We logged 2,933 messages from four gold signal providers for 60 days

By Emanuele Calcina · Published 21 September 2026 · Logging window 22 July to 20 September 2026 · Disclosure: goldsnipervip.com sells a competing gold signal service. Read this with that in mind, then go and recount it.

The short version. We joined the free Telegram channels of five gold signal providers as ordinary members and stored every message they sent for sixty days. Two of them published a losing trade at a rate that does not survive contact with their own signal count. One sent no gold at all. One is more honest than its marketing suggests. And the only claim on this page you should take at face value is the one you can recount yourself.

Why we built a logger instead of writing another list

Search "best gold signals" and you get forty pages of the same article. Ten providers, a five-star rating each, an affiliate link under every one. Nobody joined the channels. The win rates in those tables are copied off the providers' own homepages, which is roughly as useful as asking a restaurant to grade its own kitchen.

So we did the boring thing. On 22 July we joined five public free channels with a normal account and started writing every message to disk: signals, outcome posts, promos, the lot. Sixty days later there were 2,933 of them. The interesting part was never the win rates. It was the arithmetic between what a channel posts when a trade works and what it posts when a trade does not.

The scoreboard

Five checks. Each one is a thing you can confirm from a public channel in an afternoon, which is the only reason they are on the list. Advertised win rates are quoted from each provider's own site and are labelled as claims throughout.

United Kings: 506 wins, 4 stops, and a countdown that never ends

United Kings was the loudest channel in the sample by a distance. 1,978 messages in sixty days, 67 of them gold signals, which works out at 1.11 calls a day. That part is fine. The signals are properly formed, with a stop and two targets on every one.

Then you count the outcomes. 506 posts announce a win. Four posts say a stop was hit. Across 431 outcome messages, not one is a Telegram reply to the signal it scores, so there is no way to walk from a call to its result without taking the channel's word for the pairing. They published no weekly or monthly tally at all in sixty days.

The entry format is where it gets genuinely interesting, and it is the one number in this study we nearly got wrong. Thirty-six of thirty-seven calls give a $10 zone rather than a price, and the channel tells members to layer into it. Measure the reward-to-risk from the worst edge of that zone and you get 0.29 to 1, which is a damning-looking number and an unfair one. Measure it from the midpoint, which is what layering actually produces, and it is 1.00 to 1. From the best edge, 3.50 to 1. The honest reading is the middle one: on a median trade you are risking a dollar to make a dollar, and at that ratio an 85% win rate would have to hold almost perfectly to stay profitable.

SureShotFX: the one that came out better than its marketing

We expected the fixed-format channels to be the weakest and got that backwards. SureShotFX posts 50 gold calls in sixty days, a shade under one a day, and every one of them carries a 100-pip stop and a 200-pip target. That is a flat 2.00 to 1 on all 53 signals where we could measure it, the best structural ratio in the sample, and it is not close.

It also publishes losses. Not prominently, and not as replies to the original call, but inside 21 performance recap posts that deep-link back to the signal being scored. That is a real audit trail and most of the industry does not bother with one. Counting the line items in their free-channel recaps gives 105 wins against 12 losses, a claimed 89.7%, which sits at the top of their advertised 80 to 90% band, not above it.

The catch is in the size of the wins. Most closures in those recaps are partial, somewhere around 83 pips on average, while a loss is the full 100. A 2 to 1 structure that mostly closes at half a target is a 1 to 1 structure wearing a better suit. That is a real criticism and it is also the most transparent channel of the four.

WolfX Signals: honest about losses, brutal on the maths

WolfX was added to the log on 23 September and backfilled to 22 July. It sent nine gold calls in that window, roughly one a week, which is exactly what it advertises for its free tier. All nine are labelled "GOLD VIP SIGNAL", so they are samples of the paid product rather than a thinned-out freebie.

Every one of the nine carries identical numbers.

Not similar numbers. Identical ones, whether gold was trading at $4071 or $4580, with no adjustment for volatility at all. Close at the first target the way most people do and you are risking $8 to make $2, which needs to be right four times in five before it returns a penny.

