Myth: Averaging Down Is Just Buying at a Better Price

The myth: if a setup was worth taking at 100, it is worth more at 98. Adding on the way down lowers your average entry, so the trade needs a smaller bounce to come good. Institutions do it. It is called scaling in. It is not the same as being stubborn.

The test: two traders, the same market, the same driftless price. Distances below are written in G, one unit of the original stop distance, and price moves in ticks of 0.01G. Trader A takes one position with a stop 1G below entry and a target 2G above. Trader B enters the same size at the same price, adds another full unit at −0.5G, adds a third at −1.0G, stops the whole ladder at −1.5G, and exits at the average entry plus 0.35G once averaged.

Note what that means before we go any further, because it is the single most important line on this page: Trader B is risking three times as much as Trader A. Three full units, stopped 1.5G, 1.0G and 0.5G from their entries, is 3 unit·G against Trader A's 1. That is not a flaw in the test — it is the whole reason the test exists. Almost nobody who averages down decided the full ladder size in advance. They took a position, it went against them, and they took another one.

These outcomes do not need simulating. A random walk between two barriers has an exact solution, so every number below is arithmetic, not sampling.

The myth is true. It is just true about the wrong thing.

Averaging down works exactly as its defenders claim. The win rate goes from 33.3% to 72.5%. Nearly three trades in four now end green. If you judged a method by its win rate — and most traders do, at least emotionally — the ladder looks like more than twice the trader.

And the expectancy of both is exactly zero. Not approximately. Identically, to the last decimal, because in a market with no drift the price is a fair game and no arrangement of entries and exits can manufacture an edge out of it. Averaging down did not create a single unit of profit. It redistributed the same zero into a different shape: many small wins (average +1.14R) and an occasional loss of −3R, arriving 27.5% of the time.

This is the entire myth in one sentence. Averaging down does not improve your edge. It improves your win rate, which is not the same thing and is much easier to mistake for progress.

What it costs, since it does not cost expectancy

Run both through 40,000 accounts, 250 trades each, with 1R defined as 1% of equity — so Trader A risks 1% per trade and Trader B, holding three units at the stop, risks 3%.

Identical expectancy, seven extra points of median drawdown and eleven extra at the 95th percentile. And be precise about the cause: that is not the averaging structure being inherently worse, it is the tripled risk. Same edge, three times the size, and the account pays for it in exactly the place a trader feels it.

Now the version almost nobody runs

There is an honest defence of scaling in, and it deserves its own numbers. Run the identical ladder with the size pre-allocated — a third of a unit at each of the three entries, so that the full ladder at its stop loses exactly what Trader A's single position loses — and the picture inverts:

Pre-allocated, the ladder is the smoothest of the three — because its average entry is lower, so the same total risk is spread over a wider price band. It still creates no edge; expectancy is zero in all three columns. But it is not the thing that damages accounts.

So the distinction that matters is not scaling in against single entry. It is whether the total size was decided before the first entry or discovered afterwards. One of those is a position-construction choice. The other is a reaction being sized by discomfort.

(If you are wondering why both medians sit slightly below the $10,000 start on a zero-expectancy system: that is volatility drag, the arithmetic reason a 10% loss needs an 11.1% gain to undo. It is covered in the leverage myth.)

Now make the trade slightly wrong

Everything above assumed a perfectly fair market. Real traders average down for a specific reason: they believe the position is right and the market disagrees. Sometimes the market is correct.

Tilt the walk — 49.5% up ticks instead of 50%, on the 0.01G tick stated above — and the single entry loses 0.95R on average while the ladder loses 1.77R. A small adverse drift costs the averager 86% more, because the ladder has quietly arranged for maximum size to be held at maximum adverse excursion. That is the structural property nobody advertises: the averaging ladder is largest exactly when it is most wrong.

Where the real damage happens

Everything above assumed the ladder was planned: three units decided in advance, a hard stop honoured. Almost nobody runs it that way, and the versions people actually run are worse than the maths above in three specific ways.

The size was not pre-allocated. The trader entered a full position, then added a second full position, then a third. Risk did not stay at 1%. It became 3%, discovered after the fact.

The stop moved. The 72.5% win rate is intoxicating, and each save teaches the same lesson: it always comes back. The first time it does not come back, the trader who has been rewarded twenty times for holding will hold once more.

The exit rule got worse. A ladder that was going to exit at the average entry plus a little becomes a ladder that will exit “at breakeven,” then “at a small loss,” then whenever the margin department decides.

Every large trading disaster in the stories section is this pattern at scale: Barings, the London Whale, the Hunts. In every case the trader was averaging into a position they were sure about, with a win rate that had been excellent right up until it wasn't.

When scaling in is legitimate

There is a real technique here, and it is worth separating from the myth. Planned scaling means the full position size, the add levels and the invalidation are all defined before the first entry, and the total risk at the final stop equals what you would have risked on a single entry. That is a position-construction choice with a defensible rationale — better average price in a zone, less exposure to being early — and it is what the simulation above actually modelled.

Averaging down means adding size because the trade is losing. The distinguishing question is brutally simple and worth asking out loud before every add: was this level in the plan before I entered? If the answer is no, the add is not a strategy. It is a reaction, and it is being sized by discomfort.