How it works

What the screen measures, how the portfolios are built from it, and what the numbers on this site mean. No company is judged on its story or its prospects. The rules run twice a year and the output is published as it comes.

The universe

The 500 largest US companies by company size, minus banks, insurers and utilities — not because they are bad businesses, but because their accounts work differently and the cash measures below produce nonsense for them. A company that stops filing accounts stops being scored: the last published figures would otherwise stand in for a business nobody is reporting on.

What gets measured

Three measures of whether a business actually produces cash, and one of how it behaves. Reported profit can be shaped by accounting choices; cash is harder to flatter.

The fourth is not an accounting ratio at all. Filing lag counts the days a company takes to publish, and it enters the ranking backwards — slower is worse.

A company slower than its peers is a company having trouble closing its books, and it is the only signal we found that is not already contained in the other three: its correlation with each of them runs between −0.07 and +0.10. It is read off the dates on the filing, not the figures inside it, so a screen built on the accounts alone cannot see it at all. It is also what makes the tighter books work: without it, holding fewer companies made the results worse, not better.

From four measures to one list

Each company is ranked on all four measures and the ranks combined, with filing speed entering in the opposite direction because slower is worse. Companies are placed in order on each measure and the places added up, so one freak reading cannot swamp the other three. The precise weighting, the treatment of ties and the handling of incomplete filings are the parts that took the work, and they are not published.

That one ranking has two ends. The books are drawn from the top of it. The bottom 250 is the exclusion list, published free; it decides nothing about the books, which come from so near the top that the cut never reaches them.

A holding is not sold for slipping a few places. At each review a company already in a book is kept while it still ranks in the top 25 — the top 40 for Broad — and only the places left empty are filled, with the highest-ranked companies the book does not already hold. On average 2 of Concentrated's 6 companies change at a review, 5 of Balanced's 12 and 7 of Broad's 20.

Because Concentrated and Balanced keep holdings on the same terms, every company in Concentrated is also in Balanced; the rules guarantee it. Broad keeps holdings longer, so Balanced sitting inside Broad is not guaranteed — a company Balanced drops can stay in Broad. It has held at 58 of the 58 reviews in the record, the current one included.

The fund in two of the books

Concentrated puts 60% of the money, and Balanced 30%, into an exchange-traded fund that aims to return 2 times the S&P 500's daily move. Broad holds no fund. The companies share the rest of the money.

The fund is held only while the index trend is up. At each review the fund is held while the S&P 500 Total Return stands above its average of the last 10 month-ends, and sold only once it is more than 2.5% below that average. When it is not held, its share of the money goes into the companies, in the same proportions: the book is then entirely shares. Across the 58 reviews in the record the fund was held at 83% of them, and the rule switched 8 times. It is held now.

Two things follow, and neither is small. While the fund is held, the leveraged books carry more of the market than the index does: 60% of the money in Concentrated moves with 2x the index, before the companies are counted. And the rule is read only at the reviews, twice a year, so a fall that starts in between is met with the fund still held — and the fund falls about twice as fast as the index on the way down.

The fund in the record is a model, not its market price. The real fund started trading in 2010. For every date in the record, before and after that, its return is worked out from the daily S&P 500 Total Return: 2 times the day's return, less a running cost of 0.60% a year and the cost of the borrowed half at the 13-week Treasury bill rate plus 1.13%. That margin was fitted so the model matches the real fund's record from 2010 to 2026, which means it already carries the fund's actual costs, tax withheld on dividends inside its index included. Moving money between the companies and the fund at a review is charged like any other trade.

How many companies

Three books, at 6, 12 and 20 companies. The screen itself was built on 1998 to 2014, and 2015 to 2026 were set aside and looked at once, at the end. These book sizes, the rule for keeping holdings and the fund were chosen later, with the whole record in view. So the second column below is not a test the books passed unseen. It is what they did.

Read the columns against each other. The lead is smaller on the later years for every book, and smallest for Broad, the book without the fund. In a market that mostly rose, as 2015 to 2026 did, a fund that doubles the index adds to the return almost by construction; that part of a lead says nothing about how well the companies were chosen.

