ghost|path is a reading instrument for scalpers. It watches the run forming on your screen this minute, finds the moments in market history that traced the same shape, lays those old moments straight over the one forming now, and shows you where each of them travelled next. This manual teaches the whole instrument from nothing: what every mark on the chart means, why it is there, and how to read it, with a worked example on every idea. It never tells you to buy or sell. You read the rhyme; the decision stays yours.
Charts rhyme. The same handful of machines write them every day, so a shape you are watching right now has almost certainly been written before, on some coin, on some afternoon in the past. ghost|path is built on that one fact. It is a memory for the market: it takes the run forming on your screen this minute, searches history for the moments that looked the same, and draws what came next. Nothing about it is a forecast. It shows you where your shape has lived before, and you decide whether the rhyme is worth anything.
Everything begins at echo.trading. The promise is the whole product in one line: the chart remembers. When your run reaches a shape it has met before, the instrument finds those earlier moments and draws where they went next.
The grey line below is your chart up to now. The dot marked NOW is this moment. The purple line to the right of it is the closest echo — a real run from history that matched your shape, redrawn from now so you can see where it travelled afterwards. That is the whole idea before any of the detail: your run, and its memory of where this shape went the last time it appeared.
A ghost is a memory, not a prediction. It says “a run that looked like yours went this way afterwards”, which is a fact about the past, not a claim about your future. The value of the instrument is that it puts that memory in front of your eyes instantly and lets your own judgement do the rest. It never says buy and never says sell.
ghost|path lives on one page, the board at echo.trading/board.html. The chart sits in the middle with the memory drawn over it. Below is the picture you meet again and again: your live candles, the matched run laid over them (the solid purple line hugging the candles), and, to the right of now, the forward path the old run took. The rest of this manual walks through every mark on it, in the order you would meet them.
Suppose you are watching AAVE on the five-minute chart and it has carved a slow drift lower over the last several hours. On its own that is a shape you have seen a thousand times and it tells you nothing. ghost|path takes those hours, searches every coin’s entire history for the runs that traced the same path, and finds that this shape has appeared before — on TIA back in May, on CRV, even on BTC in March. A rhyme is a rhyme wherever it happened, so it does not care which coin wrote it. It lays the strongest of those old runs straight over your live candles and draws the forward path each of them took. You are no longer staring at a lone drift. You are looking at a handful of times the same shape happened before, and what the market did with it.
The instrument works the same way a careful trader works by hand, and it is worth learning as three plain steps. First it finds the matches. Then it lets you prove a match with your own eyes by laying the old run over the new one. Then it confirms whether the world around the two runs looks the same as well. Every screen on the board is one of those three steps, so once you know the law you always know where you are.
The engine never sleeps. Every minute it takes the current run on each coin and searches the stored history of every coin it tracks for the windows that traced the same shape. The search is blind to which coin and which day a match came from. Its results arrive as the live rail across the top of the board, sorted best-first, so the strongest rhymes in the whole market sit on the left.
A list of matches means nothing until you can see one. When you pick a match, the instrument draws that old run straight on top of the run forming now, lined up so both end at this moment. This is the part you judge with your own eyes. If the two lines hug each other tightly the rhyme is real; if the old line wanders away from your candles it is a loose match, and you treat it with suspicion. The machine finds the candidate; you are the one who proves it.
Two runs can trace the same shape while the situations around them are completely different. The confirm step lines up the context of the old run against the context of yours: where price sat in its range, whether it was resting on a pool of orders, and how the tape was behaving. When those agree, the rhyme is deeper than shape alone. When they do not, you have a shape that looks alike but sits in a different world — exactly the kind of look-alike worth doubting.
Find gives you candidates, prove lets your eye test the shape, and confirm tests the world around the shape. A match that passes all three is worth your attention. A match that only passes the first is a coincidence with good marketing.
You do not need to understand the whole instrument to start reading it. This chapter gets you from opening the page to reading your first real overlay; the later chapters fill in the why.
Open the board and look at the strip labelled echoes live now. Each chip is a coin that currently has a convincing match, ordered so the strongest rhyme in the market is on the far left. On each chip you get three things: the coin, the timeframe the match was found on (1m or 5m), and a similarity number.
