An experiment in what an agentic-AI research framework can do — published in full so you can check its work, not a tip service to follow. Generic research and education, not advice, not a personal recommendation, not FCA-authorised. Hypothetical book, not real money. Capital at risk.
Portfolio Lab · Research Lab
Portfolio Lab — transparent, low-cost portfolios with real out-of-sample performance, net of all costs. View the portfolios →
Research notes — work in progress. Our evidence-driven analysis of investment signals: which predict returns, which don't, and why. Verdicts evolve as data accrues. Not investment advice.

Signal Research

Which signals actually predict returns — tested on our own data, recent windows, with the founding rule that a claim earns its place only with evidence. Keep decisions and reject decisions both matter. Click a column to sort; filter with the box.

Signal ⇅Family ⇅Verdict ⇅Grade ⇅
Quality / Gross Profitability — is it still true?
Profitable companies have historically beaten the market — but our own pull of Ken French's profitability factor shows no single decade is statistically significant, the 2010s were flat, and the live long-only quality ETF underperformed the S&P. We keep it as an overlay, not standalone alpha.
 Quality OVERLAY ONLY C+
Insider Cluster Buying — our SEC data
Several insiders buying within days. From authoritative SEC Form-345 data across four regime-varied years (2017/2019/2021/2022, ~12,700 events). Measured against the S&P the signal looks value-destructive by 12 months — but that collapse was ~85% small-cap-vs-large-cap beta, not the signal. Against the correct small-cap benchmark (IWM) it is a real +2.7% one-month / +1.6% three-month abnormal return, then flat. A conviction + entry-timing strand, not a hold.
 Insider / ownership OVERLAY ONLY B-
Congressional Trading — the 'Pelosi tracker'
Does following Congress beat the market? The rigorous post-STOCK-Act study finds no edge; our pull of the tracker ETFs shows the recent 'wins' are a tech-beta tilt (Democrat fund just ahead, Republican fund well behind) that reversed in 2022 — plus a ~45-day disclosure lag. We reject it as a signal.
 Information / political AVOID F
Share Buybacks — only the real ones, only when cheap
The buyback basket has underperformed recently and the classic announcement drift is gone post-2002. What survives is narrow: a value-conditioned drift, and only where the share count genuinely falls. We verified real reductions on our names — the rule we already use.
 Capital allocation SITUATIONAL C+
Following the 'smart money' — 13F cloning
Does copying famous investors' 13F holdings work? Broad clones (GURU, ALFA) trailed the S&P by 3–4pp/yr. But cloning the right manager can beat the market: copying Chris Hohn's (TCI) lagged top holdings beat the S&P by ~2.7pp/yr for a decade — following a public filing, 45-day lag and all. These are all superb investors; where a clone lags it's the lag and what a 13F can't show, not the manager. The real edge is in selecting the right manager — a discipline of its own.
 Institutional positioning SELECTIVE B-
Free-cash-flow yield — does buying cheap-on-cash beat the market?
Does buying high free-cash-flow-yield stocks beat the market? Across the S&P 500 and ten annual rebalances we found FCF yield is U-shaped: both the cheapest and the most expensive quintiles beat the market by ~5%/yr while the middle lagged. The headline 'spread' is near-zero only because it shorts a tail that also wins.
 Value / cash flow OWN THE ENDS B+

In progress

Signals we're currently testing. Reports go live once the data is in and the result holds up — no conclusions until then.

SignalFamilyStatus
Short interest
Is heavy shorting a warning or a squeeze setup? Building it from free FINRA / exchange short-interest data — in progress.
 Positioning / sentiment Work in progress
Deep-drawdown + quality reversal
Do quality businesses that have fallen hard revert? Running the indicative version now on the price history we hold, with quality from free SEC filings — in progress.
 Mean-reversion / catalyst Work in progress
Earnings-estimate revisions
Do upgrades to analysts' forecasts predict returns? The hardest to do on free data — scoping a free estimate-trend proxy.
 Earnings / fundamental-momentum Work in progress
Spin-offs
Do newly spun-off companies beat the market? A classic special-situations claim we use in our own screen — building it from a free spin-off list + prices. In progress.
 Corporate actions / special situations Work in progress