The Pre-Launch Capacity Review

A backtest run at zero dollars of actual capital doesn't know when it's asking a market to absorb more volume than that market can quietly handle. Every real trade moves the price a little against the trader placing it — market impact — and that cost scales up with position size in a way a small backtest, sized as if it traded for free, never has to confront. The pre-launch capacity review is the step where a research team estimates, before committing real money, roughly how much capital a strategy can run before its own trading starts eating the edge it showed on paper.

The core exercise combines two things: how large the strategy's positions would need to be at a candidate allocation size, and how liquid the instruments it trades actually are. A common rule of thumb caps any single position at some modest fraction of a stock's average daily trading volume — for example, no more than 5-10% of daily volume — on the reasoning that trying to build or exit a larger position than that in a reasonable number of days starts to move the price noticeably against the strategy itself. Applying that constraint across every stock a strategy would want to hold, at a given dollar allocation, tells you whether the strategy can actually be built at that size without breaching the volume constraint on its least liquid holdings.

A strategy's capacity is rarely a single fixed number — it depends heavily on how patient the strategy can afford to be. A signal with a long holding period can spread its trading over many days, absorbing much more capital than the same signal traded impatiently over a single day, because spreading execution out reduces the daily volume footprint even though the total shares traded are the same. This is also why a strategy's estimated capacity in a capacity review is typically expressed as a range with an assumed execution horizon attached, not a single hard number, and why a strategy's most illiquid names — often a small subset of its holdings — are usually the actual binding constraint on total capacity, not the average liquidity across the book.

For example, a small-cap value strategy backtests beautifully at $5 million with no cost model, but a capacity review reveals that at $200 million the strategy would need to hold positions in several thinly traded names representing 25% or more of those stocks' typical daily volume — a level of participation that would take weeks to build or unwind without materially moving the price, meaning $200 million is well past this strategy's realistic capacity even though $5 million looked costless.

What this means in practice

A capacity estimate made before launch sets the ceiling the fund should never exceed regardless of how good live performance looks, and it should be revisited periodically rather than treated as fixed forever, since a stock universe's liquidity and a strategy's own crowding (how many other funds are running something similar) both change over time. Skipping this step is one of the more common ways a strategy that worked beautifully at a small allocation quietly stops working once it's scaled up, for reasons that have nothing to do with the original idea being wrong.