The method

How this works, and what it cannot tell you.

This page is published so it can be argued with. Every rule that decides what goes into the library, and every rule that scores a card, is written out below in the same words the product uses on screen. The second half of the page is the part most tools leave out: the limits.

1. The words

The level: The highest price a stock reached in its previous 20 trading days. It shows on the chart as a dashed orange line. Think of it as the ceiling the stock has not managed to get above recently.
A breakout — the days that MOVED 20%+: The stock pushed above that ceiling and finished the day well above it. In this collection a day is labelled a breakout when it closed at least 20% above where it opened, and above the level. There are 896 of them.
A failed breakout — the days that DID NOT: The stock pushed above the same ceiling and then fell back below it before the close. There are 2,285 of them — considerably more than there are breakouts, which is worth sitting with.
Entry: The price at which you would have bought. You mark it yourself; the product never suggests one.
Stop: The price at which you would have given up and sold, to cap the loss. You mark that yourself too.

896 and 2,285 are facts about this collection, not about the stock market. They are not a success rate, a win rate or a base rate, they are not offered as one, and they should not be read as one. They describe how many days of each kind were found inside a narrow, filtered universe over one stretch of years. Section 6 sets out how narrow.

2. What is in the library, and what is kept out

The collection is 3,181 chart days drawn from 1,502 tickers between June 2021 and September 2026. Every one is a real session as it traded. None is simulated, redrawn, or assembled from parts.

A day only becomes a candidate if the stock clears a fixed liquidity and price filter. This filter is not adjustable by the user, on purpose:

Condition	Why it is there
Price between $1 and $20	Below a dollar the chart stops being a reliable record of tradeable prices — the spread starts to dominate the candle, and a shape you learn there does not transfer.
Real dollar volume on the day, and a substantial 20-day average	A shape formed on almost no volume is not the same shape. Thin days teach a pattern that does not exist at size.
Major US exchanges only. No OTC.	Same reason, more so.
Regular trading hours. ETFs excluded.	One session shape, one kind of instrument, so that the thing you are learning to recognise stays a single thing.
Long side only, at the beginner level	Recognising the shape comes before doing anything more complicated with it.

This is a deliberately small window on the market. It is chosen so that one shape can be learned properly, not so that the library represents trading in general. It does not.

About the groups

No two charts are the same. Every card here is sorted by one rule, the same rule for all of them: what the price did before it reached the line, and what it did after. Most cards sit clearly in their group. Some sit on the edge, and a few do not fit any group well. We read the borderline cards by hand and we move the lines as we learn, so a label can change between visits. Take the group as a starting point, not a verdict. Practice is you deciding what a chart looks like and what you would have done. The label is there to get you started. The chart is there to be argued with.

How a card is labelled

The breakout line on every card is the highest high of the previous 20 sessions. A day becomes a card when its high goes above that line. It is labelled Breakout when the close finishes above the line and at least 20% above the open, and Failed breakout when the close finishes back under the line. That is the whole rule, and it is applied by a program, not by hand, so it cannot favour the ones that look good.

Market capitalisation and float are as of a stated date, never a later one. Each card carries the most recent figure dated on or before its own day. Where no figure exists, the card shows a blank rather than a number from the wrong year. About one card in four is blank for this reason.

Times are US Eastern. Regular trading hours only, 09:30 to 16:00. Charts are shown as traded, not adjusted after the fact.

3. The order of the practice

Study the labelled ones first

Both decks are open, complete, and labelled — nothing hidden. The instruction is to open twenty that moved and twenty that did not, and just look. You are not hunting for a rule. You are letting your eye get used to two shapes that begin identically.

Marking cards before you have seen plenty of each is guessing, and guessing teaches nothing. That is the most common way people waste their time with a tool like this, and the product says so before it lets you start.

Then test yourself

A test card stops at 09:35 — five minutes after the market opens. You see five one-minute bars and the daily chart of the weeks before. You do not see the ticker, the date, or the label.

