The portfolio

Those seven numbers total 100, and they are the whole instruction.

We can't tell you this beats a cheap index fund.

Set it beside an index fund that borrows as much as it does, and over 427 months it came out +2.2 points ahead, plausibly +0.1 to +4.6 a year. The smallest gap that test could reliably see was 2.83 points a year. So we measured something positive, and our instrument is not sharp enough to confirm it.

Two comparisons, two different waits. Against the index fund that borrows, your own account would need 12.2 years before it could tell. Against a plain cheap index fund holding the same mix, 31 years.

Both are longer than you have.

What we can tell you is why each line is there, where it goes, and what it will do to you in a bad decade.

The lines

What each piece is for

RSST buys you a dollar of US stocks and a dollar of managed futures for every dollar you put in. Managed futures is a rule that goes long things that have been rising and short things that have been falling, across bonds, currencies, commodities and stock indexes. It tends to be flat when stocks are fine and helpful when stocks fall slowly. This is the one line in the portfolio that is not stocks.

VTI is every US public company, weighted by size, for 3 basis points a year. A basis point is a hundredth of a percent. Cheapest thing you own, heaviest lifting.

VXUS is the same idea outside the United States, developed and emerging together.

VTV leans the US half toward cheap companies. Cheap here means a low price against what the books say the company owns, and the lean is measured rather than assumed: how much cheap-stock exposure you actually get reads +0.337 over six years of its own filed returns.

AVDV does the same thing in smaller foreign companies, which is the only place a fund can buy you that particular corner of the market. Adding it was worth +0.28 points, plausibly +0.05 to +0.56 a year against the same portfolio without it, and the smallest gap that test could reliably see was 0.29. So it misses being a real result by a hundredth of a point, which is the most annoying number on this page.

IDMO buys foreign shares that have been going up. Momentum is the least comfortable holding here and it is only 5%.

AVES buys cheap emerging-market shares. Emerging markets pay the biggest measured reward for owning cheap stocks in our data, and the least reliable one.

Two of those seven lines carry no bet at all. VTI and VXUS are the benchmark, and most of RSST is US stocks you would have owned anyway. The active part of this portfolio is worth about 3.71 genuinely independent bets, not seven.

Placement

Which account each piece goes in

Two thirds of your money sits in accounts the tax system leaves alone. That shelter is scarce, so the question is what a sheltered dollar saves you, and the answer is not the usual one.

Four things about that table are worth your attention.

Every foreign fund outranks every US fund. You have probably read the opposite advice: hold international in the taxable account so you can claim the foreign tax credit. It is wrong for someone shaped like you, and the reason is arithmetic. Foreign countries withhold tax on dividends before you see them. Inside an IRA or a Roth that money is gone for good, because a credit offsets a US tax and a sheltered account owes none. Sheltering all four foreign funds destroys 7.4 basis points a year of credit, permanently. It buys back far more than that, because foreign funds yield more and a bigger dividend in a taxable account is a bigger tax bill. VXUS saves 65 basis points per dollar sheltered against VTI's 25 basis points. The credit is not close to deciding it.

A total-market US fund is last, every time. VTI is the cheapest, broadest, calmest fund you own, which is exactly why it belongs in the account with no protection. Its 1.07% dividend is the smallest tax bill you can put on a scarce sheltered dollar.

A 5% momentum fund outranks almost everything else. IDMO turns over 105% of its portfolio a year and only a quarter of what it pays out gets the low dividend rate. ETFs normally shrug off turnover through a trick called in-kind redemption; at 105% the trick stops working. Nothing about the size of that line tells you it is second in the queue.

Roth and traditional are not the same account. The recurring tax drag is identical in both, so drag cannot choose between them. What separates them is that Roth growth is never taxed, while a traditional balance is shared with the government at whatever rate you eventually withdraw at. At a 24% withdrawal rate, $100,000 of traditional IRA is $76,000 of your money. So the Roth should hold the highest expected return, and the traditional should hold the thing you are least sure about. That puts the funds leaning toward cheap and profitable companies in the Roth and the trend fund in the traditional, which is the reverse of the usual rule. Two smaller facts point the same way. The government shares in RSST's swings as well as its average, which is a feature when that fund is the least established thing you own. And required minimum distributions force the traditional account and never the Roth. RSST after a strong trend year is exactly the position you would be trimming.

Getting this ordering right is worth +2 to +7 basis points a year against what you would otherwise have done. Not more. Numbers like "asset location is worth 75 basis points" come from comparing against nobody's real portfolio.

The thing that breaks it

If part of your tax-deferred third is a workplace 401(k) rather than a rollover IRA, its fund menu can invert this whole table. A typical menu offers a US index fund, a foreign one, and an emerging one. Of the seven funds here, only VTI and VXUS can go in it.

Write f for the share of your tax-deferred money sitting in a rollover IRA. The plan above shelters VXUS at 16% of the portfolio, so a workplace plan smaller than that is free. Below 0.52 it starts costing you. At the extreme, with the whole tax-deferred third captive, VTI gets forced into shelter at 17.3% while AVDV and AVES get evicted to taxable. That is the exact inverse of the ranking, imposed by a fund lineup. It costs 9.1 basis points a year, which is larger than the entire ordering decision you just made.

Rolling an old employer balance into an IRA buys most of that back for the cost of a form. It is the cheapest thing on this page.

The one choice

The one choice you have to make

How much RSST. We recommend 30%. The alternative is 25%, and here is the actual trade.

