TRADING HONESTLY

I Built a Crypto Signal Service in an Afternoon. That's the Problem.

Satdish · Published 11 July 2026

Yesterday afternoon, between fixing a spreadsheet and making a coffee, I built a crypto signal service. It watches six markets around the clock, and when its conditions line up it sends me an alert that reads exactly like the ones you see in paid Telegram groups: direction, entry price, stop loss, take profit. Formatted plans, professional-looking, delivered automatically.

It took a few hours. Most of that was cosmetics.

I'm not telling you this to show off. The code is genuinely mundane. I'm telling you because once you've built one, you can never look at the signal-selling industry the same way again. If a bloke in Kent can knock one up in an afternoon, what exactly are people paying £50 to £100 a month for?

What a signal generator actually is

Strip away the rocket emojis and here's the machinery: a rule, a data feed, and a messenger. Mine checks whether price is above or below a couple of moving averages on two timeframes. When the answer changes, it takes the current price as the entry and places the stop and target a fixed multiple of recent volatility away. That volatility measure is called ATR, and it's just "how much does this thing normally move." That's it. That's the whole oracle.

The commercial ones are rarely more sophisticated. The most common setup in the wild is a TradingView alert wired to a Telegram bot: indicator fires, bot posts "🚀 LONG BTC @ 64,120 / TP 65,900 / SL 63,100" to two thousand subscribers. Some groups don't even run their own rules; they quietly resell signals copied from other paid groups, a few minutes stale. And "AI-powered" almost always means the same moving-average cross with better marketing.

None of this is illegal machinery. Moving averages are fine. ATR stops are fine; I use that exact recipe myself. The problem isn't the maths. The problem is everything wrapped around it.

The tricks in the format

Multiple take-profits, with TP1 absurdly close. Watch for signals with TP1, TP2 and TP3, where TP1 sits inside the market's normal wobble. TP1 gets hit by noise, the group posts "✅ TP1 SMASHED," and that trade goes in the books as a win, even if price then reverses and takes out the stop for a loss bigger than the TP1 gain. Run that structure deliberately and you can advertise a 90% win rate while your subscribers steadily lose money. One stop-loss can cost more than five TP1s earn. It's the exact risk-reward asymmetry any serious trader would call a mistake, engineered on purpose and sold as a track record.

Results as screenshots. No timestamped record, no position sizing, no list of every signal including the ugly ones. Cherry-picked wins in an image are not a track record; they're an advert.

Volume. Ten signals a day keeps the group buzzing and the winners frequent-looking. Frequency is a retention feature, not an edge. Real edges are rare and boring.

Follow the actual money

Here's the part that reframes everything: for many groups, the subscription isn't the business. The business is the exchange referral link. Sign up to a broker or crypto exchange through the group's link and they earn a cut of your trading fees, often for as long as you trade. They get paid on your volume, win or lose. Sit with that incentive for a second: the more you trade, the more they make, regardless of whether the signals work. Ten signals a day suddenly makes perfect commercial sense.

Further down the spectrum you find pump groups, where the operator buys a thin altcoin, signals the group to buy, then sells into their buying. And in the UK specifically, most of these operations are unauthorised firms making financial promotions to retail customers, which is the FCA's territory. The typical Telegram signal seller isn't just dubious; they're usually non-compliant.

So why did I build one?

Because as a private tool, the same machinery is honest. Mine runs on my own computer, alerts only me, and keeps an unedited history of every signal it has ever produced, including the ones that would look terrible in a screenshot. Nobody pays me to believe in it, so I'm free to treat it as what it actually is: a context tool that says "conditions have lined up, come and look," not an oracle that says "this will work."

That's the entire difference between a legitimate tool and a product that harms people, and notice that it's nothing to do with the code. It's the honesty layer: knowing the exact rules that generate the signal, keeping the losers in the record, and never confusing an alert with a decision. The trade decision still belongs to you. Your levels, your risk, your process. A signal that skips that step isn't a shortcut; it's someone else's incentive wearing your account.

The uncomfortable question to ask any signal seller

Not "what's your win rate." You now know how that number is manufactured. Ask instead: show me every signal you've issued, timestamped, including the losers, with the position sizing you recommended. An honest operator can produce that in thirty seconds. Watch what happens instead.

And if you're tempted by a group because trading feels hard and you'd rather borrow someone else's conviction, that instinct is the thing to work on, not the signal source. It's the same instinct that blows up funded accounts. I know because it blew up mine, and no signal service would have prevented it.

This article is education, not investment advice. Nothing here recommends any trade, service or product. Quite the opposite, mostly.