No single company is a large part of any book. Equal amounts put 6.7% of Concentrated in each company while the fund is held, and at most 16.7% at any review in the record. Broad's best-ranked company takes 9.5%. That is a different kind of concentration from the one this site used to warn about: the risk that remains is the fund's, and the small number of companies.

Position sizing

Concentrated and Balanced: equal amounts. Every company gets the same share of the part of the book that is in companies — 6.7% of Concentrated and 5.8% of Balanced while the fund is held, and more when it is not.

Broad: by position. The best-ranked of its 20 companies takes the largest share and each place down a step less, from 9.5% to 0.5%. The weight follows a company's place among the holdings, not its rank among the 500: a holding kept low in the ranking simply takes the last place.

Both are arithmetic of the book's size, so the shape is the same at every review and never reads anyone's market value. That was the lesson of the first version of these books, weighted by company size: it once put 70.8% of one book into the thirty-first best of its thirty-five companies.

The blunter way to calm a portfolio does not work either. Dropping the most volatile companies from the list altogether — or the ones that move most with the market — steadies it and destroys what it is steadying: on the years set aside, a book built that way goes from +5.23% ahead of the index to 1.27% behind it. Choosing steadier companies throws away the thing the screen is finding.

Rebalancing

Twice a year, in January + July, back to the target weights; in between, nothing is traded and the weights drift with prices. The particular months matter more than they should. Running the identical rules — the fund's trend rule included — on all six possible pairs of review months moves the annual lead over the index by 5.4 points on Concentrated, 4.6 on Balanced and 3.9 on Broad. All eighteen combinations beat the index over the full record, the weakest by 2.0 points a year. The months this site uses are the best of the six for all three books. On the other five, Concentrated led by 3.2 to 7.3 points a year rather than 8.6, and Broad by 2.0 to 5.8. On 2015 to 2026 alone, Broad finished behind the index on 3 of the six calendars.

That spread is what a small book does. An earlier version of this method held four to six companies and no fund, and across the years set aside its six calendars finished 6.9 points apart, beating the index by more than three points on one and losing to it on another: what looked like a strategy was substantially a choice of months. Concentrated holds 6 companies again, with most of its money in the fund. On 2015 to 2026 its six calendars finished 3.9 points apart and none behind the index: narrower, and still not small.

One period, worked through

Every headline figure on this site is built from the same small calculation, repeated 57 times. Here is one of them in full, using Balanced over the six months from Jan 2025.

1. Set the book. On the last trading day of Jan 2025 the screen ranks the 500 companies. The 9 of Balanced's 12 companies still in the top 25 are kept, and the 3 empty places go to the highest-ranked companies it did not hold.
2. Read the trend rule. The S&P 500 Total Return stood 7.67% above its average of the last 10 month-ends, so the fund was held: 30% of the money in the fund and 70% in the companies, 5.83% each.
3. Hold for six months. Nothing is bought or sold until the next review. The companies' returns, dividends included and weighted equally, add up to +5.09%. The fund returned +5.50%.
4. Take the costs off. 25% of the companies changed at the review, and 0.20% is charged on that part: 0.05% of the companies' share. Resetting the split between the companies and the fund cost under a hundredth of a percent.
5. Combine them. 70% of the money earned the companies' +5.04% and 30% the fund's +5.50%: +5.17% for the period.
6. Compare it to the index. The S&P 500 Total Return did +5.64% over the same days. This was one of the 18 of 57 periods in which the index won.
7. Chain the periods. Each period multiplies the last. $100 at the start of the record was worth $7,753 by Jan 2025; +5.17% takes it to $8,154. After all 57 periods it is $9,290.
8. Turn it into a yearly rate. The compound annual growth rate is the single steady rate that would take $100 to $9,290 since 1998 — that is 17.23% a year. It is an average, not a description: no individual period looked like it.

Every other figure comes off the same series. Volatility is how much those period returns vary around their own average. The worst fall is the largest drop from any peak in that chain to the lowest point after it. The difference against the index is this yearly rate minus the index's own, computed the same way over the same days.

