One point per company, at its own most recently filed quarter. Where the US market sits right now.
Results commentary lifted from 2,954 companies’ own 10-Q and 10-K filings — searchable across 11.3M characters of what they wrote about their own quarters.
Quality gates applied to the chart and the table below.
A strategy is only worth running if it beats what you would have got by picking at random. The gate exit rule is the test case here, and this is the test.
A company is eligible only if it clears all six: revenue growth ≥ 20%, positive net income, positive free cash flow, ROIC ≥ 10%, net debt ≤ 3× EBITDA and FCF conversion ≥ 50%. (A trailing-beta ≤ 1.3 filter is also applied, which changes the risk more than the return.)
The twenty-five highest composite scores among those that pass, held in equal amounts, with no sector cap. Fewer than 25 qualify when the market is dear, so the book is smaller in those quarters.
Formed on the first trading day of each quarter. A holding is cut once it stops qualifying, and a winner is kept only while it still makes the new 25 — about 22 names at a time, held a mean of 2.2 quarters.
Beating the S&P would only show these companies are smaller than the S&P. So each calendar year is compared with 2,000 random baskets of the same size, drawn from the same scored companies over the same quarters and compounded the same way. That asks the real question: was the choosing worth anything? The S&P is drawn too.
Over the twelve full years 2014–2025 the book returned +650%. The scored companies it drew from returned +156%, so the choosing beat the pond it was fishing in by roughly ten points a year. It cleared the range of luck in four of the twelve years, and the typical year sat at the 73rd percentile of random selection — better than chance more often than not, and in no year meaningfully worse.
The S&P 500 returned +353% over the same years. $10,000 became $75,000 in the book and $45,300 in an index fund — both in dollars, since a sterling investor would also have worn the exchange rate, which is not modelled here. The book beat the S&P in eight of the twelve years.
The book’s own reckoning is more careful, and it is worth repeating. Split in half, the return advantage flips sign — about +6 points a year in 2014–2019 and about −5 in 2020–2025 — so the outperformance is not something to bank on. What is robust is the risk: the beta cap holds the book’s own beta near 1.05 in both halves and lowers volatility, at no cost to Sharpe. Read the chart as evidence the choosing beats a coin toss, not as a promise to beat the index.
Read it this way. Each marker sits at what the book actually returned that calendar year; the grey band behind it is where nine out of ten random books of the same size landed, and the dashed line is the S&P. A filled marker is a year that finished clear of the band; a hollow one is a year the book cannot be told apart from luck, whether it landed inside the band or below it. The book cleared the band in four of the twelve full years, and the median year sat at the 73rd percentile of random selection. 2026 is still running.
The return edge is not stable. It is large in some years and negative in others, and splits into a positive first half and a negative second half — about +6 points a year to 2019, about −5 after. A book that wins big in a handful of years and gives some back in others is harder to live with than a steady one, even where the long-run total looks good. The robust part is the risk control, not the outperformance.
The method was designed while looking at this data. The six gates, the composite score and the 1.3 beta line were each decided by examining this history and fixing what looked wrong. The data each year is out-of-sample by construction; the method is not. The only real cure is to fix the rules now and watch years that have not happened yet.
What is genuinely solid. The universe at each quarter is rebuilt from companies that had a price and had filed accounts on that day — including the ones that have since delisted, priced at their last real print rather than dropped. Two thirds of the companies in the price history no longer trade. Leaving them out is the single most common way a backtest flatters itself.
Ranked by composite score within the companies passing the gates above, so this table changes when the screen does. Operating companies only: Financials and Real Estate are scored on a separate, weaker model and are listed below them — the two are not comparable.
Scored on P/E, ROE, growth and payout only: free cash flow and enterprise value are meaningless for them. A leveraged mortgage REIT can look cheap on P/E while being the most rate-sensitive thing in the index.
Enter your own holdings and weights to see how the book scores against the same five factors and six gates used everywhere else on this page. Nothing is saved or sent anywhere — the calculation runs in your browser against the figures already loaded.
Weights can be percentages, fractions or any units — they are normalised to the total you enter, and the raw total is reported so you can see if it is not 100.
Enter holdings on the left, then press Analyse.
Same axes throughout. Each panel shows one sector against the full market in grey, so the panels are directly comparable.
Sorted by market cap. Blank means not computable from the filings — never zero.