Averaging Down: The Math, and When It's Just the Disposition Effect
You bought at $100. It’s at $80. Buy the same amount again and your average cost drops to $90 — suddenly the stock only needs to climb back halfway for you to break even. Averaging down feels like a cheat code, and the feeling deserves a careful audit: part of it is real arithmetic, and part of it is a documented bias doing an impression of a strategy.
The arithmetic part — all true
Average cost is a weighted average, so every purchase below your current average pulls the average toward it:
- Equal dollars at $100 and $80 → average $88.89 (the harmonic mean — equal-dollar buys weight the cheaper price by share count, so the average lands below the midpoint).
- Equal shares at $100 and $80 → average $90 (the simple midpoint).
The stock average calculator does the general case for any lot list, and its target-average solver runs the question in reverse — how many shares would it take to reach a $90 average? The solver’s formula, n = S·(a − c) ÷ (p − a), carries the sobering part in its denominator: the closer your target average sits to the current price, the more shares it takes, diverging to infinity as the target touches the price. Holding 10 shares at $100 with the price at $80, a $90 average costs 10 more shares — but an $81 average costs 90 more shares, nine times your original position.
That is the honest shape of averaging down: modest improvements are cheap; dramatic ones require betting multiples of your original stake on the same idea that is currently losing.
What the arithmetic doesn’t say
Nothing in the math above contains any information about whether the price will recover. A lower average changes where break-even sits, not whether it will be reached. The same purchase can be described two ways with identical numbers:
- “I’m buying more of a business I understand, at a better price than before.”
- “I’m increasing my exposure to my largest loser so that my screen goes green sooner.”
The first is a thesis about the asset. The second is about your reference point — and reference-point behavior is precisely what Odean (1998) documented in 10,000 real accounts: investors systematically hold losers and, symmetrically, treat “getting back to even” as an achievement, though the market has no idea what anyone paid. (The full story is in our disposition-effect article.) When the only argument for a purchase is that it lowers your average, the purchase is about the anchor, not the asset.
A three-question audit
Descriptive, not prescriptive — three questions that separate the two cases in your own trade list:
- Would you buy this at today’s price if you held none? The averaging-down purchase and a fresh purchase are the same transaction; only the anchor differs.
- What does the position become? Run the full sequence in the position calculator — concentration creeps: two equal averaging-downs make the name three times your original exposure.
- Where does it end? A pre-decided limit (“two adds, then the thesis was wrong”) is what distinguishes a plan from a ratchet. The loss-recovery table shows why the ratchet’s endgame is brutal: each further decline raises the required recovery non-linearly.
Averaging down is arithmetic in service of whatever is driving it. The calculator can show you exactly what a purchase does to your average, your exposure, and your break-even — it can’t tell you which of the two descriptions above is the true one. That part is yours.