Two sessions agreeing on a price is the evidence you want

Prior day high and prior day low are automatic on most platforms. One click and they appear. That convenience is what makes them poor levels on their own.

Crowded prices are where resting orders sit, and resting orders are what a large move needs to fill against. Price runs through the obvious line and collects the stops parked beyond it. Then it goes where it was going anyway. You get the wick, not the reversal.

There is a second problem. A prior session high is a single fact about a single day. It might be the top of a trend day nobody defended. Or the edge of a range price tested five times. The line looks identical either way and tells you nothing about which one it was.

So the level earns nothing from being yesterday's high. It has to earn a line the same way any other price does, by showing evidence.

The narrow version is worth stating exactly. Take two session highs, from two different sessions, that land close enough together on the price scale. That pair is a resistance zone. Two session lows from different sessions make a support zone the same way.

Neither one has to be a swing point. Neither has to carry a candlestick pattern. The evidence is the repetition. On two separate days, with different news and different traders, price stopped in the same area.

Draw the zone from the outer edges of the pair, wick to wick. It counts as one item against your cap of three supports and three resistances. A single line or a gap box counts the same.

A third session landing in the same area makes it stronger, but does not make it wider. Keep the box where the first two put it and note that price keeps coming back.

A high from Monday and a low from Wednesday sitting at the same price is a coincidence, not a cluster. The pair has to be the same type to mean anything.

The reason is what each one represents. A session high is where buying gave up on that day. Two sessions where buying gave up in the same area is a supply story. Mixing in a low from another day describes a different event entirely, and the two do not add together.

Use the same merge distance you would use for pattern levels on that instrument. On a large, high-priced stock that is roughly a dollar. On futures, work it out from the daily range rather than importing a dollar figure that means nothing there.

And if a real reversal pattern sits inside the pair, the pattern gets its own line. A pattern outranks a cluster every time, because something specific happened at that price rather than just a stop.

This is where the two markets split, and getting it backwards will scatter your levels.

For stocks, use regular hours only. The extended session prints thin, wide candles on small size. A high made in that liquidity is a price nobody defended. Include them and you will draw levels that price walks through at 09:31 without slowing down.

Futures are the exception and trade nearly around the clock. The overnight range is real trading with real participation. The session for a futures contract runs from the open through to the close. Two overnight highs landing together are as valid as two daytime ones.

Which means the phrase previous session means two different things depending on the instrument. Set your chart's session template once per market, then leave it. Half your levels moving because a setting changed is a debugging problem you do not need.

Ten minutes, once, on the timeframe that shows several sessions on one screen. Work backwards from the current price as usual.