We publish method, and we publish what would disprove it. We do not publish positions. These are the first four; they set out how the desk measures things, which is the part worth agreeing on before any result is discussed.
A Sharpe ratio reported without the number of strategies tried to reach it is not a measurement. What the correction costs, and why the honest version of a research process is slower than the dishonest one.
Read the noteCapacity is usually quoted as a fund-level number. It is not: it is set by the liquidity of the instruments a signal trades and the speed at which it must trade them. Equities, futures and digital assets differ by orders of magnitude on the same nominal AUM.
Read the noteEvery forecast decays. How fast determines the rebalance frequency, the turnover, the cost, and ultimately whether the signal is implementable at all — which is why we estimate it before we estimate the return.
Read the notePrivate-market datasets promise genuinely independent bets. Most of that promise is destroyed before the modelling starts, by a data problem that inflates the reported information ratio and leaves no trace in the backtest.
Read the note