Now the part that surprised us, because it cuts the other way. WolfX publishes its losing trades more openly than anyone else in this study. On 13 August: "Yes, Gold hit Stop Loss. Second SL of the entire month." On 20 August: "Third SL of the entire month." It announces the loss and then counts it. Its weekly recaps carry the losing days with the negative attached. And its website advertises no win rate at all, which after the GoldSignals.io section below is worth pausing on.

The catch is that its own tallies cannot be added up. On 13 August the channel reported 11 of 13 signals in profit "during August 2026", on 20 August 15 of 18, then on 30 August "8/8 SIGNALS IN PROFITS". Eight of eight cannot be the August total when the same channel had already counted eighteen signals and three stop-losses. No window is ever stated. Across the twelve tallies we logged the claims run from 82.8% to 100%.

Which leaves a thin and specific problem. At 0.25 to 1 you need 80% to stand still, and WolfX claims something in the mid-eighties. If that is accurate it works, with about three points to spare. Nothing available to a free member lets you check it, and thirty-four of its fifty outcome posts refer to VIP trades you never see.

GoldSignals.io: 90% advertised, 72% in their own weekly reports

This one needs no parser, because they do the work themselves and then nobody reads it.

GoldSignals.io sends its actual signals as images, so there were zero machine-readable calls in our log out of 167 messages. What it does send as text is a weekly PIP report, every Saturday, listing each trade for the week with its result, then a stated total. We captured nine consecutive weeks of them, 20 July to 18 September.

Their site advertises 90%. Their own reports, added up, give 72.1%. Count the individual trade lines instead of trusting the stated totals and you get 213 trades at 70.9%, because five of the nine weeks reconcile exactly and four do not. Both readings land in the low seventies.

Here is the part worth saying plainly: 72% is a decent win rate. There is nothing wrong with it. The problem is that it is eighteen points below the number on the homepage, and the evidence gap is published by the provider, weekly, in the channel they invite you to join. Anyone who joined and read for two months already knew.

ForexGDP: no gold, and the cleanest record-keeping in the sample

ForexGDP's free channel sent 15 signals in sixty days and not one of them was gold. The pairs were EURNZD, EURUSD, AUDJPY, USDCAD, CADJPY and USDCHF. This is not a gotcha. Gold sits behind their Supreme tier at $217.19 a month, so a free-channel gold drought is the product working as designed, and anyone arriving from a "best gold signals" list should know that before they join.

On the one check they can be measured on, they beat everyone including us on format: 22 of their 28 outcome posts are replies to the original call. You can click a result and land on the signal that produced it. For a study that keeps complaining about untraceable outcomes, it would be dishonest not to say that ForexGDP has the best traceability of the four.

Where that leaves us, including the parts that do not flatter us

We are the fifth entry in this comparison and we wrote it, so the ranking is worth exactly what you think it is until you check the underlying record. Here is ours, stated the same way we stated everyone else's.

Between 25 May and 4 September 2026 the VIP channel closed 221 trades. 168 won, 45 lost, 8 finished at break-even. That is a 79% win rate and +14,154 net pips, with 15 of 15 weeks in profit. Every one of those 221 rows is on the results page with its date. The free channel scored 86.6% across 149 calls in 2026, and it publishes 1 to 2 signals a day permanently, with no card and no clock on it.

Now the uncomfortable half. Our 79% is self-scored, exactly like theirs. We ran the parser, we set the rules, we published the output. That is a structural weakness in every win rate on this page, ours included, and no amount of transparency language changes it.

What we think is actually different is narrower and more boring than a percentage. Our losing trades are individually listed, not folded into a weekly summary. Our outcome posts are replies to the calls they resolve, so a member can walk the chain backwards. The monthly figure is published even when it is bad. July 2026 came in at 74% over 67 trades and September is running at 69% over 16, both well under the headline. And there is no countdown on the pricing.

