It arrives at ten past nine in the morning in a group of about two hundred people. A daily chart, a rising channel drawn with two clean touches on each side, a small arrow marked "entry" near the lower rail, a green rectangle above marked "target", and four words underneath: posted here last month. Nobody in the group disputes it, because there is nothing to dispute — the candles are real, the channel does contain them, and the price did reach the box. The question worth asking is not whether the picture is honest. It is what a picture of a chart is capable of showing, and the answer is: a good deal less than it appears to.
This lesson is the module’s last because it uses all of the others. Everything the previous four established — that a price must sit on a tick, that a size must be an integer, that the axis counts sessions, that the instrument may be a proxy — is invisible in an image, and each of them is a way a chart can look perfect and have been untransactable.
A three-minute highlights reel of a batsman is entirely genuine footage. Every shot in it was played. It still cannot tell you his average, because the deliveries he missed and the innings he did not score in were never filmed for the reel — not out of dishonesty, but because a reel is a selection and selection is what makes it watchable.
A posted chart is a highlights package of one setup in one stock over one window. The candles are real and the annotation may be sincere. What the format cannot carry is the denominator: the other times the same setup appeared and did nothing, the trades taken alongside it, and the position size that decides whether any of it mattered.
Seven things a picture cannot carry
| What is missing | Why it decides the answer | The question that recovers it |
|---|---|---|
| When the annotation was drawn | A channel drawn today across last month’s candles is a description. The same channel drawn last month is a prediction. The two are pixel-identical and only one of them is evidence. | Is there a version of this image timestamped before the move? If not, this is a drawing about the past. |
| Where the left edge was placed | The start of the window is a choice, and a different start supports a different structure. Two years of history behind that channel may show four earlier breaks of it. | What does this look like with three times the history, on my own screen? |
| Which scale the axis is on | A trendline that holds on a linear axis can be broken on a logarithmic one over the same data, and the picture does not say which was used. | Redraw it on both. If the conclusion flips, the conclusion belonged to the axis. |
| Whether the marked prices exist | An entry marked at a computed level is not a price. The arrow may sit between two ticks, or inside a wick that consisted of one small print. | Round the marked entry to a valid tick and check the volume in that candle at that price. |
| Whether the marked size was available | The implied position may exceed what the book could absorb, which converts a clean fill into several worse ones. | Compare the implied quantity with the stock’s ordinary daily volume. |
| What else was posted that week | One winning chart out of thirty posted charts is a different claim from one out of one, and the image is identical in both cases. | Scroll back through the source and count the posts, not the wins. |
| What it cost to hold | Costs, the spread crossed twice, and the nights held through are all outside the frame, and on a short-horizon claim they can be the whole result. | Subtract a realistic round trip from the marked move and see what survives. |
The ten-minute audit
None of this requires you to assume bad faith, and arguing about intent is a waste of an afternoon. There is a mechanical procedure that settles it, and it works the same way whether the chart came from a stranger, a broker’s research note or your own folder.
- 1Reproduce it on your own data
Same stock, same timeframe, same indicator settings, same adjustment convention. If you cannot reproduce the picture, stop here — the disagreement is about data, and no amount of discussion about the pattern will resolve it.
- 2Extend the right edge
The part of the series after the screenshot’s last candle is the only genuinely out-of-sample evidence in the whole exchange. Everything to the left of that edge is the data the line was fitted to.
- 3Count the other occurrences in the same name
Scan back through the same stock for every instance of the same setup over two or three years. You are recovering the denominator that the image omitted, and forty occurrences is where a hit rate starts meaning anything — the Sample size problem does not go away because the one example was vivid.
- 4Check transactability at the marked prices
Round the entry to a tick. Check whether the candle it sits in had volume at that price. Work out the size your own rule would have taken and compare it with the day’s turnover. A claim that fails here failed before the analysis started.
- 5Ask what would have made it a loss
Move the stop a quarter of an ATR tighter and see whether the trade survives. A result that depends on a stop placed to the paisa is a result about that stop.
- 6Write down what you concluded, with the date
If you do not, the next chart from the same source arrives with a clean slate, and no source ever accumulates a record.
The mirror: your own charts are somebody else’s screenshot
Every failure in the table above happens inside your own workflow with no one else involved. You open a chart, draw a line, take the trade, and three weeks later review it — on a chart that has since been redrawn, on a window that now starts somewhere else, with the outcome already known. The review is not dishonest. It is structurally identical to reading somebody else’s screenshot, with the additional handicap that you already know how it ended.
- Save the annotated chart at the moment you decide, not afterwards. An image with the date on it is the only artefact that can later distinguish what you saw from what you now remember seeing.
- Write the two or three sentences that would have made you wrong. They cost a minute and they are the only part of the record that Hindsight bias cannot rewrite.
