Fractals and Swing Points, Defining Highs and Lows Properly

[Fairman]

Fractals and Swing Points, Defining Highs and Lows Properly

Every SMC concept depends on identifying swing highs and swing lows. Yet traders often disagree on what counts as one. A clear definition improves consistency.

diagram.png (39.48 KiB) Viewed 43 times

A common objective method uses fractals:

- Swing high: a candle whose high is higher than the highs of the candles immediately before and after it (commonly one, two, or three candles on each side).
- Swing low: a candle whose low is lower than the lows of the candles on both sides.

Short-term, intermediate-term, and long-term swings:

- Short-term high/low: the simplest three-candle pattern.
- Intermediate-term: a short-term high with lower short-term highs on both sides.
- Long-term: an intermediate high with lower intermediate highs on both sides.

Why it helps:

1. You mark structure objectively instead of by feel.
2. You can compare charts consistently across timeframes.
3. It makes backtesting more reliable, because your marks follow rules.

How to apply it:

- Choose one definition and use it for all your analysis
- Use higher-order swings for bias and lower-order ones for entries
- Don't mark every tiny wiggle

Common mistake: changing the definition when it conflicts with your opinion. That's bias disguised as analysis.

Pick your rule, apply it for 30 days, and notice how much clearer your charts become.

[Shadow Trader]

I lock fractal width before I mark a single swing. On M5 London structure I use two bars each side, so a swing high needs a higher high than the two candles before and after it.

One bar each side was too noisy for the same window. It marked every tiny wiggle as structure, and my bias flipped three times before 09:00 on days that were clearly still ranging.

I sketched a synthetic EURUSD M5 chart below with swing highs and lows labelled at two bars each side. The same segment with a one bar rule would have painted extra swings that never mattered for the later break.

img_12711_r1.png (56.34 KiB) Viewed 21 times

Higher order swings still come from nesting those two bar marks, not from switching width mid session. Changing the definition when a level conflicts with my opinion is the bias Fairman warned about.

I keep the same width for thirty days before I judge the rule. Do you stick to one fractal size across M5 and M15, or widen on the higher frame?

[Shadow Trader]

Fairman, confirmation lag is the fractal detail that still trips me on fast London opens. A two bar swing high is not real until the second right side bar closes, so for ten minutes I am marking a candidate, not a level I can hang a stop on.

I used to lean on forming fractals and then watch the right side bar take them out. Now pending structure stays in a separate colour until confirmation, and entries that need that high wait.

News spikes create temporary fractals that pass the width rule and fail the next hour. I tag any swing whose defining bar sits inside a Tier-1 minute as suspect and refuse to use it for bias until a later clean swing replaces it.

Nesting still builds higher order swings from confirmed marks only. Nesting candidates was how I invented intermediate structure that never survived the morning.

Thirty days on one width remains the review cycle. The lag rule sits on top without changing the width, and I refuse midweek experiments.

Do you trade against unconfirmed fractals at all, or is closed right side bars a hard gate like mine?

[FXS Prop Desk]

Defining swings by fractals is a good way to make SMC objective. One consequence worth mentioning: a 2-candle fractal on each side is only confirmed two candles after the high/low, so any rule using swings is always a little late. Traders who expect to see the swing in real time often get confused when the indicator "repaints" it. It's not repainting, just confirmation delay.