What do the three numbers on a PipGems alert card mean?

Each card carries the pair and direction with a countdown, then two figures underneath. The first is gauge agreement - how many of 21 indicator readings currently point the same way as the signal, printed as agrees or disagrees with the count. The second is the hourly heatmap, shown as this hour followed by a percentage, which is how often signals on that pair at that hour of the day have finished in the money over the chosen lookback.

Does the app really tell you when its own gauge disagrees?

Yes, and it is the most unusual thing about it. In the session one card reads gauge disagrees 15 of 21 with a red cross, meaning the indicator panel leaned the opposite way to the signal it is attached to. Most signal products present a single confident direction and never surface internal disagreement at all. Being shown the conflict is what lets you decline the trade.

Is the hourly heatmap percentage a probability for the next trade?

No. It is a historical base rate for that pair and that hour over a lookback window, not a forecast for the specific trade in front of you. It also has no sample size printed next to it, so you cannot tell whether the figure rests on hundreds of past signals or a couple of dozen. Read it as a reason to prefer one hour over another, not as a number to size a trade against.

How does 71% compare with measured hour-of-day data?

They are measuring different things. Our own study of 23 years of hourly candles found raw hour-of-day up-rates running roughly 45% to 60%, with EUR/USD peaking near 54.7% at 02:00 UTC across about 5,900 samples. The app heatmap counts signal outcomes rather than raw candle direction, so a higher figure is not automatically wrong - but the gap is large enough that the sample size behind it matters a great deal.

Which pairs should be switched on in the filter?

The ones your broker is paying well for right now. The session sets 10 pairs in the app filter and matches them against the pairs showing the higher percentages on Quotex, so alerts only arrive for assets worth trading. In the frames the open pairs pay 93%, 85% and 82%, and the payout is what sets the accuracy you need to break even.

Do you need the gauge switched on to use the alerts?

No. The floating gauge bubble is optional and the agreement count already appears on the alert card itself, so the extra dial is a convenience rather than a requirement. The bubble is useful if you want to check a pair you have not been alerted on, since arrows on it switch pairs without leaving the broker app.

PipGems Signal Heatmap on Quotex — What “this hour: 71%” Actually Means

Rated 4 out of 5

Most signal apps tell you one thing: buy or sell. This one tells you three, and the third is the interesting one — because sometimes it says the app's own indicators disagree with the signal it just sent you.

That is rare enough to be worth a post on its own. It is also the part most likely to be ignored, including in the session below.

Three numbers on every card

Two PipGems alert cards on Quotex mobile: EURGBP PUT reading gauge disagrees 15 of 21 with hour 71 percent, and EURUSD PUT reading gauge agrees 13 of 21 with hour 65 percent

Two alerts from the same hour, side by side. Read the small line under each one:

Card	Signal	Gauge	Hourly heatmap
EUR/GBP	PUT, 55s left	✗ disagrees 15/21	71%
EUR/USD	PUT, 24s left	✓ agrees 13/21	65%

So each alert carries direction and a countdown, then two independent readings:

Gauge agreement. How many of 21 indicator readings currently point the same way as the signal. A red cross means the indicator panel is leaning the other way.
The hourly heatmap. How often signals on that pair, at that hour of the day, have finished in the money across the lookback window.

Showing the disagreement is genuinely unusual. Most signal products present one confident arrow and never surface internal conflict at all, because conflict does not sell. Printing "gauge disagrees 15/21" on the card is the app arguing with itself in public, and that is what makes it possible for you to decline the trade.

What the app is doing behind that

PipGems mobile app settings: the Live Gauge bubble with a 21-indicator dial, the alert throttle set to one minute, and the pair filter grid with ten pairs selected

Three controls do the work, and they are worth setting before the session rather than during it.

The pair filter. Ten pairs are lit in the grid out of roughly thirty. The rule stated in the video is to match those against the pairs Quotex is currently paying well on, so alerts only arrive for assets actually worth trading. In the frames, the open pairs pay 93%, 85% and 82%.

The alert throttle. Set to at most every 1 minute, with options for every signal, 2 and 5 minutes. On ten pairs this is what stops the phone becoming unusable.

The gauge bubble. A floating 21-indicator dial that sits over the broker app, with arrows to flick between pairs. It is optional — the agreement count is already on the card — and it is most useful for checking a pair you have not been alerted on.