Click any chip, or choose a coin from the selector at the top. The chart jumps to that coin, draws the best match over the live run, and frames the view for you. Underneath, the FIND cards show the top three matches, each with the coin it came from, when it happened, its similarity, and a small chart of the old shape against yours. Tap a card to lay that particular match over the live run.
The number, such as 0.98, measures how closely the old shape and your shape line up, on a scale where 1.00 is a perfect tracing. Around 0.95 is a strong, tight overlay; around 0.80 is a loose one you should eye carefully. It is a measure of visual likeness and nothing more, so read it as “how good is the rhyme”, never as a probability that a trade will work.
If no coin is rhyming convincingly, the rail says so plainly and the engine keeps hunting every minute. An empty rail is an honest answer, not a fault. The instrument is built to stay silent when there is nothing worth showing, which the method chapter covers in full.
The chart carries several layers at once, and each one has a job. This chapter takes them one at a time, so that by the end nothing on the screen is a mystery.
The engine does not find just one match, it finds the best three at once. Ghost 1 is the closest, drawn as the bright solid line. Ghosts 2 and 3 are the next two, drawn as fainter dotted lines. They start hidden so the screen stays calm, and their legend chips are the switches that bring them in. Turning them on is worth doing when you want to see whether the three best memories agree about where the run travels next: three echoes pointing the same way is a stronger read than one.
The legend under the chart names every layer in plain words, and each item in it is a switch, not a label. Tap ghost 2 or ghost 3 to bring the dotted echoes in or out; tap the others to hide the overlay or the band while you study the candles alone. The chips are how you control clutter, so the chart only ever shows what you asked for.
Around the forward path sits a soft shaded band. It is the room the instrument gives the ghost to be right — a margin of normal wiggle around the path. While the live price stays inside the band, reality is still tracking the memory; when price walks out of the band and stays out, the instrument flags that the ghost is diverging, which the lifecycle chapter covers. For now, read the band as the ghost’s allowance: inside it, the rhyme is holding.
Each ghost carries a small ring where it was born, with its age beside it (for example 14m). A ghost freezes at birth and never repaints, so as price drifts the frozen path stays put while the candles move. That growing gap between the path and the price is the divergence itself — not a glitch, but the instrument being honest that the memory and reality have parted.
Two runs can trace the same shape while one moved twice as far as the other. The instrument matches on shape first, then reports the difference in size on each match card — a note such as “past moved ×1.5 bigger” or “×0.8 the size” (see the cards in Fig 3.2). The overlay you see is drawn amplitude-matched by default, scaled into your run’s space so your eye can judge the rhyme cleanly. The Align overlay switch on the toolbar turns that off, redrawing the old run at its true original size so you can see how far it really moved.
If you have zoomed or scrolled the chart and lost the overlay, the Fit proof button reframes the view so the matched run and your live run fill the screen together again. It is the one-tap way back to the picture the instrument wants you to see, and the fastest way to compare several matches in a row without fiddling with the chart.
You can study any part of the chart closely. Roll the mouse wheel over the price scale on the right to stretch or squeeze the prices, and drag vertically to move the view up and down. Whenever you take manual control of the price scale a small AUTO pill appears, and tapping it hands the scaling back to the instrument. One thing never changes under your hands: only the real candles decide the automatic framing, so turning ghosts or bands on and off never jumps the view.
The board opens on the five-minute chart, which is the timeframe most runs are read on, and a click moves you to the one-minute for quicker runs. Whichever one you are on, the engine tells you how long a run it matched against the past — for example matched on the last 8 hours — and it chooses that length itself, trying the longest run first and stepping down until it finds one that genuinely rhymes, because a longer run that still matches is harder to fake and therefore stronger evidence.
Set apart from all of that is the WINDOW control on the toolbar, which offers 1h, 4h, 8h and 12h and simply governs how much recent history the chart draws around the live run. It opens at 4h and behaves the same way on every timeframe, so you can pull the view out to twelve hours to see where the current run sits in the day, or tighten it to an hour to read the last few candles closely. It is worth holding the window and the matched length apart in your mind, because they answer different questions: the window is how much chart you are looking at, while the matched length is how far back the engine reached to find the rhyme, and the two move independently of one another.
A ghost tells you what may happen. It does not, on its own, tell you when, because timing depends on the structure of today’s market rather than on the shape of an old run. The WHEN layer draws that structure — the zones, the rails and the live order-flow events — and reads the ghost’s turns against it. You reach it with the Structure switches above the chart and the WHEN readout below.