Our win rate is also the lowest advertised number on this page. GoldSignals.io claims 90%, ForexGDP claims up to 95%, United Kings claims 85% and over. We claim 79%, and 79% is the figure that came out of 221 trades you can open and count. A lower number you can verify is worth more than a higher one you cannot, and any provider who cannot show you the losing months is showing you a marketing asset rather than a track record.

If you want the short version of this study as a ranked list, it is on our best gold signals page, and our own 221 calls are laid out on the best XAUUSD signals page.

Five checks, ten minutes, any provider

Nothing here needs a logger. Join the free channel of whoever you are considering and run these before you pay anybody, including us.

1. Count the stop-loss posts. Scroll back thirty days and count wins, then count losses. A channel at a genuine 80% should show you roughly one loss for every four wins. United Kings showed four in sixty days against 506 wins. A channel with no visible losses is not a channel with no losses.

2. Check whether results reply to signals. Tap an outcome post. If it does not thread back to the original call, nobody can reconstruct which trade it scored, and that is usually the whole point of the format.

3. Read the entry format, not the win rate. A $10 zone with a $4 stop is a different trade depending on where you fill. Work out the reward-to-risk from the middle of the zone, because that is what layering actually gets you.

4. Add up their own published tallies. GoldSignals.io publishes everything needed to contradict its own homepage, every Saturday. Providers who post weekly numbers are handing you the audit for free. Do the addition.

5. Watch the pricing for sixty days. If "2 spots left" is still there two months later, you have learned something about the operation that no win rate will tell you.

Run those five on this page too. We published the counts, the window, the parser definition and the fact that we sell a competing service, which is enough rope to hang us with if the numbers do not hold up.

Questions we get asked about this study

Which gold signal provider is best in 2026?

On the five checks in this study, Gold Sniper VIP scores highest, and we wrote the study, so treat that as a claim to test rather than a verdict. The defensible version: we publish 221 individually listed closed trades at 79% including 45 losses, and our outcome posts reply to the calls they resolve. SureShotFX is the strongest competitor in the sample on transparency and has the best structural reward-to-risk at 2 to 1. ForexGDP has the cleanest outcome traceability but sends no gold on its free channel.

Are these win rates verified?

No, and nobody else's are either. Every percentage on this page is self-reported by the provider that published it, including ours. What we verified is narrower: how many signals each channel sent in sixty days, how many outcome posts announced wins against losses, whether outcomes link back to their signals, and what each provider's own tallies add up to.

Why does GoldSignals.io advertise 90% when its weekly reports average 72%?

We cannot answer for them. The figures are theirs: nine consecutive weekly PIP reports between 20 July and 18 September 2026 state 204 trades, 147 wins and 56 losses, which is 72.1%. Counting the individual trade lines gives 213 trades at 70.9%. The reports are posted publicly in their own channel every Saturday.

Is United Kings a scam?

That is not a claim we make or need to make. What the sixty-day log shows is a channel that posted 506 win announcements and 4 stop-loss announcements, published no periodic tally, sent no outcome post that replies to its signal, and ran a 'only 2 spots left' promotion on all 61 days logged. Draw your own conclusion, and note that we sell a competing service.

How many gold signals a day should a provider send?

Fewer than most of them advertise. The measured rates here were 1.11 a day for United Kings and 0.83 for SureShotFX, and our own VIP channel runs at around 5 a day with 1 to 2 on the free channel. Anything advertising 10 or more gold signals a day is describing volume, not selection, because gold does not produce that many clean setups in a session.

Can I reproduce this study?

Yes, and that is the point of publishing the method. Join the five public free channels, scroll back to 22 July 2026 and count. The window is 22 July to 20 September 2026, the parser rules are described above, and the signal, outcome and promotional counts are all recountable by hand from messages that are still in those channels.

Disclosure and method. goldsnipervip.com sells a competing XAUUSD signal service, so this comparison is not neutral and is not presented as neutral. Competitor channels were joined as ordinary members through links published on each provider's own website, and only public free-channel messages were stored. Every competitor win rate quoted is that provider's advertised claim, not a verified figure. Trading carries risk and past results do not predict future results.