- Record the settings, the data source and the rounding. Everything this module has covered — tick, size, session count, instrument — is a variable, and an unrecorded variable makes a review unreproducible.
- Keep the charts of the setups you did not take. They are the denominator for your own method, and they are the only ones nobody keeps.
The chart in the group, at ten past nine
A rising channel, an arrow at an entry near ₹247.63, a target box about 9% higher, and a caption saying it was posted a month ago. The market opens in five minutes and three people in the group have already said they are in.
Module checkpoint: what you can actually transact
5 questions. Answers are revealed once you submit all of them.
1.You enter a share at ₹9.40 on a five-paise tick, and your written rule places the stop 1% below the entry. What are your actual choices?
2.A ₹4,00,000 account risks 1% per trade. A candidate is quoted at ₹18,400 with a stop ₹1,900 below the entry, so the sizing formula returns 2.11 shares. What is the effect of taking three shares instead of two?
3.One bar in your yearly series was produced by the Muhurat session — roughly an hour of trading rather than a full day. What does leaving it in a fourteen-bar ATR do?
4.An exchange-traded fund holding overseas securities has traded at a premium of several per cent to its indicative NAV for weeks. What is the most likely explanation?
5.A screenshot shows a rising channel with an entry arrow and a target box, captioned as having been posted a month ago. Which single check does the most work?
Coaching ke bahar bees bachchon ke photo lage hain, sab asli, sab pass. Poster jhootha nahi hai — bas usme woh ginti nahi likhi jo sabse zaroori hai: kitne baithe the. Group mein aane wale chart ka bhi yahi haal hai. Candle asli hain, channel sach mein bana hua hai, target bhi lag gaya. Jo tasveer mein aa hi nahi sakta: line kab kheenchi gayi — chalne se pehle ya baad mein; chart ka baayan kinara kahan se shuru kiya gaya; aur us hafte kitne aur chart post hue the jo chale nahi. Sabse sasta test ek hi hai — wahi chart apne app pe kholo aur screenshot ke daayen kinare ke aage badha do. Us kinare se pehle ka poora hissa toh woh data hai jispe line fit ki gayi thi; saboot sirf uske aage wala hissa hai.
- A chart image is a sample of size one, selected after the outcome was known — hindsight supplies the annotation and survivorship supplies the selection.
- The only out-of-sample evidence in a screenshot is whatever happened after its right edge, so extending it is the highest-value check available.
- Reproduce, extend, count the occurrences, test transactability at the marked price, and see whether a slightly tighter stop kills it.
- Everything this module covered — tick, whole shares, session count, the instrument being a proxy — is invisible in an image and each one can make a perfect chart untransactable.
- Your own review is the same problem: save the annotated chart with its date, and keep the setups you did not take, because they are your denominator.
Mark it done to track your progress through the curriculum.
Common questions
Short, direct answers to what people ask about this topic.
- the tendency to see a chart setup as obvious once the outcome is known is called
- Hindsight bias. It is what makes a line drawn after the move look inevitable, and it leaves no mark on the image — a channel fitted to candles that have already printed looks exactly like one drawn in advance. Survivorship bias does the other half of the work by deciding which chart got posted at all, so what you are shown is a sample of size one, selected after the result was known.
- how do I verify a chart screenshot posted in a trading group
- Reproduce it on your own data first — same stock, timeframe, indicator settings and adjustment convention — and if the picture will not reproduce, the disagreement is about data and the pattern discussion is pointless. Then extend the right edge past the screenshot’s last candle, because that stretch is the only genuinely out-of-sample evidence available. None of this requires assuming bad faith; it is a mechanical check that works the same on your own old charts.
- how many occurrences before a chart pattern hit rate means anything
- Around forty instances is where a hit rate starts carrying information; one vivid example carries none, however clean it looks. Scanning two or three years of the same stock for every appearance of the setup is how you recover the denominator a screenshot leaves out — the times it appeared and did nothing. A small sample does not become larger because one of its members was memorable.
- what does a chart image not show you
- When the annotation was drawn, where the left edge was placed, whether the axis was linear or logarithmic, whether the marked entry sits on a valid tick with volume behind it, what size was actually available, what else the same source posted that week, and what holding it cost. Four of those are variations on the same problem: the picture was framed after the outcome was known. Every one of them is a way a chart can look perfect and have been untransactable.
- is it legal to post stock tips with entry and target in india
- Recommending securities to others is a regulated activity in India, and offering buy, sell or target calls generally requires registration with SEBI as a research analyst or investment adviser — the framework deliberately separates education about a method from advice about a security. A chart carrying an entry, a target and a stop is a recommendation whatever the caption calls it. A chart posted publicly can also be the trade itself, if the poster benefits from others acting on it.