Turn on Zones and Rails and the chart marks the ground price respects. Zones are shaded bands — the value area where most business was done, the premium and discount halves of the range, and the pivot zones around recent swings. Rails are the important price lines, tagged on the axis: BSL and SSL for buy-side and sell-side liquidity, POC for the volume point of control, and pivots. When two rails share a price they merge into one tag.
| Rail | What it marks |
|---|---|
| BSL | Buy-side liquidity: the highs above price, where stops from short sellers rest. |
| SSL | Sell-side liquidity: the lows below price, where stops from buyers rest. |
| POC | The volume point of control: the single price where the most business was done. |
| PROT / EXT | A protected pivot and the leg extreme — the structural turning points of the recent move. |
Turn on Events and the chart marks what the order book and the tape are doing this minute. A wall is a large resting order tagged on the axis with its size. A liquidity grab, marked with a ⚡, is a fast sweep that pierced a high or low and then closed back inside — the market reaching past a level to take the stops resting there. Icebergs (a hidden order that keeps refilling) and imbalance arrows (buyers or sellers running through several levels) round out the set.
Every bend in a ghost’s forward path gets one question asked of it: does this turn land on a real level of today’s chart, or in the middle of nowhere? A green dot means the turn has a reason to happen there, because a level is waiting for it. A grey dot means a pretty shape with no appointment, so at that bend you trust the direction more than the timing. Lots of green means the past and the present are telling the same story; a run of grey means the shape may be right while its clock is unreliable.
You can see the dots on the forward path itself. In the view below the turns carry green dots, and several of them sit close to the tagged rails — the memory’s bends line up with the levels live right now, so both the shape and the structure point to the same moments.
Below the confirm panel the instrument writes the same reading in plain language, one line per turn. A line such as “its turn sits on the discount band and a 2h pivot zone” tells you the next bend has two reasons to happen where it does. A green tick and “sits on” means the turn has an appointment; open space means it does not.
This chapter is about the single decision that keeps a scalper alive: where a stop can sit without being fed to the market. ghost|path draws a safe-stop area and one or more target bands for whichever direction you are considering, built from the same structure the WHEN layer already showed you. It never says place your stop here. It shows you the reading, and the trade stays yours.
Stops cluster in obvious places. Everyone who is long puts their stop a little below the same recent low, so that low becomes a pool of stop orders in plain sight. The liquidity rails — BSL above the market and SSL below it — are the map of exactly those pools, and a liquidity grab, the ⚡ on the chart, is the machine that harvests them: price reaches down through the obvious low, triggers the crowd’s stops, then closes straight back above. If your stop was in that pool, it was not unlucky, it was the target.
The most natural place to put a stop, a little beyond the recent low or high, is the exact place the grab is aiming for, because that is where the pool is. A safe stop has to sit beyond the pool, on the far side of the price the market has to defend, and it has to be at least as wide as the coin’s own normal noise. Anything tighter is not a stop, it is a donation.
Turn on Risk geometry and pick if long or if short. The instrument draws nothing until you choose a side, because the geometry of a safe stop is different for a long than for a short, and choosing the side is your call.
With a side chosen, the instrument shades a safe-stop area built from three tests at once: it sits beyond the stop pool, past the BSL or SSL level where the crowd’s stops rest; it sits past the defended wall, beyond the large resting order the market is protecting; and it is at least the coin’s noise floor wide, so ordinary wiggle cannot reach it. The target bands mark where the move could reasonably travel, drawn where the echoes’ typical travel meets a real level such as a gap, the value area or the volume POC.
Every coin has a natural jitter, a width below which price is simply breathing and any stop is certain to be clipped by noise alone. The instrument measures that width per coin and calls it the noise floor, and it will not draw a safe-stop area any tighter than that floor. The number differs a lot between coins: a calm one has a floor near 0.35%, a middling one such as HYPE sits closer to 0.5%, and a wild one such as WLD measures around 1.2%. A stop that is perfectly safe on one coin is noise-bait on another, and the floor is how the instrument keeps you honest about it.
For a short the whole picture flips. The safe-stop area moves above price, past the BSL and the level the market defends on the way up, while the target bands drop below, where the echoes’ travel meets the value area and the volume POC. The reading is the mirror image, side for side.
