Screener

Filter the latest scan's published facts — sector, size, yield, valuation — or rank them with a preset: cheapest, fastest growing, most profitable. Every view shows what is in the filings and prices; none ranks opportunity. The model behind the statuses showed no predictive power, and nothing here is a recommendation.

Scan of 2026-08-31 · 2,409 names scored · prices as of that date

Cheapest

Lowest valid-in-sector trailing valuation multiples today, latest scan: each name's sector percentiles on earnings yield, FCF yield, P/S, P/B, P/TBV and P/CFO, averaged (only the metrics valid in its sector — a REIT is never ranked by GAAP P/E here). “Cheapest” is a fact about trailing multiples versus sector peers, not a forecast that they re-rate.

Tested the unforgiving way: a composite ranking built from these same point-in-time filings (2017-01-31 to 2026-05-29) showed no measurable predictive power (mean IC 0.0168, t = 1.21), and its top quintile did not beat holding everything even before costs. Ranks like these describe today's record; they do not predict tomorrow's return.

161 names — page 1 of 4 · sector: Industrials · Export CSV

#	Ticker	Company	Sector	Cheap composite	P/E (trailing)	P/S (trailing)	FCF yield	Own 5y
1	GBX	The Greenbrier Companies, Inc.	Industrials	96.1	9.4×	0.5×	11.9%	+1.3σ

Three lenses

Never the absolute multiple alone. A P/E of 9 is a fact; whether it is low depends on what you hold it against.

Absolute: The figure itself, from trailing filings against today's price.
Sector percentile: Where it sits among validpeers in its own sector. On a preset this is the rank column itself — the composite is the mean of its components' sector percentiles, so it has no separate one.
Own 5-year range: Where it sits against its own history — the Own 5y column, in sigma.

The five-year window is the empirically privileged one. Cheap against its sector and cheap against its own history are different claims, and a stock can be one without being the other — which is the whole reason both columns are shown rather than blended into a single score.

Left out on purpose

Forward P/E, PEG, anything NTM: No SEC filing contains a forecast. Every forward multiple is somebody's estimate wearing a filing's authority.
EV/EBITDA: Long-term-debt tags cover about a third of the universe. A ranking computable for a third of the names is a lie of omission — the names with the tag would silently become the list.
P/B outside financials: Buybacks have pushed hundreds of sound companies to negative book equity. The ratio is meaningful for banks and misleading nearly everywhere else, so it is masked rather than printed.

Pigglet's screener re-cuts the same monthly scan shown everywhere else on this site and is provided for information only. It is not investment advice, not a recommendation to transact, and not personalised to your circumstances. The ranking behind the statuses was tested on survivorship-free, point-in-time data (2017-01-31 to 2026-05-29) and showed no measurable predictive power (mean IC 0.0168, t = 1.21); its top quintile did not beat a plain equal-weight portfolio even before trading costs. The metric presets rank descriptive facts about trailing filings and carry the same finding. Past performance does not predict future results.