Portfolio 3 of 6
Full lean
Nine stock funds. 60% in the US and 40% abroad, with the money leaned toward companies that are cheap for their profits and toward shares that have been rising. It has beaten a plain index by a little on both histories and trailed it for a decade at a stretch.
Probably The simulations favour the lean. What it earns from here is uncertain.
What you buy
| Fund | What it holds | Share | Fee on $10,000 | Account |
|---|---|---|---|---|
| VTI Vanguard Total Stock Market ETF | The whole US stock market | 25% | $3 | Taxable account |
| AVLV Avantis U.S. Large Cap Value ETF | Large US companies that are cheap and profitable | 15% | $15 | Taxable or retirement |
| SPMO Invesco S&P 500 Momentum ETF | The hundred large US companies whose shares have been rising | 10% | $13 | Either |
| AVUV Avantis U.S. Small Cap Value ETF | Small US companies that are cheap and profitable | 10% | $25 | Retirement account, Roth first |
| DFIV Dimensional International Value ETF | Large foreign companies that are cheap and profitable | 10% | $27 | Taxable or retirement |
| AVDV Avantis International Small Cap Value ETF | Small foreign companies that are cheap and profitable | 10% | $36 | Retirement account, Roth first |
| IDMO Invesco S&P International Developed Momentum ETF | Foreign companies whose shares have been rising | 10% | $25 | Either |
| VXUS Vanguard Total International Stock ETF | Every stock market outside the US | 5% | $5 | Taxable or retirement |
| AVES Avantis Emerging Markets Value ETF | Cheap companies in emerging markets | 5% | $36 | Retirement account |
Less than half the money, VTI and VXUS, is the plain market. The rest is three ideas. The first and largest is cheapness with a profit test: AVLV and AVUV buy US companies, large and small, that are cheap for what they earn; DFIV and AVDV do the same abroad, and AVES in emerging markets. A plain value fund such as VTV buys cheap companies whether or not they make money, and over the last ten years that version lagged the market by almost 3% a year while funds with the profit test kept up with it. The second idea is momentum: SPMO holds the hundred large US companies whose shares have risen most over the past year, and IDMO does the same abroad. Cheap stocks and rising stocks tend to have their good years at different times, which is why both are here. The third idea is the split: 60% US and 40% foreign, against the world market's 63/37, because foreign stocks are far cheaper and the US lead since 1990 was mostly rising prices rather than rising profits.
Buying rising shares is a different rule for picking stocks. What it adds depends on the fund, on the stocks sold to buy it, and on how far its holdings overlap the rest of the portfolio. SPMO is held at ten percent. Against a five-percent line, moving the second five percent out of VTI added about a tenth of a percent a year on both simulated histories, did not deepen the worst fall, and still added something when the past reward for momentum was cut in half. Ten is a working choice, not a weight anyone has shown to be best. IDMO is also at ten percent, paid from VXUS, since the same test was run on it: about two tenths of a percent a year on both histories, no deeper fall, still positive with momentum's past reward halved, and still positive after charging it an extra one percent a year for the trading its 105% turnover implies. That last charge is the reason it was held at five for a day; the fund model does not price turnover, so the test had to.
For fewer funds, hold the simple lean: the same three ideas with four funds, and about half the edge. DFLV or DFUV can replace AVLV, and AVIV can replace DFIV, at similar cost. A five-fund version of this portfolio, VTI 35, AVLV 20, SPMO 10, DFIV 20 and AVDV 15, matched the nine funds on both simulated histories to within a tenth of a percent a year. It drops the small-value, foreign-momentum and emerging-market lines, so more of the money rides on fewer funds and fewer countries, which the simulations do not price. Hold it if nine funds is the reason you would not hold this at all.
Two things were tried and lost. Replacing the two foreign value funds with the plain foreign index cost about half a percent a year on both histories. Replacing SPMO with a growth fund, one that buys fast-growing companies whatever the price, cost about a third of a percent a year, and trailed in four of every five ten-year stretches. Cheapness and rising prices are the two ideas that pay; growth for its own sake is the one that does not.
AVUV at 15% instead of 10%, paid from VTI, added about a quarter to four tenths of a percent a year in the simulations and made the worst fall on the long history almost two points deeper. Most of that gain rides on small companies simply being small, which is the weakest of the ideas here, so ten is the published weight. Fifteen is a fair choice for someone who wants the largest lean this fund set offers without borrowing.
