Patterns and signals — course checkpoint
You began this course able to name shapes on a chart. You finish able to say what each shape encodes about behaviour, and — more valuably — how much weight it can honestly bear. Let's assemble the whole thing.
The shapes, and what each one is a picture of
- Continuation patterns (rectangles, flags, pennants, triangles) are pictures of a pause: profit-taking meeting patient bidders after a fast move, on falling volume. Triangles add one detail — a flat side suggests concentrated supply or demand parked at a specific price.
- Reversal patterns (head and shoulders, double tops and bottoms) are pictures of a handover: demand thinning while supply arrives progressively earlier. The mechanism is finite buyers — every recruit becomes future supply — plus the market's memory of levels that previously rejected the price.
- Candlestick patterns are pictures of one session, compressed into four numbers. The body is the net verdict; the wicks are rejected territory. Engulfing candles say a full session was reversed, usually because size or news arrived. Long wicks say a price was visited and didn't hold.
- Gaps are pictures of a closed market: a hole in the record where the exchange wasn't open while information arrived. They're largely an artifact of trading hours — and in unadjusted data, sometimes just a dividend or a split.
- Breakouts are the moment resting orders get consumed. They fail often because a visible level is where stop orders cluster, and stops firing looks identical to genuine demand right up until it stops.
The four honesty rules that outlast the shapes
- Seeing a pattern is not evidence. Random walks produce every shape in this course, abundantly. Your eye would have found something either way.
- Hindsight rewrites charts. Cover the right side before you judge. The clarity you feel is usually coming from the future, not the pattern.
- Definitions create the statistics. Failure rates, success rates, "works 80% of the time" — all of them follow from choices about what counts as a pattern. A number quoted without its definition is decoration.
- Compare to the base rate, out of sample, after costs. A pattern must beat what would have happened anyway, on data you didn't tune it on, by enough that the edge still survives the spread. Most don't.
What patterns are genuinely good for
After all that scepticism, something real remains. Patterns are an excellent vocabulary for describing order flow. "Participation fell across three successive peaks." "The level that was defended twice stopped being defended." "The break happened on a single volume spike that didn't persist." Those are checkable statements about what buyers and sellers did, and being able to make them precisely is a genuine skill.
What patterns are not is a machine that converts shapes into expectations about the future. This course has never given you an entry, an exit, or a target, and that omission is deliberate — not a gap to be filled elsewhere.
Before you sit it
Each of these is a minute at your desk. Any one that is not names the lesson to reopen first.
- Name the two things a breakout needs before it is one — What actually happens when a consolidation breaks?
- Say what the three peaks of a head and shoulders encode about a fading trend — What does a head and shoulders encode about a fading trend?
- Say what an engulfing candle tells you about who arrived that session — What does an engulfing candle say about who showed up?
- Describe how you would test a pattern for edge, in three steps — How would you test whether a pattern has any edge?
Try it now
- Write a five-sentence description of the chart below using only this course's vocabulary — one sentence each for a continuation shape, a reversal shape, a notable candle, a gap, and a breakout or its failure. Describe only; predict nothing.
- Take the single most convincing pattern in that description and apply the Unit 4 test to it: define it mechanically, fix a window and write the dates down before you start, count every occurrence inside it, and compare against the unconditional base rate over the same window — how often the market did that thing anyway.
- Failure hunt, one last time: find the clearest case in that chart where a textbook pattern completed and led nowhere. Keep it. It's a better teacher than any of the ones that worked.
Checkpoint quiz next, then on to the risk and discipline course. Nothing in this course was a recommendation to buy or sell anything, and no pattern here was a signal — tickers were illustrations, and what you've gained is a way to describe price behaviour honestly, including the honest admission of how often it misleads.