A stock screener you can backtest
To backtest a stock screen honestly, reconstruct each name's metrics as of the historical date using only data available then - no restated numbers, no dropped delistings - and measure how the screen would have performed. Cutonce lets you define a screen, validate it against point-in-time history, and then run the same logic forward on a schedule.
To backtest a stock screen honestly, reconstruct each name's metrics as of the historical date using only data available then - no restated numbers, no dropped delistings - and measure how the screen would have performed. Cutonce lets you define a screen, validate it against point-in-time history, and then run the same logic forward on a schedule.
The pipeline
How to build it
- 1Define the screen once
Filters, scores, and thresholds that encode your thesis, saved as a pipeline you can test and reuse.
- 2Test on point-in-time data
Reconstruct metrics as of each historical date so the test uses only what was knowable then, avoiding look-ahead bias.
- 3Keep the delisted names in
Include companies that were later acquired or delisted so survivorship does not flatter the result.
- 4Run the validated screen forward
Once it holds up, the same logic runs on today's universe on a schedule and delivers a scored shortlist.
A backtest report (hit rate, false positives) plus the same screen running forward, delivering a ranked shortlist on your schedule.
Frequently asked
How do I backtest a stock screen without bias? Use point-in-time data and keep delisted names in the set. Reconstruct each metric as of the historical date so the screen sees only what was available then - that removes look-ahead and survivorship bias.
Why does my screen beat the benchmark on paper but not live? Almost always look-ahead or survivorship bias in the backtest, or overfitting to the tested period. An honest point-in-time test surfaces that before you trade it.
Can the same screen run forward after testing? Yes. Once validated, the screen runs on the current universe on a schedule and delivers a ranked shortlist.
Build this pipeline in Cutonce
Chain data, filters, scoring, and AI nodes on a visual canvas. Free to start.