How to Vet a Signal Service in Ten Minutes.
By CryptoTraders · Strategy · 2026-09-08
The signal-service industry has a structural honesty problem: the product is easy to fake, the marketing is screenshots, and the audience is hopeful. None of that means every service is a scam. It means the burden of proof sits with the seller, and you should know exactly what proof looks like before you pay anyone. Here is the ten-minute due-diligence process we would run on any crypto signal group, including ours.
Are crypto signal groups legit?
Some are. Most cannot prove it, and the reason is the business model rather than the people. A signal group gets paid when you subscribe, not when you profit. Many also earn exchange referral commissions on your trading volume, which rewards them for more trades and more leverage regardless of how those trades turn out. The cheapest way to sell a subscription is a feed of winning screenshots, so that is what most marketing looks like.
At the far end are outright schemes: "VIP" groups that pump thin coins into their own subscribers, paid groups that post the same calls as a free channel a few hours later, and accounts that impersonate a well-known trader. Both the SEC and the CFTC have published investor alerts about trading schemes promoted through social media and chat apps. The checks below will not catch every bad actor, but a service that passes all five is very unlikely to be one of them.
Why screenshots prove nothing
Every signal service on the internet can show you a winning trade. Screenshots are free, selection is invisible, and the losers simply never get posted. If you have ever wondered why every Telegram channel looks profitable while most of its subscribers are not, that is the whole answer: you were shown a highlight reel and sold it as a season.
It gets worse than selective posting. Messages can be edited after the fact to tighten a stop or move an entry. Calls that went wrong can be deleted. And the oldest trick in the industry is running several channels with opposing calls, then promoting whichever one happened to be right. None of that is visible from a screenshot, which is why a screenshot is marketing, not evidence.
The five checks
One: complete history. Ask where the full trade log lives: every signal, timestamped, losers included. If the answer is a highlights channel or a monthly PDF, the history is curated, and curation is the problem. What you want is a running record you can scroll backward through bad weeks.
Two: risk-normalized accounting. Results in percentages or dollars without stated risk are theater: a +80% month at unstated leverage is not a result, it is a lottery ticket that happened to cash. The honest unit is R, profit measured against the risk defined at entry, and any service that cannot express its record that way has not measured itself. R-multiples make a careful service and a reckless one comparable.
Three: faithful execution modeling. Ask how outcomes are scored. Against the alerted entry and stop, at prices a subscriber could plausibly get, with the stated management rules applied mechanically? Or reconstructed afterward at best-case fills? The gap between those two methods is routinely the entire claimed edge.
Four: live continuation. A record that stopped updating is a record that stopped flattering. What you want is performance tracked forward, in public, updating as trades close, so the provider committed to the number before knowing the outcome. This is the single hardest check to fake, which is why so few services offer it.
Five: drawdown disclosure. Every real strategy has losing streaks; the math guarantees it. A service that shows no red is showing you fiction. Find the worst stretch in the record and ask whether you could have sat through it. If they will not show you the worst stretch, they have answered a different question.
| Check | What to ask | Red flag | What good looks like |
|---|---|---|---|
| Complete history | Where is every signal, losers included? | Highlights channel or monthly recap | A scrollable, timestamped log |
| Risk-normalized results | What unit are results measured in? | Percent gains at unstated leverage | Results in R against a defined stop |
| Execution modeling | How is each outcome scored? | Best-case fills reconstructed later | Alerted levels and rules applied mechanically |
| Live continuation | Can I watch it update as trades close? | A static PDF or an old spreadsheet | A public record that updates on its own |
| Drawdown disclosure | What was the worst losing stretch? | No red weeks anywhere | Losing streaks visible in the record |
Why win rate alone tells you nothing
The most quoted number in signal marketing is win rate, and on its own it is close to meaningless. A 90% win rate with tiny targets and wide stops can lose money steadily, because one loss erases nine wins. A 40% win rate can be excellent if the winners are three times the size of the losers. Without the size of wins and losses in the same unit, the percentage cannot be judged.
Our own record shows why. As of October 20, 2026, Momentum Algo had 235 wins and 236 losses, a win rate of almost exactly 50%, and stood at +116.0R. A coin flip on hit rate, clearly positive on expectancy, because the average winner is larger than the average loser. Any service quoting a win rate without the matching R figure is asking you to judge it on the one number that cannot tell you whether it works.
How execution modeling inflates results
This is where honest-looking records quietly go wrong. Each of these practices is common, and each one makes a record look better than anything a subscriber could have achieved:
- Scoring an entry range at its best price, when most subscribers filled somewhere in the middle.
- Counting a trade as a win when a target was touched, even if the stop was hit first on the same candle.
- Scoring the outcome at the furthest target reached while ignoring the stop that the stated rules would have moved to breakeven.
- Leaving out signals that never filled, but only when they would have been losers.
- Quietly changing stops or targets after the alert, then scoring against the edited levels.
To test it, take three closed trades from the record and rescore them yourself from the original alert, using the exchange's historical candles. If your numbers match theirs, the accounting is honest. If you cannot find the original alerts to check against, that is your answer.
Red flags that end the evaluation early
- Guaranteed returns, or any phrase like "risk-free".
- Urgency: limited spots, countdown timers, a price that rises at midnight.
- Lifestyle marketing in place of a record: cars, watches, beach laptops.
- Pressure to sign up for a specific exchange through their referral link before you can see anything.
- Requests to send funds to a person, or to hand over exchange API keys with withdrawal permissions.
- A record that starts recently, after a rebrand, with no history before it.
The meta-signal
The strongest tell is not in the data but in the posture. Services with real edges compete on verification, because verification is the moat competitors cannot screenshot. Services without them compete on urgency. The moment the pitch shifts from evidence to pressure, the evidence has told you what it contains.
The ten-minute vetting process
- Find the full trade history. If you cannot locate it within two minutes, stop here.
- Check the unit. Results should be in R or tied to a stated risk per trade.
- Rescore three closed trades yourself from the original alerts.
- Find the worst drawdown in the record and decide honestly whether you would have stayed subscribed through it.
- Check that the record is still updating, ideally with trades that closed this week.
- Read the pitch for urgency and guarantees. Then decide on the evidence alone.
Applying it to us
We built our public algorithm track record to pass this list rather than argue with it. Every trade from BTC Scalp Algo, Momentum Algo and Nexus Algo is logged, and since June 1, 2026 every closed trade is on the public ledger with full per-trade detail, losers included. Each trade is tracked against live exchange prices from the alerted entry to resolution, with the same management rule members see in the alerts applied mechanically: the stop moves to breakeven after the first target, and the outcome is scored in R by the furthest target actually reached.
As of October 20, 2026, the combined record stood at 549 wins and 525 losses across 1,074 trades, a 51% win rate and +265.0R, an average of about +0.25R per trade. The red stretches are on the public equity curve too: Momentum Algo has given back more than 20R from a peak before recovering, and you can see exactly when. Apex Algo, the newest, shows results accumulating rather than an early number, because a handful of trades is not a sample. Figures update as trades close, and past performance does not guarantee future results.
The human side is held to the same standard. Every analyst and member call is captured the moment it is posted in Discord and tracked from entry to close on the Trade Calls dashboard, wins and losses alike, with full multi-year records per trader for members to inspect. Run the five checks on us first. Then run them on everyone else.
Educational content, not financial advice. Trading leveraged products carries a high risk of loss.