← All 25 termsQuant methodology
CSCV / PBO
QuantCombinatorially Symmetric Cross-Validation, the standard method for estimating Probability of Backtest Overfitting. Tells you how many of your backtest winners would have been selected by chance alone.
Full definition
CSCV (Combinatorially Symmetric Cross-Validation), introduced by Bailey, Borwein, López de Prado, and Zhu (2014), is the standard method for estimating the Probability of Backtest Overfitting (PBO). It partitions the backtest time-series into N blocks, then iterates over all 2^N combinations of training/test splits. The proportion of splits in which the best in-sample strategy underperforms the median out-of-sample is the PBO estimate. A PBO of 50% means half your 'best' strategies would have lost out-of-sample by chance. On this site, CSCV is gate G22 of the G1 to G31 evaluation stack; every quant project reports its PBO with the block count and split count disclosed.
Related terms
Each link opens the long-form definition of a closely related term.
See this term in action.
The /methodology page maps every term to the gate or artifact that implements it on this site.