Candlestick Patterns
Three-Bar Reversal / Railroad Tracks
Two candles of nearly matching size and opposite color print back to back after an extended move, like a pair of parallel rails stopping the trend cold.
Lock it in before the next bar prints
The tape below scrolls forward on its own, one new bar at a time. The readout under the chart scores the two most recent bars live. Click Lock it in while a qualifying-looking pair is still current, the instant a new bar prints, that pair is gone. A few spots on this tape are decoys, similar size and opposite color, but without a real move behind them.
Waiting for the tape to start...
Click Lock it in when you think the two most recent bars qualify.
How it works
- Size match and opposite color both have to be true together. Two large candles of the same color are continuation, not reversal, no matter how big they are.
- The bars have to be big relative to what came before, not big in absolute terms. That's why the recent-average comparison is part of the mechanical test here, not just a visual impression.
- Adjacency is required. This is specifically back-to-back bars, which is also why this drill runs in real time instead of letting you pick any two candles at leisure.
- Timing the lock is part of what's being tested. A qualifying pair stops being the most recent pair the instant a new bar prints.
- The pattern flags exhaustion, it doesn't guarantee the reversal sticks. What happens over the next several bars is tracked separately, on its own stat.
Where this breaks
Strip out the extended move and two big opposite candles mean nothing
The "after an extended move" condition is easy to skip, treating any two opposite, similar-sized candles as a signal on their own. The size-vs-recent-average check exists precisely because railroad tracks only mean something after a real run. Two big opposite candles printing in the middle of an already-choppy, directionless stretch aren't exhausting a trend, since there's no trend there to exhaust. The decoy spots built into this page's tape exist to show exactly that, same shape, wrong context, no real signal, and the mechanical lock test still passes them, because ratio, color, and size relative to average are all it actually checks. Strip out the real-move context and what's left is just two large candles, which happens often enough on its own to make the naive version of this pattern nearly meaningless standing alone.