Momentum Signals

What it shows

Coins that are accelerating RIGHT NOW (not the ones with the biggest 24-hour gain). Our own scanner records the price of hundreds of coins every minute in a short rolling memory and compares the pace of the last dozen and last few dozen minutes with how that coin normally moves. When the pace clearly stands out, the move is recorded as a signal and tracked until it fades. The widget shows moves happening now, coins just below the threshold and the history of closed moves. This describes a tendency in the data: it is not a recommendation and not a price forecast.

How to read it

Start with the four numbers at the top: how many moves are running now, the weekly hit rate and average result, and how many coins we track at all. Then open the "Confirmed" tab; each row is one running move with a mini price path, its peak and how much it has already given back. The "Building up" tab tells you whether things may get busier shortly. "History" shows how previous moves ended, and "Track record" indicates whether the scanner is in a good or a weak period. Read it as a description of market state: when many moves run and last long, the rising tendency is broad; when they flip to "Fading" quickly, the market gives gains back as fast as it makes them.

Parameters

Active now
How many moves are running right now: the coin crossed the pace threshold and is still holding it.
Positive reading: A dozen at once usually means the rising tendency covers a broad market, not a single coin.
Neutral reading: Zero is a normal state of a calm market, not a failure: a signal appears by itself once the pace picks up.
Rising / Fading / Closed
The life stage of a move. Rising = the pace continues. Fading = price gave back part of the way from the peak, or stalled. Closed = the move ended and its result is calculated.
Neutral reading: "Fading" describes the state of the data, not an instruction; it shows the pace is weakening.
Peak and distance from peak
The peak is the highest point of the move measured from detection. Distance from peak says how much of that path the price has already given back.
Neutral reading: A big gap between peak and result means the move was fast and short, which is information about the character of the move.
Building up
Coins about halfway to the threshold. The bar shows how close they are, providing a preview of what may become a signal.
Neutral reading: Appearing on this list is not a signal and very often ends in nothing.
7-day hit rate
The share of closed moves from the past week in which the price rose by at least the required threshold after detection: the peak reached is what counts, even if the move later gave part of it back. What remained at the close is shown separately as “held to the end”.
Neutral reading: With only a few closes this number is very unstable; read it together with the number of closed moves.
Average result and average peak
The average change at close and the average highest change during the move, both measured from the price at detection.
Neutral reading: Past results describe the past and say nothing about the next moves.
To signal
How much of the way is left before the coin reaches the threshold at which we open a move. One hundred percent means the threshold has just been reached and the coin moves from the watch list into the signals.
Neutral reading: The raw percentage change would say nothing here, because every time window has its own threshold. We therefore compute the fraction of the threshold in whichever window is closer, and the row tooltip states how many percentage points are missing and which window it refers to. Eighty percent of the way does not promise the coin will cross the threshold: it can just as well turn back.
24h change
The plain price change of the coin over the last day, the same one shown in the market tables. It plays no part in the signal decision.
Neutral reading: This number and the “to signal” column speak about different things: a coin can have a weak day and still sit just below the threshold, because the threshold is measured over a short window, not over the whole day.
Thin market
A marker on a coin whose 24-hour turnover is below the credibility threshold. Such a move can be the effect of single trades.
Negative reading: Coins with this marker do not enter the main hit rate. Otherwise, they would inflate the picture. We still follow the move to the end and show its result: once we detect it, we do not hide it from you. Those results are counted separately as a second statistic: number of moves, hit rate and median result.
Neutral reading: Confirmation and exit rules are the same as on a liquid market, so such an entry can stay on the list longer than a false start, and its result still does not enter the main track record, we count it separately. In exchange, a thin market can get more patience: a separate setting extends how long such a coin may go without a new high and how long the signal may live, because small coins move in jumps. The selected settings profile drives this, and the thin-market profiles make that time two to three times longer.
Data stalled
The last price reading is older than several scan cycles. We show it instead of pretending an old price is current.
Neutral reading: A dash instead of a number always means missing data, never zero.
Daily signal cap
The maximum number of moves we record in a day. The cap exists so that one turbulent day does not flood the list with dozens of entries about the same thing.
Neutral reading: Once the cap is reached the scanner keeps watching the market, it simply stops adding entries that day.
Coins tracked
How many coins have a short rolling price memory from which we compute pace. The second number says for how many of them the reading is live rather than from a periodic fetch.
Neutral reading: A coin outside this list cannot produce a signal even if it rises sharply, as we simply do not measure it.
Strongest move
The highest peak reached by a single move in the period, measured from the price at detection. It is an extreme value, not a typical one.
Neutral reading: One strong move says nothing about the rest, which is why we always show the average and the full coverage next to it.
Data coverage
A full account of every closed move of the week: how many entered the hit rate and how many dropped out, and why. Without this breakdown the percentage would be a number without a sample.
Neutral reading: Nothing disappears from the history: a move rejected from the count still shows on the list, just with its reason given.
Closed in total
Every move that came to an end over the past week (including the unsuccessful ones and those that ultimately did not enter the percentage).
In the statistic
Moves that met every measurement condition and from which the hit rate is computed. This is the denominator of that number.
Neutral reading: The smaller this number, the less the percentage means, as a handful of closes can shift it by tens of points.
False starts
Moves that fell apart in the first minutes after detection. The pace did not hold long enough to treat them as a started market move.
Neutral reading: A false start is a stumble of our own entry, not a market outcome, so it counts neither as a success nor a failure. It stays visible on the list and here.
Too short
Moves that closed faster than the minimum measurement time. Over such a short stretch the result is more noise than tendency.
Bad data
