Key takeaways
- Channels claimed a 95% win rate on the signals they reported results for. Real prices put the same signals at 63%.
- Of 27 channels with enough results to judge, 17 lost money after a normal spread. Taking every signal from every channel would have lost 347 times the amount risked per trade.
- Of 3,164 signals that hit their stop loss, channels mentioned 40. The other 3,124 were never acknowledged.
- Only one channel combined a clear edge with a clean record. Several of the best-looking ones had deleted 40–66% of their posts.
Search Telegram for forex signals and you’ll find thousands of channels promising 90%-plus win rates, screenshots of profits and “another TP hit” every few hours. Almost none of them publish a record you can check.
So we checked. For a year, our tracking system recorded every signal from 39 Telegram channels the moment it was posted, and tested each one against real market prices: did it reach its entry, did it hit the stop or the target first, and what did the channel tell its subscribers afterwards?
This is what 10,285 signals, posted between 19 October 2025 and 18 October 2026, really did.
The headline numbers
| Channels found and screened | 428 |
| Channels tracked for 12 months | 39 |
| Signals recorded | 10,285 |
| Signals that reached their entry price | 7,719 (75%) |
| Signals on gold (XAU/USD) | 7,074 (69%) |
| Channels with 30+ measurable results | 27 |
| … that made money after costs | 10 |
| … that lost money after costs | 17 |
Claimed win rates are fiction
Most channels post updates: “TP1 hit”, “TP2 smashed”, “running +80 pips”. We matched every update to the signal it referred to and compared the claim with the price.
On the 4,313 signals where a channel posted a result, it claimed 95.0% were winners. The prices say 62.8% were. And that’s the flattering subset: the channels chose which signals to report on.
The gap isn’t mainly invented wins. It’s missing losses:
- Stop-outs never mentioned: 3,124. Of 3,164 signals that hit their stop loss before any target, channels acknowledged 40. Losing trades simply disappear from the feed.
- Targets claimed but never reached: 2,001. The channel announced a target hit, often TP2 or TP3, that the price never touched before the stop.
- Posted after the move: 1,931 (19% of all signals). When the signal went out, price had already passed TP1, having traded through the entry in the hour before. These signals look perfect in screenshots and can’t be traded.
- Edited after posting: 1,629 (16%). The signal was edited more than 10 minutes after it went out, by which time the market had moved and subscribers had acted on the original.
- Never filled: 2,565 (25%). Price never came back to the entry. Channels rarely mention these.
What following them would actually have made
To measure results, we followed each signal the simplest repeatable way: enter at the stated price, use the stated stop loss, close the whole position at the first target, and charge a typical retail spread (1.5 pips on currency pairs, $0.50 on gold). Results are in R, multiples of the amount risked, so a channel trading gold and one trading EUR/USD can be compared.
Of the 27 channels with at least 30 measurable results:
- 17 lost money. The biggest loser, with an audience in the tens of thousands, averaged −0.39R a trade over 493 trades and won only 25% of the time: a loss of 190 times the risk per trade over the year.
- 5 made only a sliver, between 0 and +0.07R a trade. One extra pip of slippage per trade would erase most of that.
- 5 made more than +0.1R a trade. But one of those had only 34 trades, and three of the other four had deleted 42%, 55% and 66% of their posts. We can only check what was still there when we recorded it. Losing signals deleted within minutes never reached us, so those three records are best cases.
That leaves one channel (#05 in the chart) with a clear edge on a meaningful sample and almost nothing deleted: 224 trades, an 89% win rate at the first target and +0.44R a trade after costs. Even it claimed a later target the price never reached 128 times.
And if you’d taken every signal from every channel, risking the same amount each time? −0.047R a trade across 7,415 trades: a loss of 347 times the risk per trade. Leaving out signals posted after the move or edited later doesn’t rescue it: −0.065R a trade.
TP1, TP2, TP3: the target ladder
Most channels post three or more take-profit levels. Followers tend to remember the TP3 screenshots.
Show the data
| TP1 | 53.9% |
| TP2 | 32.9% |
| TP3 | 19.2% |
Of the signals that filled, 53.9% reached TP1, 32.9% TP2 and 19.2% TP3. A channel that posts three targets will usually be able to show a “TP3 hit” a few times a week. That tells you nothing about whether following it makes money.