The instrument says safe-stop area, target band and stop pool. It never says place your stop at this price or take profit here, because it is teaching you to read where risk lives, not making the decision for you. Read the area, read the band, weigh the noise floor, and then choose. The reading is the product; the trade is yours.
The overlay proves the shape. The confirm panel tests whether the world around the two runs is the same as well, because a shape that rhymes inside a matching situation is a far stronger read than a shape that rhymes by itself.
For the match you have selected, the panel lays the context of the old run (then) beside the context of your run (now) across three rows. Each row gets a tick when the two agree and a cross when they do not, and the header counts the agreements.
| Row | What it compares (then vs now) |
|---|---|
| Position vs EQ | Where price sat in its range — premium, discount, or around the midpoint. Here it was at equilibrium then, but sits in the discount now: they disagree. |
| Near liquidity | Whether the run was resting on a pool of orders. Neither run was, so this row agrees. |
| Tape character | How the order flow was behaving and which way delta leaned. Both lean down, but the old run was notable while yours is quiet: they disagree. |
The panel exists to catch exactly one mistake: being seduced by a beautiful overlay whose context does not match. The AAVE overlay above is a tight 0.98, yet only one of three context rows agrees — a convincing shape sitting in a different situation. The panel says so in its own words, because all it does is confirm or dispute the mimicry rather than instruct you.
The panel re-reads the live context roughly every thirty seconds so it keeps pace as your run develops, and it carries an updated at stamp so you know how fresh the read is. It is change-aware: when nothing has actually changed the panel does not flicker or redraw, it simply sits still. Movement on the panel therefore means real news, not restlessness. When a piece of context cannot be read freshly — most often the tape row when the live tape is too quiet to judge — the instrument shows that row as absent rather than filling it with a stale value, and the agreement count is measured only over the rows it could compare.
A ghost has a life. It is born frozen, it is given a margin to be right, and eventually either it fades quietly or reality walks away from it and the instrument flags the break. Understanding that life is what lets you trust the overlay, because it is what stops the instrument from cheating.
The moment a ghost is drawn, its forward path is fixed. It is drawn once, from the prices at its anchor, and never re-drawn to flatter the live candle. This matters more than anything else on the board, because a line that quietly re-fits itself to whatever just happened would always look right and would therefore be worthless. ghost|path refuses to do that, and it says so on the chart in its own words.
While a ghost is live the instrument grades its health and shows it in a small chip. 🟢 On script means price is still tracking the path. 🟡 Symptoms flags an early warning, such as a liquidity grab against the path or the tracking error starting to rise. 🟠 Warning is stronger — a missed appointment or the tape turning against the run. 🔴 Diverged means reality has clearly left the path. Tap the chip and it spells out the reason in plain words.
When the live price sits outside the band for enough of the recent bars, the instrument raises a divergence banner. It states the fact plainly — how many of the last bars sit outside the path — and hands the decision to you with the words your call. It does not retire the ghost for you, because a break of the path is information you may want to act on in either direction.
On the chart you can see why the banner fired. The frozen forward path holds its line while the live candles pull away from it, the birth ring marking where the ghost was anchored. Because the path never moved to follow the candles, the break is real information rather than a redrawn excuse.
Any tool can draw a line that looks right in hindsight. The only line worth anything is one that was drawn before the outcome and then left alone to be judged. That is what “frozen at birth” buys you: every overlay on the board is a claim the instrument made in advance and cannot take back, and the divergence flag is it owning the claim when reality disagrees. Invalidate a ghost and the next-best match takes its place at once; the retired ghost fades but is never erased, because the record of how ghosts actually performed is the honest scoreboard.
This is the chapter that explains why you can trust the instrument. Everything ghost|path does is built to tell you the truth about a rhyme, including the truth that sometimes there is no rhyme at all — and this is the method it holds itself to.
The similarity number measures how closely two shapes line up, and nothing more. Around 0.95 is a strong, tight overlay and around 0.80 is a loose one, so it is a good measure of how much to trust your eyes. It is not a probability that a trade will work, it is not a forecast, and it carries no odds. The honest edge, if there is one, lives in your shape-sense and your reading of the levels; this number simply tells you how good the rhyme in front of you is.
