I will validate your stock screen against a same day random draw control

Most backtests answer the wrong question.

"My screen made +4% over 20 days" tells you what those 20 days

did. It does not tell you whether the screen selected anything.

I run one comparison: your picks against same-day random draws

of the same size, from the same pool you could have picked from.

WHAT I NEED FROM YOU:

  picks CSV:   date, name

  universe CSV: date, name, fwd_ret

  I am not a licensed financial adviser and hold no financial

  licence or certification. This is a statistical comparison of

  numbers you supply. It states what that comparison supports. It

  does not tell anyone what to do with their money, and it does

  not predict any future result. No specific return, earnings

  figure, or profit rate is stated or implied anywhere in the

  deliverable.

SEE IT BEFORE YOU BUY. The sample report in my portfolio is exactly

this package run on a real pair of files: 99 picks over 8 days, 197

names priced, picks varying 6-22 per day against a pool of 25-59.

Observed -0.4961% against a control mean of -0.4096%, percentile

0.4775, 2,000 draws. The verdict printed was NOT ESTABLISHED. I ship

the failures too.