What the funds themselves did
The histories below are simulated from index data, because most of these funds are young. Here is the one check that uses the funds' own returns, from October 2021 to March 2026: a six-fund value portfolio built on VTV, the plain US value fund, with AVUV, SPMO or both added. Each row is the money left at the end against leaving it alone. A positive number means more money; these are differences over the whole period, not yearly returns.
| Change | Ending wealth difference | 95% uncertainty range |
|---|---|---|
| Replace the 15% VTV holding with AVUV | −0.3% | [−4.6%, +4.4%] |
| Move 5% of the portfolio from VTI to SPMO | +1.4% | [−0.6%, +3.4%] |
| Make both changes | +1.0% | [−3.2%, +5.8%] |
Exploratory · Fund filings and portfolio simulation · as of 2026-09-05
SPMO helped in this short history and AVUV did not. Both are held at ten percent, on the century of index data rather than on these four years. Every range includes no improvement: four years of fund returns cannot settle a question of decades, and the long simulations are what the portfolio rests on.
What this comparison includes
The portfolios start with cash and rebalance yearly. Fund returns already include fees and internal trading costs. We add an assumed 0.05% roundtrip cost for investor trades, including the initial purchase. Personal taxes and a final sale are excluded.
The ranges come from resampling stretches of the same history and rerunning the portfolios. They describe uncertainty in this short sample, not a range of future outcomes.
What $10,000 did
- $10,000 became
- $554,000
- Against $314,000 for a plain world stock index
- Grew a year
- 12.1%
- Against 10.3% for the index
- Worst fall
- −52.5% $4,750
- October 2007 to February 2009; back to even in 62 months
- Costs a year on $10,000
- $18
- 0.18% fee
| Series | Start | Finish |
|---|---|---|
| Full lean | $10,000 in 1990-10 | $553,929 in 2025-12 |
| Simple lean | $10,000 in 1990-10 | $391,970 in 2025-12 |
| Plain world stock index | $10,000 in 1990-10 | $314,164 in 2025-12 |
Best calendar year: 2003, up 40.8%. Worst: 2008, down 37.6%. In 2008 it lost 37.6%; in 2022 it lost 11.6%, against 17.8% for the index.
The worst fall
−52.5% $10,000 fell to $4,750 and took 62 months to get back.
| Series | Worst fall |
|---|---|
| Full lean | −52.5% |
| Plain world stock index | −52.7% |
It was back above $10,000 in December 2012, 62 months after the October 2007 peak. The plain index fell to $4,730 and took 63 months. The lean falls with the market; the hard part is the quiet stretch. Cheap US stocks ran 54.3% behind over 17.7 years from September 2008, and have not caught up. Foreign stocks ran 69.0% behind over 18.2 years, measured against US stocks.
What it beat, and by how much
In the simulation above, this portfolio grew 12.1% a year against 10.3% for the plain stock index. On the longer 1932–2025 history, built from US data, it grew 13.5% against 11.3%. These are modeled outcomes after assumed costs, not returns the funds earned.
The two histories overlap, and both apply today's assumed fund holdings to older markets. Neither can tell you how large a future advantage would be. Weaker rewards, funds that change what they do, and taxes can each leave the lean behind a cheap index for years. Do not plan around those numbers.
Who it is for
Someone who can watch a plain US index fund pull ahead for ten years without selling. Cheap stocks trailed expensive ones between 2007 and 2020, and foreign stocks trailed US ones for most of the same years; a portfolio like this trailed the market by about a third over that stretch, and that is inside what it has done before, not evidence anything is broken.
How to hold it
What could go wrong
The lean can lose for longer than you can wait, and the stretch arrives with no warning. In the worst tenth of stock months the value funds move together far more than usual, so five ideas shrink to about three. A fund that only buys cheap stocks, and cannot bet against expensive ones, captures about half of the paper gap by arithmetic.
IDMO and SPMO are the funds most likely to disappoint. Buying what has been rising beat the market on paper by about 7% a year since 1994, but a paper version's trading cost is as wide as the whole gain, and all three regions crashed together in 2009. IDMO is also more than half Japan at the moment, and its 42% year in 2025 was mostly that. Both are held at 10%. IDMO is the dearer and the more concentrated of the two, which is why its ten had to survive an extra trading charge before it was adopted.
The 60/40 split between US and foreign stocks is a judgement, not a measurement. On 1990 to 2025 a portfolio with more foreign stocks did worse, because US stocks won that window; on the century of US data the same shift did better, because there it is more cheapness rather than more foreign exposure. The split sits between those two answers and leans on the fact that US stocks in 2026 are priced higher, relative to their earnings, than in all but 18 months since 1881.
The long history
Over 1932 to 2025, with foreign markets mapped to US stocks and the cheapness and momentum leans built from US data, $10,000 at the March 1937 peak fell to $4,550 by March 1938, a fall of −54.5%, and took 74 months to get back; all US stocks fell −50.3% and took 52 months. Over that window the portfolio grew 13.5% a year against 11.3% for all US stocks. Nobody earned these returns, and nobody sat through those falls knowing how they would end.