Is your backtest too good to be true? — PSR · DSR · PBO, the overfitting stats most backtests never show
● runs 100% in your browser — your data never leaves this page
input
drop a CSV here or click to choose optional date col + return columns first column = your chosen strategy extra columns = variants you tried
verdict
Upload returns or load the sample. The sample is 30 columns of
pure random noise with the best-looking one picked as "the strategy" —
exactly the mistake most backtests make. Watch it get caught.
PSR/DSR: Bailey & López de Prado (2012, 2014) · PBO: CSCV — Bailey, Borwein, López de Prado, Zhu (2015)
· from Advances in Financial Machine Learning
Research & education only. Not investment advice.