What we believe

When a quantitative fund blows up — and they do, regularly — the explanation from the surviving managers is almost always the same phrase: a once-in-a-lifetime event.

The phrase has been deployed for the 1987 crash, LTCM in 1998, the 2000 dot-com unwinding, the 2008 financial crisis, the 2010 flash crash, the 2015 Chinese devaluation, the 2018 volatility blow-out, the 2020 Covid liquidity event, and the 2022 rates-and-tech compression.

Whose lifetime?

The reason the phrase keeps working is that the underlying risk apparatus — ARCH, GARCH, VaR, the standard statistical machinery — is largely calibrated to ordinary-market conditions. Ordinary conditions are the dominant regime, so the apparatus performs well most of the time. The problem is the tail.

The pattern persists for a structural reason. The standard incentive arrangement in institutional asset management compensates the manager well for taking the kind of risk that pays the manager well most of the time and fails the investor catastrophically when it doesn't. The asymmetry is in the manager's favor. The incentive structure makes ordinary-period stability look sufficient, while leaving tail-risk costs to appear only later.

The point of the koan is to refuse the single number — not to claim a different one. VS Asset Management's approach is not the discovery of a better risk number. It is a different question, set out below.

Two further observations follow.

First: alpha decays. Strategies that work become known, crowded, and arbitraged away. Building a firm around a specific trade is building a depreciating asset. (See The Emperor Has No Alpha.)

Second: static drawdown thresholds are not risk management. The fashionable "cut at –X%" stop-loss apparatus, common in institutional risk management, can convert temporary losses into permanent ones. The argument and supporting analysis are developed in the literature. (See Varma, The False Promise of Drawdown Rules: New Evidence and a Better Framework, Journal of Portfolio Management 52(1), 2025, pp. 145–161; and the plain-language version, The Stop-Loss That Stops Gains.)

The alternative to all of this is not a better prediction. It is a different question. Don't ask what is the risk. Ask what regime are we in, and how should we be exposed accordingly? Don't predict risk; classify it. Classification has different failure modes than prediction.

VS Asset Management was founded in 2001. The Risk Timing® methodology has been in use since 2016. The specific implementation is proprietary.