Advertisement IC Markets

Minimum Swing Size on M1: Filtering Noise Before Calling a Structure Break

Discuss 1-minute to 15-minute price action setups, fading intraday momentum, key support/resistance zones, and proven short-term trading methodologies.
Post Reply
FXS Prop Desk
Posts: 1693
Joined: Wed Oct 07, 2026 9:52 am

Minimum Swing Size on M1: Filtering Noise Before Calling a Structure Break

Post by FXS Prop Desk »

On M1, almost every few candles create a new "swing high" or "swing low". Without a minimum swing size, structure breaks happen so often that they carry little meaning.

Ways to define a meaningful swing:
  • Fixed pips: e.g. at least 3 pips from the previous opposite swing. Simple, but does not adapt to volatility.
  • ATR-based: at least 1 × M1 ATR(14), or 0.5 × M5 ATR.
  • Bar count: a swing high must have at least N lower highs on each side (fractal logic).
Once the swing definition is fixed, a structure break means a close beyond the last valid swing — not just a wick.

Effects of a stricter definition: fewer signals, but each one reflects a larger shift in order flow. For scalpers, an ATR-based swing usually works better across sessions than a fixed pip value, because Asian and London volatility differ so much.

How do you decide which M1 swings count when reading structure?
FXS Official Team | The Idea Lab & Strategy
Recommended broker for automated trading & scalping IC Markets
Shadow Trader
Posts: 5594
Joined: Wed Sep 09, 2026 5:19 pm

Re: Minimum Swing Size on M1: Filtering Noise Before Calling a Structure Break

Post by Shadow Trader »

I use the ATR version but with a floor tied to the spread. My rule is the larger of one M1 ATR(14) and four times the current spread, because late in the Asian session M1 ATR on EURUSD can drop to about 1.2 pips, and a swing that size is mostly bid and ask noise.

On GBPJPY the same floor does more work. A 2.5 pip spread in thin hours sets the minimum at 10 pips even when ATR reads lower.

There is a lag to expect with any size filter. A swing only becomes valid once price has moved the threshold away from it, so the latest high on screen is often unconfirmed, and tools that draw it early repaint.

On the break itself, I want one more condition than a close. If the next M1 bar closes back inside the old swing, I treat the break as failed and stand aside rather than flip direction.

Fractal counting worked worst for me. Two bars each side on M1 produced roughly as many swings in a dead hour as in the London open, which is the opposite of what a filter should do.

Do you reset the swing reference at session opens? I start fresh at 07:00 London.
Shadow Trader
Posts: 5594
Joined: Wed Sep 09, 2026 5:19 pm

Re: Minimum Swing Size on M1: Filtering Noise Before Calling a Structure Break

Post by Shadow Trader »

One thing I would add on which swings count is to keep the small ones on screen rather than deleting them. I mark swings that fail the size test in grey, because stops cluster under them and they are where most false breaks start.
img_13739_r2.png
img_13739_r2.png (84.15 KiB) Viewed 73 times
The chart shows a typical EURUSD M1 morning with my threshold at 3.5 pips. The low at 1.1731 counts because the rally off it ran 13 pips, while the dip to 1.1737 only bounced 2.5 pips before rolling over, so it stays grey.

A close under 1.1737 is a warning for me, not a break. The structure only turns when an M1 bar closes under 1.1731, and that bar came six minutes later.

The threshold is fixed at the moment the swing forms, using ATR as it stood then. If I recalculate with the current ATR, a quiet Asian swing can lose its status at the London open simply because volatility doubled, and the structure rewrites itself behind me.

Does anyone treat a run of grey swings breaking in sequence as a structure break in its own right? I have seen three in a row before the real one more often than I expected.
Shadow Trader
Posts: 5594
Joined: Wed Sep 09, 2026 5:19 pm

Re: Minimum Swing Size on M1: Filtering Noise Before Calling a Structure Break

Post by Shadow Trader »

Raw price streams from liquidity sources contain errors. Fat-finger quotes, delayed packets, crossed markets and isolated outliers all appear from time to time. Before prices reach a client terminal, many brokers run them through filters designed to remove such noise.

Filters come in several forms. A simple one rejects any tick that deviates from the previous price by more than a set threshold unless confirmed by subsequent ticks. Another throttles update frequency, passing along only the latest price within a fixed interval. Others check each tick against several feeds and discard values that disagree with the consensus.

Filtering protects against false stop-outs and bad fills caused by a single erroneous quote, which is a real benefit. The trade-off is that legitimate rapid moves can also be smoothed, delayed, or shown with fewer intermediate ticks. Two platforms connected to the same pair can then display slightly different micro-structure.

For scalpers working with tight targets, this matters. A strategy built on tick-level patterns may behave differently on a throttled feed than on an unthrottled one, and spikes visible on one chart may never exist on another. Comparing tick counts per minute and maximum gap sizes across feeds gives a measurable view of how much filtering is applied.
Post Reply