Most signal providers will show you the wins. Almost none will show you the losses, the drawdown, or the months where they were down 12%.
That's why we built Channel Ratings: a public, free directory at [/channels](/channels) where every Telegram (and Discord) signal channel we've ever processed gets a public A through F grade based on real, executed signals over a 90-day rolling window. Grades update nightly. No hand-picking. No paid placement.
This post explains exactly what the grade measures, how to read it, and why it changes the way you should subscribe to providers.
The five inputs that build a grade
A grade is a weighted blend of five public metrics. Every metric is computed from signals the channel actually posted — not back-tests, not screenshots.
| Metric | Weight | What it measures |
|---|---|---|
| Parse rate | 15% | What percentage of posted messages were valid, executable signals (vs. updates, news, "good luck team" posts). |
| Execution rate | 10% | Of the valid signals, how many actually got filled when AutoTrade was on. |
| Win rate | 25% | Closed trades that hit a TP before SL — over the last 90 days. |
| Profit factor | 25% | Gross profit ÷ gross loss. Above 1.5 is healthy. Below 1.0 means the channel lost money. |
| Risk/reward ratio | 15% | Average reward per unit of risk. A 50% win rate with R:R 2.0 still makes money. |
| Max drawdown | 10% | Peak-to-trough equity drop. Punishes blow-up channels. |
Why "parse rate" is in the grade
A channel can post the world's most accurate signals — but if 60% of the messages are unparseable noise ("guys we're up 200 pips this week 🚀🚀🚀"), your copier won't execute them. The grade reflects what you'd actually trade, not what the channel claims to call.
A channel that posts ten clean signals a week consistently beats a channel that posts forty messages where only six are real entries.
The AI-written review
Beside each grade we publish an AI-generated, plain-English review. Claude reads the last 90 days of signals on the channel and writes a paragraph covering:
The review rebuilds nightly, so a channel that was great in Q1 but went off the rails in Q2 will see its review (and grade) reflect that within 24 hours.
How to use the ratings before subscribing
The page at [/channels](/channels) is sortable and filterable. Two recommended flows:
If you want a new provider:
If you're already subscribed to channels:
How channels can improve their grade
We get this question often from signal providers. The honest answer:
The point of all this
Most signal-copier tools just execute signals. None of the others tell you whether the signal channel is any good. That's how traders end up paying $50/month for a $30-channel subscription that loses them $400 — and never realising the channel was the problem.
A public grade, refreshed nightly, with the math shown in the open, changes that. Before you connect a channel, [check its grade](/channels). After 90 days, check it again. Drop the F's. Keep the A's.
Trading is hard enough without the provider lying to you.