Why we capped the fund

The most honest number a manager can publish is the one they refuse to grow past.

Every fund tells you its target return. Almost none tell you the asset level at which that target stops being achievable. The second number is more informative than the first, because the first is an intention and the second is an admission.

Ours is $500 million. We do not intend to raise past it. This is not modesty and it is not scarcity marketing. It is the output of an execution study, and the number moved down, not up, when we ran it.

Size and return are not independent variables

There is a persistent idea in asset management that a strategy is a thing you own, and capital is a thing you pour into it. Add more capital, get proportionally more profit. The strategy is the constant; the size is the dial.

This is wrong in a specific, mechanical way.

Every order you send is a request for liquidity that someone else has to supply. At small size, that liquidity is already resting on the book and you pay nothing beyond the spread. At larger size, you consume the resting liquidity and the price moves against you while you are still filling. Past a certain point, you are no longer a participant in the auction. You are the auction.

The cost of this is not linear. It compounds against you, and it compounds precisely at the moments that matter most, because the moments where an edge is largest are the moments where the book is thinnest.

The evidence is not controversial

Renaissance Technologies' Medallion Fund is the most successful trading operation in recorded financial history, with widely reported gross returns averaging around 66% annually over three decades. It has been closed to outside capital since the early nineties. It caps its asset base in the region of $10 billion and returns profits to its partners every year rather than compounding them.

Consider what that means. The people with the best strategy anyone has ever documented, with unlimited access to capital, chose to give the capital back. Not because they lacked ambition. Because they had measured where their edge ended, and they respected the measurement.

Warren Buffett has said the same thing from the other direction. Asked what return he could generate on a small asset base, he answered that on a million dollars he could make 50% a year, and that his size was the anchor on his results.

The largest multi-strategy platforms in the world today run tens of billions of dollars and deliver returns in the low-to-mid teens net of fees. This is not a criticism. Those are outstanding institutions and, on a risk-adjusted basis, superb products. But they are a different product. They have solved for capacity, and capacity was solved by accepting a lower return per dollar. That was a rational trade. It was still a trade.

In digital assets, the constraint binds earlier

Bitcoin is a multi-trillion dollar asset. That figure is almost irrelevant to a systematic trader.

What matters is not market capitalisation. What matters is the depth of the perpetual futures order book at the price you actually want, at the minute you actually want it. That number is orders of magnitude smaller, it varies enormously by venue and by hour, and it collapses exactly when volatility expands.

Our system operates around dislocation: liquidity sweeps, funding capitulation, value rejection, compression breakouts. These conditions exist because the book is temporarily unable to absorb what is being thrown at it. That is the entire source of the opportunity. A strategy that feeds on thin liquidity cannot then deploy size as though liquidity were deep. At sufficient scale, the fund stops trading the dislocation and starts being it.

Our internal execution work puts the ceiling on a single venue, using patient limit-maker execution, at roughly $150 to $300 million. Beyond that band, the arithmetic still works, but only with multi-venue distribution and a dedicated algorithmic execution layer, which is a different firm with different risks and a different cost base. We have not built that firm. So we do not price ourselves as though we had.

$500 million is the honest middle of a range we measured, not the top of one we hoped for.

The obvious objection

Every emerging manager says they are capacity-constrained. It is the oldest line in the business, because it converts the inability to raise into the appearance of discipline, and because it costs nothing to say.

So the claim is worthless unless it is structural. Three things make ours structural rather than rhetorical.

It is a hard close, not a soft one. There is no mechanism by which a sufficiently large allocator reopens the fund. No side letters, no capacity carve-outs, no separately managed account that runs the same book alongside the fund.

Profits above the ceiling are returned rather than compounded. If performance carries the asset base through the cap, capital goes back to investors. Compounding a fund past its own measured capacity is the most common way good strategies quietly become mediocre ones, and it happens without any single decision being taken.

The fee structure is aligned with the constraint. A capped fund collects a small fraction of the management fee available to a large one. We built the economics to pay us for return rather than for assets, because a manager paid on assets will always find a reason why the capacity estimate was too conservative.

The question to ask any manager

At what asset level does your edge stop working, and how did you measure it?

If they have a number and can show the work behind it, they have studied their own execution. If they have no number, they have never looked. And if they tell you the strategy scales indefinitely, they have told you something important without meaning to: what they are selling is not an edge, it is exposure.

Capacity is a constraint, not a virtue. There is nothing admirable about being small. What matters is which way the causation runs. We sized the fund to fit the edge. The alternative (and it is the industry norm) is to size the edge to fit the fund, and to discover the consequences slowly, over years, in returns that drift toward the index while the fee base grows.

We would rather publish the ceiling now than explain it later.