Auction Market Theory: Why a Handful of Levels Do All the Work

Most traders draw too many lines. Fibs, pivots, trendlines, supply zones, until the chart looks like a knife fight. Almost none of them can tell you why any single line should matter to anyone but them.

Auction Market Theory is the answer to that question, and it is the reason a small handful of levels appear in nearly every serious intraday framework, mine included.

What AMT actually says

Strip the jargon and AMT is one idea. The market is an auction, and its only job is to find the price where business gets done.

Price moves for exactly one reason. The current level has run out of business. If buyers and sellers are happy to transact here, price stays here and volume builds. That is acceptance.

If one side refuses to play at this price, the auction has to move to find them. That is rejection, and it is why price ever goes anywhere at all.

So every move on your chart is the auction doing one of two things. Advertising a price to see who shows up, or leaving a price because nobody did. Trends are a failed search for the other side. Ranges are a successful one. That is the whole theory, and once you see it, you cannot unsee it.

Why levels matter at all

If the market is an auction, then certain prices are not just lines. They are the auction's public record.

A session high is where the auction went looking for sellers and found them. A session low is where it found buyers. The market wrote down, in actual traded prices, where the other side lives. Those records matter because the auction keeps referring back to them. When price returns to a level where sellers showed up last time, the open question is simple and binary. Are they still there?

Two things can happen, and both are tradeable information. The level holds, sellers are still there, and the auction rejects and rotates away. Or the level breaks, the sellers are gone, and the auction has to travel to find the next batch, often quickly, because between old levels there is nothing but empty space and stop orders.

This is also why level breaks and level fails are the raw material of nearly every mechanical setup I test. Sweeps, reclaims, breakouts, fades. They are all just the auction asking its one question at a recorded price, and the trade is your bet on the answer.

Levels are magnets, not just walls

There is a second property, and it is the one most level-drawers miss entirely. These prices do not just react when touched. They pull.

The mechanism is not mystical. Beyond every meaningful level sits a pool of resting orders. Stops from the trapped, entries from the waiting, targets from the positioned. Order pools are business, and the auction's entire job is finding business. So when the current price runs dry, the auction does not wander randomly. It moves toward the nearest place it knows orders are resting, which is, by construction, the last recorded extremes. Unfinished business behaves like gravity.

You see this everywhere once you name it. The overnight high that gets tagged an hour into the session. The gap that fills back to yesterday's close. The old low that price grinds toward on no news at all. None of it is coincidence. It is the auction doing its rounds of the places where counterparties are known to live.

Practically, this is the difference between using levels as reaction points and using them as destinations. A level ahead of price is a candidate target and a directional pull. A level behind price is a reference the auction may come back to test.

I combine these levels with whatever setup I am trading for exactly this reason. The setup tells me the auction is moving. The map of untested levels tells me where it is most likely moving to, and a trade with a known magnet ahead of it is a different bet from one travelling into empty space.

Why PDH, PDL, ONH and ONL beat the rest

Here is the part most level-drawers never think about. A level is only as good as the number of people watching it.

Your trendline is yours. Your fib retracement depends on which swing you anchored, and the trader next to you anchored a different one. These levels are private. The auction does not owe them anything, because almost nobody is positioned around them.

The previous day high and low, and the overnight high and low, are different in kind, for three reasons.

They are objective. Every trader, every algo, every desk on the planet computes the same four prices to the tick. No anchoring, no interpretation, no settings. If a level's power comes from shared attention, these have the maximum possible amount of it.

They are where the unfinished business is. The prior day's extremes are, by definition, the prices where yesterday's auction stopped. Someone got rejected there. Stops accumulate beyond them, from yesterday's winners protecting profit and yesterday's losers hoping for redemption. The overnight extremes are the same story told by the session that cannot finish its own business, because the overnight auction is thin, and its extremes are IOUs that the deep day session frequently comes to collect.

They reset daily. A level from three weeks ago has been fought over, absorbed, and forgotten. PDH, PDL, ONH and ONL are fresh every single morning, which means the positioning around them is live, not archaeological.

None of this makes them magic. It makes them crowded, and crowded is what you actually want, because the auction's behaviour at a price is driven by how many participants have business there.

A level is a location, not a signal

Knowing where the auction will ask its question tells you nothing about the answer. Price touching PDH is not a short. Price breaking ONH is not a long. The level is the venue. The trade is whatever the auction does there, and that behaviour is measurable rather than guessable.

This is exactly where my own testing lives. When I tested raids on these levels mechanically, the interesting finding was never that the levels get hit. It is what happens after. Failed raids on known levels reverse at rates meaningfully above chance in some sessions and not in others, and the same level behaves differently by weekday, by session, and by how far the raid travelled.

The location is free knowledge. The behaviour at the location, measured across years of sessions, is the part that decides whether you have a trade.

The AMT Levels indicator draws the full set on your chart automatically, and it is free on the Tradestar marketplace. The behaviour statistics, what actually happens at each level, in each session, updated as sessions close, live in Market Edge. The lines are the map. The odds are the product.

A level is a fact about the past auction, not a promise about the next one. Trade your own plan.