Range trading has a built-in assumption: price oscillates between two boundaries, so the trade is to sell the top and buy the bottom. We tested that assumption directly on EURUSD — every H1 touch of a range boundary during a sideways H4 trend, 3,111 of them across 16 years — and asked which way price actually resolved.

The base rate is not what "trade the range" assumes

Outcome	Count	Share of all touches
Breakout (H1 closes beyond the range with a real body)	2,114	67.9%
Reversion (price reaches the opposite extreme)	362	11.6%
Unresolved after 24 hours	635	20.4%

Strip out the touches that never resolved cleanly within a day, and the split among the rest is stark: 85.4% break out, 14.6% revert. A range boundary on EURUSD is, more often than not, a level about to be broken — not a wall price bounces off.

The wick tells you which one you're looking at

Not every touch is created equal. The distance the wick pushes past the boundary before the H1 candle closes is a real discriminator:

Sweep distance past boundary	Resolved as breakout	Resolved as reversion
0–2 pips	64.4%	12.7%
2–5 pips	71.9%	10.4%
5–10 pips	70.9%	10.1%
10–30 pips	76.0%	10.0%
>30 pips	85.7%	4.8%

A shallow poke past the level — the kind that looks like a clean rejection wick — is the closest thing to a real reversion signal here, and even that is still a minority outcome. A deep sweep, the kind that looks most like a "stop hunt," is in fact the strongest breakout tell. This echoes what we found in the liquidity-sweep study: the market rarely punishes itself for going too far.

Session and range maturity shift the odds, but never invert them

Reversion rate is highest during the Asian session's quieter hours and around the 07:00–09:00 UTC London handover (peaking near 15% at 08:00 UTC), and lowest in the 03:00–05:00 UTC dead zone (down near 7%). Range maturity — how many times the boundary has already been touched — matters too: young ranges (6–8 prior touches) revert about 16% of the time, while ranges that have been tested 19+ times revert only 6% of the time and are close to expanding for good.

None of these conditions flip the base rate. They move it from "mostly breaks" to "even more overwhelmingly breaks."

What this means

The exact entry triggers and stop placement we trade stay private, but the shape of the finding travels:

A range boundary touch is a pre-breakout setup far more often than a reversion setup. The playbook instinct to fade the edges needs a strong filter behind it, not a default assumption.
Sweep depth is diagnostic, not damning. A wick that pushes hard past the level is closer to a breakout tell than a stop-hunt tell — the opposite of how it's usually read.
Maturity and session context are amplifiers, not switches. They shift a 6-to-1 bias toward 8-to-1 or 12-to-1; they never make reversion the favorite.

If you're trading EURUSD ranges, the higher-probability read at the boundary is "this is about to break," and reversion is the trade you take only when a specific, narrow set of conditions stack in its favor.