Market Structure (ICT)
Multi-Timeframe Confluence
The same entry signal means something different depending on which way the bigger picture is leaning. Two synced charts make that comparison directly instead of asking you to imagine it.
Watch both timeframes at once
The lower chart is the real tape, six of its candles roll up into every candle on the higher chart above it, the same price data at two resolutions. Press play and the lower timeframe streams in. Every time it fires a breakout signal, the tape pauses so you can take it or skip it, and your call gets logged as with or against whatever the higher timeframe's own trend is doing right now.
Higher timeframe (6 lower-timeframe bars per candle)
Bias: —Lower timeframe (the real tape)
Press play to start the tape.
How it works
- The higher timeframe isn't separate data, it's the same tape rolled up. Every 6 lower-timeframe bars become exactly one higher-timeframe candle, open of the first, close of the last, high and low of the group.
- Bias is read off the higher timeframe's own recent direction. It compares the last two completed higher-timeframe closes, nothing more exotic than that.
- The signal itself never changes based on bias. A lower-timeframe breakout is flagged the same way whether it agrees with the higher timeframe or fights it, the bias only labels the trade after the fact.
- A win here just means price kept moving that direction for 6 more lower-timeframe bars. That's a short, specific, and somewhat arbitrary window, a different one could tell a different story from the same tape.
- Confluence is a filter, not a separate signal. This page never invents extra entries when the timeframes agree, it only sorts the same lower-timeframe signals into two buckets after they've already fired.
Where this breaks
A short random walk barely has a higher timeframe to align with
Multi-timeframe confluence works best when the higher timeframe reflects a real, sustained trend, something with enough history and enough participants behind it to actually mean something. A few dozen bars of a random walk rolled up into a handful of higher-timeframe candles produces a "bias" that can flip on a single new candle, which is a much shakier signal than a real higher timeframe built from months of actual price history. Treat the bias chip on this page as a demonstration of the mechanic, comparing a signal's outcome against a rolled-up trend read, not as evidence that this specific bias calculation would hold up on a real chart with real structure behind it.