Draw two lines on a chart, rotate them until they roughly frame the price action, and call it a channel: that is how most channels get drawn, and it is why most channels do not hold up the second time price approaches either rail. A valid channel is not two lines that fit after the fact. It is two lines that were already there, touched independently by real swing points, before you needed them to mean anything.
What makes a channel valid
A channel needs two parallel rails, each touching price at least twice, drawn from actual swing highs and lows rather than adjusted to make the shape look neater. The lower rail connects genuine swing lows; the upper rail connects genuine swing highs. If forcing the second rail parallel requires ignoring an obvious third touch that sits outside the lines, the channel is not real yet, it is a guess wearing a ruler.
The midline's role
The line running equidistant between the two rails works as a decision point inside the channel. A pullback that holds at the midline and resumes toward the far rail suggests the trend leg is intact. A pullback that pushes through the midline and keeps going warns the channel may be widening or breaking down entirely, well before either rail is actually touched. Watching the midline gives an early read on which rail is more likely to matter next, and it also gives a practical place to take partial profit on a rotation trade that is working, banking part of the gain before price has to travel all the way to the far rail to prove the trade right.
Rotation trades from rail to rail
Inside a valid channel, the basic play is buying near the lower rail with confirmation and targeting the upper rail, or the reverse in a downward channel, with the stop placed just beyond the rail being traded from. This only works with confirmation at the rail itself: a rejection candle, not just a touch, because price can pierce a rail briefly without the channel actually failing.
Sloping channels versus horizontal ranges
A rising or falling channel is really a trend with guardrails, and rotation trades inside it work with the underlying direction rather than against it: buying the lower rail in an ascending channel is buying a pullback within an uptrend, which is a fundamentally different bet than buying the bottom of a flat, horizontal range with no trend behind it. A horizontal range channel carries a higher risk of a clean fakeout through either rail, since there is no directional momentum discouraging a false break. Sloping channels earn a little more trust for rotation trades precisely because the trend gives the rail touch a reason beyond the geometry itself.
A worked example
EURUSD builds an ascending channel: the lower rail rises through 1.0780, 1.0810 and 1.0840, three genuine touches; the upper rail rises through 1.0850, 1.0880 and 1.0910, also three touches, roughly 70 pips of width. Price pulls back to the lower rail near 1.0845 and prints a clear rejection candle. Entry sits at 1.0850. The stop goes below the rail at 1.0833, a 17 pip risk. On a $10,000 account at 1 percent, $100 divided by 17 pips at $10 a pip gives roughly 0.59 lots. The target sits near the upper rail at 1.0905, a 55 pip move, close to 3 to 1. Invalidation is a confirmed close below the lower rail with follow-through, which signals the channel is breaking rather than simply being tested.