Roughly half the time, a coin has no run in history that convincingly matches its current shape, and when that happens the instrument publishes nothing for it and says so. An empty read is a real result. A tool that always found a match would be flattering you, and ghost|path would rather be silent than invent a rhyme that is not there. In the same spirit, when a live feed goes quiet the instrument shows the read as absent rather than dressing up an old value as current — an absent read beats a stale one every time.
At the bottom of the board the instrument names itself: a reading instrument, not a signal — you judge the rhyme. That is the whole contract. It finds, it proves and it confirms, it freezes its claims and owns them, and then it stops. What to do with the reading is, always, your decision.
Under the panels sits the record, the scoreboard of how frozen ghosts actually performed once the market caught up to them. It is built forward, from ghosts frozen before their outcomes were known, which is the only kind of track record that means anything. Grading arms as the first cohort of ghosts ages past its horizon, so the record fills in honestly over time rather than being back-fitted to look good on day one.
If you want the instrument to tap you on the shoulder when a strong setup forms, Ghost Summons sends a tiered come-and-look alert to your phone. It is an invitation to open the board and read, never a signal and never a direction — it tells you a rhyme worth looking at has appeared, and stops there.
This chapter puts the whole instrument together as one continuous read, in the order you would actually use it. Follow it once and you have the rhythm of the board.
You open echo.trading/board.html and look at the rail. The strongest rhymes sit on the left, so you start there — the coin on the far left is rhyming with the past most tightly right now, and it is the one already drawn on the chart.
You open the coin and look at one thing: how tightly the purple line hugs the candles. It hugs well, so the shape genuinely rhymes. If you want a second opinion you tap ghosts 2 and 3 to see whether the next two memories agree about the path.
You drop to the confirm panel. A high agreement count means the situation rhymes as well as the shape; a low count is your cue to be careful even though the overlay looked good. Here it reads one of three, so you keep your guard up.
You read the WHEN rows to see which turns have an appointment with live structure, then decide you are reading this as a long, turn on Risk geometry and pick if long. The safe-stop area appears beyond the pool and the target bands where the echoes’ travel meets structure. You weigh the safe-stop area against the coin’s noise floor and make your own decision.
You keep the coin open and watch. While price stays inside the tolerance band the memory is holding. If price walks out and stays out, the divergence banner fires and says your call, and you decide whether to invalidate the ghost and let the next memory take its place. Either way the instrument has told you the truth: it made its claim in advance, and it owned the moment reality disagreed.
While you watch, the health chip keeps grading the memory for you, so you do not have to stare at the band yourself. Green means the run is still on script; the first symptoms — a grab against the path, the tracking error creeping up — show up here before the banner ever fires.
Scan the rail, judge the overlay with your eye, ask the confirm panel whether the world agrees, read the timing and the risk for your side, then watch the ghost and honour the break. That is the whole session — five reads, each one on the board, and every decision along the way yours.
A short set of the things new readers ask, with plain answers.
Almost certainly not. An empty rail means nothing in the market is rhyming convincingly right now, and the instrument keeps hunting every minute. It is designed to stay silent rather than show you a weak match, so an empty rail is the instrument working correctly, not failing.
Two honest reasons cover almost every case. The coin may have no convincing echo in history at this moment, which the instrument states plainly, or its recent run may not hold enough structure to match against. Both are real answers. Also check your timeframe: the fifteen-minute view carries no ghosts on purpose — it is the context view, there to show you the trend and the swing structure around your scalp. Drop back to 1m or 5m to read the memories.
The live structure feed has gone quiet for a moment, and the timing frame returns when the feed does. The instrument would rather tell you the frame is momentarily absent than draw timing dots from data it cannot currently trust.
That row could not be read freshly, most often the tape row when the live tape cannot be read honestly right now. The instrument shows the read as absent rather than filling it with a stale value, and the agreement count is measured only over the rows that could be compared.
You have taken manual control of the price scale, which is why the AUTO pill appeared. Tap the pill and the instrument reframes the candles for you again.
Your preferences — which ghosts you show, which ghosts you have invalidated — are remembered per device rather than in an account. Open the board on a different device and it starts from the sensible defaults again.
ghost|path finds where your shape has lived before, proves the rhyme to your eye, and confirms the world around it, and then it stops. It is a reading instrument, never a signal. You judge the rhyme, and the decision is always yours.