whole game. It is the difference between "trust this screenshot" and "here is the report, run it yourself and watch it print the same total." A number you can regenerate is evidence. A number you cannot is decoration with a dollar sign on it.
What a record you can actually check looks like
Turn every tell around and you get a checklist, and the one below is it. Above all, a record you can trust can be reproduced, which is why this site publishes the gross figure of $274,406 over 88 months: gross is the number you can regenerate in the NinjaTrader 8 Strategy Analyzer, which makes the whole claim checkable rather than merely attractive.
| What to check | Fake record | Verifiable record |
|---|---|---|
| Drawdown | Not shown, or a near-flat curve | Stated in dollars ($4,379) and visible on the curve |
| Account size | Missing | Stated ($50,000) |
| Contract count | Missing | Stated (1 micro contract) |
| Live or backtest | Ambiguous "results" | Labeled plainly as a backtest |
| Period | A short, flattering window | Full history, day by day (88 months) |
| Costs | Only net, no assumption stated | Gross published, cost assumption stated |
| Reproducible | No inputs, no way to run it | Load it into the Strategy Analyzer and rerun |
None of this guarantees a system will make money, and a verifiable backtest is still only a backtest. But it moves the claim from "believe me" to "check for yourself." If you are weighing one system against another, the full method is in how to choose a trading bot, and the broader catalog of dishonest tactics is in trading bot scams and red flags.
Frequently asked questions
Is a smooth, always-rising equity curve a good sign?
Usually the opposite. Every real strategy has drawdowns, so a curve that rises in a near-straight line with no meaningful dips normally means the drawdown has been hidden, the timeframe cherry-picked, or the result is a curve-fit that will not repeat. Ask any equity curve where its worst drawdown is and how large it was in dollars; a record that cannot answer has removed the most important information on the page.
How can I tell a backtest from live results?
Look for an explicit label, and be suspicious when you cannot find one. A backtest is a simulation over historical data, while a live record reflects real orders, fills and slippage. Honest records say which they are in plain language; doctored ones use the vague word "results" or dress a simulation up to look live. When the distinction is unclear, assume the figures are hypothetical and discount them.
Can a track record be technically true and still misleading?
Yes, and most misleading records are. Cherry-picking a flattering start date, showing only winning months, omitting the account size, or quoting a percentage return without the drawdown all mislead with numbers that are individually accurate. The defense is not to check whether each figure is true but whether the whole record is complete and reproducible. Missing context is the tell, not a false number.
What single question exposes most fake records?
"How do I reproduce this myself?" A verifiable system can be loaded into the same platform and run over the same dates until it prints the same total, which is why this site publishes a gross figure you can regenerate in the Strategy Analyzer. A record with no platform, instrument, period or settings cannot be reproduced, because there is nothing behind the picture. If the honest answer is "you cannot," you have your answer about the record too.
Does a verifiable backtest mean the system will be profitable?
No. Reproducibility proves the past result is real and checkable, not that the future will resemble it. A backtest can be honest and complete and still fail going forward, because markets change and past performance, real or simulated, does not guarantee future results. Verifiability lets you judge a claim on evidence instead of trust, which is necessary but never sufficient.
In short: a fake trading bot track record sells you a smooth green curve while hiding the account size, the contract count, the drawdown and the losing months, and it never lets you reproduce the result. A record worth trusting does the opposite: it states what it is, shows the pain along with the profit, and hands you the inputs to run the same test yourself. Publishing a gross figure of $274,406 over 88 months is not a boast but an invitation to check, and being checkable is the only property that separates evidence from decoration.
Sources and further reading
- CFTC, "Customer Advisory: Be Cautious of Trading Systems and Robots."
- CFTC Regulation 4.41, hypothetical performance results disclosure requirements.
- NFA Compliance Rule 2-29, "Communications with the Public and Promotional Material."
