Representativeness is pattern-matching mistaken for probability. A chart that looks like the textbook example feels like it must behave like the textbook example, even when the base rate says most such charts fail.
In practice, a stock that resembles a past ten-bagger gets treated as a likely ten-bagger. A three-day winning streak in a system feels like proof the system works, though three trades tell you almost nothing about expectancy. A company in a hot sector inherits the sector story regardless of its own numbers.
The antidote is to ask what fraction of all similar-looking cases actually worked out, then to weigh the resemblance against that number. Resemblance is a hypothesis, not a result.
Related: base-rate-neglect, sample-size, narrative-fallacy, hot-hand-fallacy