Reading the figures

“15.20% a year”

The compound annual growth rate: the single steady yearly rate that would take you from the start value to the end value over the 57 six-month periods since 1998. Real years are nothing like steady.

“Worst 6 months: -25.11%” and “worst fall”

The worst six months is the worst half-year the portfolio had in the whole record. The worst fall is the largest drop from a peak to the bottom after it, and on this site it is measured on the half-yearly marks, the values at each review. Marked at every month-end instead, the falls run deeper: -43% for Concentrated, -41% for Balanced and -43% for Broad, against -37%, -37% and -38% on the half-yearly marks. Both miss anything that fell and recovered inside a month, and a book holding a leveraged fund can do a good deal of that.

The benchmark

“the S&P 500” is the S&P 500 Total Return index, with dividends reinvested and nothing deducted for fees or tax. It is measured with dividends reinvested and nothing deducted — no fund fee, no trading cost, no tax — which makes it a harder line to beat than any tracker you could actually buy. Our own figures are net of trading costs, so the comparison runs against us.

Costs

Every figure is after 0.20% on the part of the book that actually changes hands — companies entering and leaving, and money moved between the companies and the fund — and after a 30% loss booked on any holding that stops trading. The fund's own running and borrowing costs are inside its modelled return. A review means about 5 orders for Concentrated, 10 for Balanced and 14 for Broad, counting companies in and out and the fund, plus a top-up or trim on the holdings that stay. Your own broker, and your own tax position, will differ.

Where this is weakest

Every item below inflates the historical record, or understates how it would have felt. Real results would have been worse.

• It is one path through history. 57 six-month periods from Jan 1998 sounds like a lot and is not: they cover the dot-com crash, the 2008 crisis, the long bull run after it and the 2022 fall. Four or five distinct market conditions, not 57 independent tests. A method that happened to suit those conditions would look good here.
• Backtests flatter, and these books were chosen with the whole record in view. Rules tested on the past are chosen, consciously or not, because they worked on it. Published edges lose roughly a quarter of their return when tested on fresh data and more than half once the research circulates. Eleven years were held back and looked at once for the screen. The book sizes, the rule for keeping holdings, the fund and its trend rule were settled afterwards, so for them there is no untouched test.
• The trend rule is the kind of rule the register rejects. Of the timing rules tested here, none survived, and levered overlays showed no advantage that carried over to the years held back. This one did not go through that test. It is a design choice for holding less of the market in a falling one, and its part in the record is not evidence that it works.
• Leverage cuts both ways. A fund that doubles each day's move loses ground in a market that swings without going anywhere, and falls about twice as fast as the index in a crash. The rule is read twice a year, so it cannot step aside from a fall that starts in between. Marked at month-ends, the deepest fall in the record was -43% for Concentrated.
• The fund is modelled. Before 2010 it did not exist, and after 2010 a real holding will differ from the model a little every year. The fund could also be closed or changed by its provider.
• The delisting penalty is an estimate. A holding that stops trading is booked at a 30% loss. Some are taken over at a premium and some go to zero, so 30% is a reasonable average and not a measurement.
• The edge is modest and the ride is not. The worst six months in this test was -25.11%. A few points a year is invisible next to that, and most people who abandon a method do it in exactly that moment.
• Six companies is a small book. The earlier research here found that small books look best on the full record and are the most exposed to luck, and set fifteen as the floor. Concentrated now holds 6, beside the fund; the rebalancing test above is the check on what that costs.

Two things that commonly flatter backtests are handled: the figures include dividends, and the universe includes companies that later went bust or were bought out, not just the ones still trading today.

Sources and timing

Company accounts and daily prices for US-listed companies, including ones that have since been delisted, and for the fund model the daily S&P 500 Total Return and 13-week Treasury bill yields. A company's results are only used once they were actually filed with the regulator, plus a further 60 days of caution — using figures before they were public would let the backtest know things nobody knew at the time. The trend rule reads only month-end index levels up to the review. The record runs to 31 July 2026, the last rebalance it covers.