Five more things we found
Bigger channels did a little better, but not reliably. Across the 27 channels, audience size and results had a weak positive relationship (a rank correlation of +0.36). Two of the largest, each with over 100,000 subscribers, lost about 100R apiece over more than 1,500 trades.
Some channels are the same channel. In two pairs of channels, 99–100% of the signals matched each other’s (same instrument, direction and entry, within 10 minutes): one source, two audiences. One of those pairs is the two 100,000-plus channels above.
Gold dominates. 69% of all signals were on XAU/USD. Gold moves fast enough to produce frequent “TP hit” posts, and its wide price swings make loose stop placement easy to hide.
A few channels don’t use stop losses at all. Overall only 5% of signals had no stop, but three channels posted almost none. Without a stop, a signal can’t be measured, which may be the point.
Deleting is routine. On average, 13% of posts in a channel’s history had been deleted by the time we recorded it.
How to check a channel yourself
You don’t need our system to protect yourself. Before you follow, or pay for, any signal channel:
- Record it yourself for at least a month. Screenshot or copy each signal when it is posted, with the time. Don’t trust the channel’s own summary.
- Check every signal against a chart. Did price reach the entry? Did it hit the stop or the target first? Most charting platforms let you scroll back to the minute.
- Count the silence. If you see far more “TP hit” posts than stop-outs, losses are being left out.
- Watch for edits and deletions. Telegram shows “edited” on changed posts. Note gaps in the message history.
- Work out the result in R, after costs. A high win rate with tiny targets and big stops can still lose money. Our guide to reading a forex signal explains how.
- Treat “VIP” upsells with suspicion until the free channel’s record checks out. See how to spot forex signal scams.
Method
- Finding channels. 428 channels were found through Telegram search for forex and gold signal terms and through links between channels. We tracked 39 that were active and posted signals in a consistent, parseable format. This isn’t a random sample: it leans towards larger, active channels, and channels that shut down before we found them aren’t included. Both biases make the results look better than the market as a whole.
- Recording. Each channel’s history was read for the 12 months to 18 October 2026, with edits and gaps in message numbering recorded. Posts deleted before we read them can’t be seen.
- Signals. Entry (or entry zone), stop loss and targets were parsed from each post, with follow-up updates linked to their signal.
- Prices. Every signal was checked against Dukascopy 1-minute bid prices. Buys enter on the ask and sells exit on the ask, using a typical spread, and a target only counts when price actually reaches it. A signal must reach its entry within 24 hours or it is “never filled”. Anything still open after 72 hours is closed at the market. If one minute touched both the stop and a target, we counted the stop.
- Results. Each trade is closed at the first target, or at the stop or 72-hour mark, and expressed in R after a typical spread (1.5 pips on currency pairs, $0.50 on gold). Results are given for channels with 30 or more measurable trades.
- Anonymity. Channels are shown as numbers, ordered by result. The behaviour this study measures is common across the market, channels change hands and rebrand, and a named ranking would be out of date within months, while the method stays useful.
The full anonymised dataset, channel by channel, is available on request. We plan to repeat this study every year.
Frequently asked questions
Are Telegram forex signals profitable?
Most of the ones we tracked were not. Of 27 channels with at least 30 measurable results over a year, 17 lost money once a normal spread was charged, and only five made more than 0.1 times their risk per trade. Taking every signal from all of them would have lost money overall.
How accurate are the win rates Telegram signal channels advertise?
Not very. On the 4,313 signals where channels posted a result, they claimed 95% were winners. Checked against real prices, 63% were. Channels mostly stay silent about losses: 3,124 of 3,164 stop-outs were never mentioned.
How can I check a forex signal channel before following it?
Record its signals yourself for at least a month, from the moment they're posted, and check them against a price chart. Count stop-outs and signals that never reached the entry, not just the wins it announces. Be wary of channels that edit or delete old posts.
Why don't you name the channels?
Because the findings are about how signal channels behave in general, and naming a few would turn a study into a ranking that is out of date within months. The method below lets you run the same checks on any channel yourself.
This article is for education and information, not investment advice. Trading forex and CFDs on margin is high risk and most retail accounts lose money. Read our risk warning.