One nice touch: the settings sync with the web version, so changing the throttle on the phone changes it on the desktop too.

The number that needs care

The gauge count is self-explanatory: 15 of 21 is a headcount, and you can see the dial that produced it. The heatmap percentage is different, and it is worth being precise about what it is.

It is a historical base rate for that pair at that hour, over a lookback window. In the session the window is set to a year. It is not a probability for the trade in front of you, and the video's phrasing — "the heat map confirms a 71% probability for this hour" — is looser than the number can support.

The specific thing missing is sample size. A 71% figure built on 400 past signals and one built on 25 look identical on the card. There is no way to tell them apart, and they deserve completely different levels of trust.

For scale, here is what the same question looks like when it is measured properly. We ran 23 years of hourly candles across 19 markets and published the sample size and significance next to every cell:

Measured hour-of-day, EUR/USD	Up-rate	Samples
02:00 UTC — strongest hour	54.7%	5,902
20:00 UTC — weakest hour	45.5%	5,902
Honest range across all markets	45–60%	~6,000 each

Those two things are not measuring the same quantity — raw candle direction is not the same as signal outcomes, and a good signal should beat the raw base rate. But the distance between 54.7% across 5,902 samples and 71% across an unstated number is the distance between a figure you can lean on and a figure you should treat as a hint. Both hours in this session, 71% and 65%, sit above the entire measured range for raw hourly direction.

Our own research post says it plainly: testing 24 hours across many pairs guarantees some cells look brilliant by chance. That is a warning about our data too, and it applies here.

The trade we would not have taken

Two Quotex mobile charts during the session, the EUR/USD chart dropping sharply after a PUT signal and the EUR/GBP chart with the trade running

Both trades in the session closed green. The second one is unremarkable in a good way: the gauge agreed, the heatmap read 65%, the trend was clearly down, and the PUT worked.

The first is the one worth talking about. The gauge said disagrees, 15 of 21 — meaning the indicator panel actually leaned towards a call — and the trade was taken as a put anyway. The reasoning given on camera is: "we do not depend on the indicator, it's just an optional filter."

It won. That does not make it a good decision, and the two facts are independent.

Here is the problem in one sentence: a filter you honour when it agrees and override when it does not has been switched off without being removed. The occasions a confirmation tool disagrees with you are the only occasions it is doing any work at all. If the gauge is genuinely optional, the honest version is to turn it off and stop reading it. If it is not optional, then "disagrees 15/21" is a skip, including on the trades you want to take.

There is a defensible middle: decide in advance that the gauge governs one class of setups and not another, and write that rule down before the session. What is not defensible is deciding per-alert, in the moment, based on how much you like the look of the chart.

What the session gets right

The pair filter is synced to payouts. Alerts only arrive for assets worth trading. This is the single highest-value habit in short-expiry trading and we went through the arithmetic in the payout filter post.
The throttle is set before the trading starts. Ten pairs without a throttle is a phone you cannot use.
The stake is flat. $20 on every trade, no increasing after a loss.
It runs on a demo account. The DEMO watermark is visible through both charts, and that is the correct place to be establishing whether an hour bias is real.
The block is time-boxed. One hour, then stop — which is a far better session structure than trading until something goes wrong.

How we would use the heatmap

As a ranking tool, not a permission slip. When two alerts arrive close together and you can only take one, the higher hourly figure is a reasonable tiebreaker. Using it to justify a trade the gauge is arguing against is asking it to do a job it cannot do.

The strongest combination on the card is the boring one: gauge agrees, heatmap above average, payout high. All three lined up is a smaller number of trades than most people want to take, and that is rather the point.

Why 4 and not 5

The app deserves real credit here. Surfacing its own gauge disagreement on the alert card is the opposite of what this category normally does, the pair filter and throttle are sensible controls rather than decoration, and the session runs on demo with a flat stake and a fixed time box.

Two things hold it back. The heatmap percentage is presented without a sample size, which makes a well-supported figure indistinguishable from a thin one — adding n to that line would cost nothing and would settle the question. And the session's one clear teaching moment, an alert the gauge openly contradicted, was resolved by overriding the gauge rather than skipping the trade. Two winning trades on one demo hour is a demonstration of a workflow, not evidence about it.