The ticker and date are hidden because with the symbol you could pull the chart up elsewhere in ten seconds and the repetition would be worth nothing. They come back at reveal.

You mark your entry, then your stop, then reveal. The rest of the day is not in your browser. It sits on the server and is not sent until your mark is saved to disk. That is not a promise about our good intentions; it is the only code path there is, and the product ships a command that checks it.

A card can be marked once. Changing your answer after seeing the outcome is not a repetition.

4. How a card is scored

After you reveal, the card is scored against a rule that was fixed before any of this was built. From wherever your entry actually filled: did the price rise 10% first, or fall 10% first, by 11:30?

TARGET: It gained 10% from your fill before it lost 10%.
STOP: It lost 10% first. If both happened inside the same one-minute bar, it is scored as a stop — a one-minute bar cannot tell you which came first, and assuming it went your way is how people fool themselves.
NEITHER: Neither happened by 11:30. Recorded, and counted as neither.
NO FILL: The price never reached your entry, so there was no trade at all. This is a real outcome and a common one. A great many plans simply never happen, and a tool that quietly ignored them would be flattering you.

Those four words are the whole scoring vocabulary. There is no equity curve, no simulated account balance, and no running total of pretend money anywhere in the product.

5. What is measured about you

One thing: whether your reading of a chart is getting better — how you place a level, an entry and a stop, and how that changes across repetitions.

One card tells you nothing about yourself. Twenty marks is roughly where a trend in your own reading starts to show, and the product says that on screen rather than congratulating you at three. A counter that only ever goes up is a slot machine, and that was a deliberate decision not to build.

6. The limits — the part that is usually left out

The test deck is deliberately half breakouts and half failures, so that guessing cannot beat reading the chart. The real collection is nothing like half and half, and the market is nothing like either. Do not read the mix in the test as how often a breakout works.

This measures chart reading. It does not measure whether you would make money, and it cannot. Position size, timing of the real fill, slippage, commissions, taxes, the state of your account and the state of you on the day are all outside it. A good score here is a good score here.
Reading a chart well is not readiness to trade. Nothing in the product, and nothing on this site, should be taken as a signal that you are ready to risk real money. It will never tell you that you are, and it is not in a position to know.
The universe is narrow by design. Section 2 lists the filter. What you learn here is one shape, in one price band, on one kind of stock, in regular hours. It is not a map of the market.
Five years is one stretch of history. June 2021 to September 2026 is a particular set of market conditions and it will not repeat. A shape learned inside it is a shape learned inside it.
The labelled decks contain hindsight and there is no way around it. Once a day is labelled, the outcome is in the picture, and studying it feels far easier than the real thing. That is precisely why the test cards exist and why the rest of the session is physically withheld.
A marked entry is an assumption. The scoring uses where your entry would have filled in the historical bars. Real fills are not that obedient, and NO FILL exists because plans miss.
Repetitions on history are not the same as repetitions live. There is no money at stake here, no clock pressure beyond the one built in, and no consequence. That difference is real and the product does not pretend otherwise.
No rate is shown to a beginner. Not a win rate, not a hit rate, not a base rate. A number like that read early, without the sample size and without the context, does more harm than the information is worth.

7. Why the failures are the point

Almost every breakout explanation you can watch uses examples that worked. They have to — a clean illustration needs a clean outcome. The result is that a beginner learns what a breakout looks like afterwards, which is a different skill from telling one from the other at 09:35, when both look the same.

The 2,285 days that did not move are not a bonus feature. They are most of the library, and they are the reason the other 896 are worth studying at all.

Trading Reps is educational software for learning to read charts. It is not financial advice, not a signal service, and not a recommendation to buy or sell anything. Nothing here predicts what any stock will do next. Past price movement does not indicate future results, and practising on historical charts does not make anyone a profitable trader. Trading stocks involves risk, including the loss of the whole amount invested.