Every number we publish, and where each one comes from
Coverage, the backtest and the live portfolio as one table of figures, each linked to the page that explains its method. A figure without its method is a claim, not evidence.
Last updated September 15, 2026
Every number we are willing to stand behind, in one place, so nobody has to reassemble them from eight pages. Each one links to the page that explains it, because a figure without its method is a claim, not evidence.
The work behind the number
- 13
- years tested, point-in-time
- 12,038
- filing dates pulled from EDGAR
- 973
- companies re-scored every month
- 0
- manual overrides in the test
These four are the headline. The rest of this page is the detail behind them.
Coverage
| Figure | Value | Where it comes from |
|---|---|---|
| Companies in the universe | 973 | What we cover |
| Sectors represented | 11 | What we cover |
| Primary listing markets | US, with Canada, the UK, Europe and Asia | What we cover |
| Data cadence | End of day. No intra-day prices, by design | How fresh the data is |
| Fair-value methods run in parallel | Up to 4 | Fair-value methods |
| Pillars measured on every company | 5 | The Returnolio score |
| Insider and political history per company | 12 months | Smart Money |
The test
| Figure | Value | Where it comes from |
|---|---|---|
| Years tested | 13 (2013 to mid-2026) | The backtest |
| Filing dates pulled from EDGAR | 12,038 | The backtest |
| Manual overrides during the test | 0 | The backtest |
| Trading costs modelled | 20 bps round trip (0.2%, the buy and the sell together) | The backtest |
| Capital invested during the test period | None. It is a retrospective test | The backtest |
What the test returned, including the parts that went badly
| Figure | Value | Where it comes from |
|---|---|---|
| Model, full period, after costs (measured 9 Oct 2026 on the 9 Oct data) | 26.2% a year | The backtest |
| Passive blend (a plain mix of indexes), same period | 16.2% a year | The backtest |
| Calendar years the model was ahead of the blend | 9 of 13 | The backtest |
| Rolling five-year windows (overlapping five-year stretches) ahead of the blend | 114 of 114 | The backtest |
| Sealed hold-out, 2022 onward | 29.5% a year vs 12.3% | The backtest |
Read the whole table, not the first row
Backtested results on historical data, not live returns. Single years are noisy: four of the thirteen went to the blend. The number also moves when the underlying financial data is corrected, which is why each figure carries its measurement date. The backtest page gives all of it, year by year, and says what the published list adds that the test does not: a person who reads every name before it goes out.
The portfolio
| Figure | Value | Where it comes from |
|---|---|---|
| Names in the Monthly Top 10 | 10 | How the Top 10 is chosen |
| Position sizing | Bought at equal weight, then left to run | How the Top 10 is chosen |
| Maximum companies from one sector | 4 | How the Top 10 is chosen |
| Borrowed money used | None | How the Top 10 is chosen |
| Holdings disclosure | With the monthly edition, after the fact | How the Top 10 is chosen |
Common questions
How many stocks does Returnolio cover?
A little under a thousand companies across 11 sectors, mostly US-listed, with Canada, the UK, Europe and Asia represented. The list is deliberately short: a ranking only means something if every company in it has been measured the same way, and plenty of listed companies do not have the filing history to allow that.
How long is the backtest, and is it point-in-time?
Thirteen years, 2013 to September 2026, scored to the day each filing actually reached the SEC rather than to the quarter it covers. 12,038 filing dates were pulled from EDGAR so the model never sees a number before the market did.
Did the model beat the market in every year?
No. It was behind the passive blend in four of the thirteen calendar years. All of it is published, because a record with the bad years removed is not a record.
Thirteen years tested, 12,038 EDGAR filing dates, five losing years published, and a person reading every list before it goes out.
Educational and proprietary. This explains what our research does, not the exact formula behind it, and it is not personalised investment advice. See the full disclaimer.
Was this page helpful?