The test
average entry = (units₁ × price₁ + units₂ × price₂) ÷ total units
risk after = |average entry − new stop| × total units
Compare that against the risk you had before. A correct pyramid leaves it flat or lower.
10,000 units at 1.0850 with a stop at 1.0800 risks $50. Adding 5,000 at 1.0910 gives an average of 1.0870 across 15,000 units. With the stop raised to 1.0870, the risk is $0 — the position is free and three times the original size in exposure.
Leave the stop at 1.0800 instead and the risk becomes $105. That is not pyramiding. That is a larger trade at a worse average price, and the fact that it is currently in profit does not change the arithmetic.
Why the stop must move with the size
The two things have to travel together. Size up without moving the stop and risk grows linearly with the addition; the position is now betting more on the same idea, at a price that is less favourable than the original entry.
The version that works: each addition is accompanied by a stop that has moved far enough to keep total risk at or below the original figure. Done properly, a trend position can end up several times its starting size while never risking more than 1% of the account at any point — and can eventually risk nothing, having locked in the original entry's profit.
The calculator turns red when risk went up, because that is the only failure mode that matters here and it is invisible while the trade is winning.
Pyramiding is not averaging down wearing a different word
They look symmetrical and are opposites.
Pyramiding adds to a position that is right, on a stop that is moving in your favour. Each addition is funded by profit that already exists.
Averaging down adds to a position that is wrong, usually with no predetermined limit. Each addition increases exposure to a thesis the market is currently disagreeing with.
The arithmetic of the average price is identical. The direction of the evidence is not, and that is the whole difference.
Two practical constraints
Additions should shrink. Adding the same size each time makes the average creep towards the latest price, so a normal retracement takes out a position that was deeply profitable. Halving each addition keeps the average close to the original entry.
Liquidity and slippage compound. A position built in four pieces pays four spreads and four commissions, and the later entries are often taken in faster conditions. On a small stop that cost is a meaningful share of what the pyramid is meant to earn.