It looks for solid companies that have fallen sharply because of a problem that looks temporary, on the idea that the market may have overreacted. This is not a buy recommendation: it is the starting point for your own research.
Every day it checks the price of thousands of companies and keeps the ones that fell at least 8% in a single session, or 12% cumulatively over the last 48 hours. Either gate qualifies; it only counts if it happened within that 48-hour window.
It checks which news item or official filing explains the fall, and that it came before the fall rather than after it. With no cause found, the company is dropped.
It runs 10 checks on cash, debt, margins, the company's response and the sector. Every check cites the figure it is based on.
The score comes from 7 objective factors computed in code, not by AI. Not every company that clears gates 1 and 2 gets published: the funnel below shows how many get fully researched and how many cross that bar on the latest run. Every discarded company is recorded with its reason.
Run on 2026-09-04 20:33.
Today
Tracking
What happened to the previous day's detections over the following sessions. This is after-the-fact measurement with market data, not a forecast or a promise.
Selected
History
Discarded
Showing the most recent ones. See the full list with each company's page.
Results
Theoretical tracking figures, not a track record: with only a few days of history and a few dozen cases, any percentage can be statistical noise. Useful to watch whether the filter works, not to promise returns.
Of the ones Radar Caídas reviewed and did not select, how many still rose the next day. This checks whether the filter is too strict.