Screen performance, explained

Calculation

For each 3-month and 12-month window, methodology version 4 evaluates a screen after the signal session closes, using only fundamentals, prices, and financial statements dated as available by that signal. Every preset enters at the following shared SPY session close. Qualifying securities form a fixed, equally weighted basket. We calculate the mean adjusted-close return through the common exit trading date, with no rebalancing.

Financial statements are selected by Tiingo's public release date, not just the fiscal period they describe. Flow metrics use the trailing four consecutive reported quarters; when that history is incomplete, they fall back to the latest annual filing available at the signal. Balance sheet, ratio, share, and per-share metrics use the latest quarterly filing, with annual data used when no quarterly filing exists for the release event. Later amendments become new availability events and do not rewrite an earlier signal.

Moving-average screens use adjusted closes through the signal session and require complete 50- and 200-session windows. No preset assumes it could observe a close and transact at that same close.

Benchmark and exclusions

SPY is shown as an S&P 500 proxy over identical trading dates. Excess return is the percentage-point difference between the screen and SPY; it is not statistical alpha. Securities without a positive adjusted close on both dates are excluded.

Important limitations

Results depend on point-in-time data coverage and may still be affected by survivorship in the current ticker universe. They do not include trading costs, taxes, liquidity constraints, dividends not reflected in adjusted prices, or intra-period rebalancing. These are historical illustrations, not tradable backtests or investment advice.