Moves measured on a suspicious price reading (a gap in the data, a single-exchange spike or a source failure). We do not pretend that was a real market move.
Manually excluded
Closed moves the operator removed from the count after finding a faulty reading. They stay visible in the history together with the reason.
Unconfirmed
Moves that closed before their pace was ever confirmed. We do not know enough about them to score them either way.
Typical range
How far moves of THIS character usually travelled from the moment of detection. We take the median peak of every closed move in the same parameter group over the last two weeks: the median, not the average, so one blow-off does not inflate the picture. The scale ("narrow / medium / wide") is measured against the threshold at which we call a move a hit at all.
Neutral reading: This describes the group's past, not a forecast for this move, that is, a single coin can end up entirely differently.
Risk marker
What works against this group of moves: how often they end in a false start, how widely the results scatter and whether the group holds the target hit rate. A thin market on the coin itself raises the risk regardless of the group, because there a handful of trades can draw the whole move.
Neutral reading: The thresholds come from the scanner settings (hit threshold, target hit rate, minimum turnover), so they move together with them, meaning they are not hard-coded numbers.
A dash instead of a marker
A dash means the group does not yet have enough closed moves to say anything about it. We would rather admit the gap than compute an "average" from three cases, because such a number looks like knowledge but is not.
Neutral reading: The threshold is the same one used by the mechanism that mutes weak groups: if a sample is too small for a decision, it is too small for a promise.
Futures market backdrop
Conditions on the futures market for this coin at the moment of the signal: the cost of holding a long position (funding), the 24-hour change in open positions and the ratio of long to short accounts. The same numbers as on the Derivatives page, computed by one shared function so two screens never disagree about the same coin.
Negative reading: When the reading is older than a few hours or the exchanges report nothing, the row simply disappears. Missing data is silence, not zero.
Neutral reading: This is only the BACKDROP of a move, not its premise: it does not enter the detection thresholds or the numeric assessment of a signal. Expensive longs alongside rising open interest can at most raise the risk marker by one step, because such a market reacts more violently to a turn.
Correlation leader
The coin this one tracked most closely at the moment the signal appeared: its symbol, the strength of the link (0–1) and how much the leader's price has moved since then. The link comes from correlations computed on daily prices, not recalculated per signal. Bitcoin wins ties of comparable strength, because the market reads most moves through BTC anyway.
Negative reading: When two coins share too few days of history, the link is too weak or the measurement is stale, the row does not appear at all. No leader does not mean "this coin moves on its own"; it means we have no measurement worth trusting.
Neutral reading: This is an observation of a historical link, not a forecast: the leader does not enter detection thresholds or the numeric assessment of the move. The record is frozen at detection, so recomputing correlations later never rewrites what the signal saw at the time. With an inverse link, the backdrop worsens when the leader RISES.
Coins in scanner range
How many coins the scanner considers at all at the moment of a signal: the range is set by the global market-cap rank limit in the configuration.
Neutral reading: This is the denominator for every other number: "3 signals today" means one thing across 500 coins and another across 50.
Dropped by the liquidity floor
How many coins in that range fail the minimum turnover and market cap, so their moves cannot become a full signal. On a shallow market a single order can draw a chart nobody could repeat.
Neutral reading: Those coins are still visible as observations, since the filter concerns scoring, not hiding.
Unconfirmed
The move is fresh and awaiting confirmation: it has to hold the pace for a few minutes before we start scoring it.
Neutral reading: The entry is visible immediately but joins the hit rate only after confirmation, as we do not hide it until then.
Operator correction
An entry manually removed from the count because it rested on a faulty reading. It stays in the history together with the reason, so nobody has to take our word for it.
Move from history
A move reconstructed after the fact from historical quotes rather than caught live by the scanner. It describes the market but does not prove we would have caught it in real time.
Neutral reading: That is why, next to the overall figure, we separately show the hit rate computed only from moves caught live.
What the history shows
The history contains every move, including the failed ones and those that ended at a loss; we remove nothing after the fact. The hit rate is computed more narrowly: only from confirmed moves on a liquid market with a trustworthy measurement, and every skipped entry states plainly why it was left out.
Neutral reading: Past results describe what already happened and are not a promise of future moves.
How it is computed
The signal engine is a proprietary CryptoPulsar script: a model learning from many signals at once, from seasonality patterns and market-cycle parameters, tuned on our own data. We publish the entry and fade thresholds and the coin universe; the internal weights, feature selection and the way signals are combined remain our proprietary part.
Neutral reading: The model is tuned continuously, so thresholds and accuracy change over time, which is why we show a day-by-day history.
Sonar band
The narrow band under the header shows the same state as the list, only in motion. The arc on the left pulses to the scan rhythm, and each disc is one open move: the further right, the later the phase (freshly detected, building up, near the peak, fading). A longer tail means a faster pace, a dimmed disc means a drawdown from the peak. Green is an active move, gold is near the peak, ash is fading, pink is a close below zero. When nothing is open, the band shows watched coins as points whose height and brightness grow with the score.
Neutral reading: Motion on the band describes only what the scanner sees in the data. Clicking a disc jumps to that move's row, and the effect itself can be switched off with the icon next to the header.
Notifications
The Pulse Alpha plan can receive a bell and a phone notification about a new move, its fading and its close. Everything is off by default; you switch it on in notification settings, together with thresholds (minimum move, minimum turnover, watchlist only, daily cap).
Neutral reading: A notification describes an event in the data and never contains an incentive to trade.

Example

You see 9 active moves, a weekly hit rate of 54% over 37 closes and an average result of +2.1% with an average peak of +6.4%. Reading: the rising tendency is currently broad, but the gap between peak and result says the moves are short and give back part of the path quickly. If half the rows read "Fading" at the same time, that is a cooling impulse, not an accelerating one.

Where in the app

All methods