One measurement favours 30%, and it is worth being exact about what it compared. The higher-trend portfolio in that test was an earlier eight-fund draft; the lower-trend one was the seven funds above. The 30% version won by 0.50 points, plausibly 0.23 to 0.77 a year and the smallest gap that test could reliably see was 0.39 points a year, which makes it the only whole-portfolio comparison in this entire project that clears its own resolution. Almost everything else here is a shrug.

Two things it does not settle. The pair differed in four of its holdings as well as in how much trend it held, so the gap is not purely the five points. And what the five points buy is mostly borrowing rather than trend: the 30% version leaves you exposed to 132% of your money against the 25% version's 127%, and over a stretch where US stocks returned 9.83% a year, more exposure wins.

Nobody has run these exact seven funds at 30%. The test that would settle the question is the same seven funds at both weights, and it is on the list of things this project has not done.

The cost is that you might sell it. Over thirty years, at a trigger where you quit once the fund has run 20% behind, a 30% weight ends the position in about 17% of simulated paths against 11% at 25%. If trend following has stopped paying entirely, 30% ends it in 66.7%. Read that last number properly. Whether you can hold the position and whether it pays are the same bet, made twice.

We say 30% for you, for three reasons.

1. The comparison that favours it is the only resolved one we have, imperfect as it is. Throwing that away for four methods that disagree with each other is choosing the weaker evidence.
2. You are contributing 5% to 15% of the portfolio a year. New money at low prices is the thing that carries a position through a drought, and it covers the one rebalancing move this portfolio cannot make internally two and a half to seven and a half times over.
3. Zero is the worse mistake. Through the one flat decade in our data, March 1999 to February 2009, US stocks returned −2.55% a year and a 30% trend holding added about +9.5 points a year above the stocks it displaced, turning that decade positive. Against +0.21 points a year in ordinary decades, that one decade is what decides whether a plan survives.

Take 25% if you think you would sell it. A portfolio you keep beats a better one you abandon, and the half point a year you give up is invisible against the 6.0% your returns already wander from the comparison in a typical year. The switch is one swap: RSST 25, VTI 24, everything else unchanged.

Cost

What it costs

The funds charge 38 basis points a year in total, weighted by how much of each you hold. After the money the funds earn back by lending out their shares, the real number is 36 basis points. Almost all of it is RSST, at 0.99% on 30% of your money. Six of the seven funds cost you under 31 basis points each and three cost under 3.

Then there is the spread, which is the gap between the buy and sell price, and you pay it once. RSST's is 0.09%, which for this category is fine. The whole entry cost across all seven lines is roughly 3 basis points of the portfolio, one time. Worth knowing, not worth optimising.

One fund we did not pick is worth a line. CTAP does the same job and advertises a 0.10% fee; its real all-in cost is about 0.81%, because a 0.75% affiliated fund hides inside a swap. Its spread is 0.33%, so a round trip costs 66 basis points. Against RSST on a one-year hold that is a 24 basis point a year difference, on a shelf where the whole fee gap between the candidates is 18. RSST was already the pick on age and size.

The last cost is new and it might be the largest. On 1 June 2026 Fidelity began charging $100 per purchase on ETFs from issuers that refused to pay it platform fees. Schwab has said it will do something similar by the end of 2026. On a $10,000 buy that is 100 basis points on day one, which is bigger than any annual fee here except RSST's. The issuers exposed are the small ones, and RSST's issuer is small. No shelf issuer appears on any published list, but Fidelity's live list refuses automated reads, so treat this as unchecked rather than clear.

Check it before your first purchase. If your broker charges it on RSST, buy that line once or twice a year and point your monthly contributions at the index funds. Twelve $100 fees a year on a $200,000 portfolio is 60 basis points; once a year is 5.

Operating it

How to actually run it

Read this bit twice.

The seven numbers in the first table are shares of the money you put in. They are the only numbers you ever type into a brokerage screen. Elsewhere you will see this portfolio described by what it leaves you exposed to: 66% US stocks, 36% international, 30% managed futures, 132% in total. Those are audit figures. They are recomputed every quarter from RSST's latest filing, and they add to 132 rather than 100 on purpose, because RSST gives you two dollars of exposure for every dollar you put in.

Type the exposure figures into a screen that wants weights adding to 100 and it will quietly scale everything down by 24%. Your 30% trend holding becomes 22.7%. You will have bought roughly three quarters of the thing you decided to buy and you will never notice, because nothing will error and the pie chart will look right.

Rebalancing. Once a year, on a date you pick and never move. Compare each line against its target and act only when one has drifted a quarter of its own weight away or further, so a 30% line at 22% or 38%. That policy held mean drift near 1 percentage point at roughly three trades a year and zero realised tax across our test.

Do every trade inside the Roth or the traditional account. All four foreign funds and the whole trend holding already live there, so selling them realises nothing. The one move you cannot make internally is selling US stocks to buy international, because your US holdings are mostly in the taxable account. It needs about two points of the portfolio a year, and your contributions cover it several times over.

So: point new money at whatever is furthest below target, fill the dollar-limited accounts first, and do not sell in the taxable account. Never selling there is worth about 14 basis points a year on its own, which is larger than the entire ordering decision above. Leave a point or two of VTI in the Roth rather than taxable, so a rebalance can never force a taxable sale. That costs a fraction of a basis point.

The bad decade

What it will feel like when it is losing

Not a crash. A long grey decade where the thing you own trails the thing everyone else owns, monthly, for years, with no obvious moment to act. Here is what these three engines have already done, inside the history this portfolio is built from.

None of those is a tail scenario. All three are in the record. The whole portfolio's worst peak-to-trough fall was −50.3% against −64.6% for a control that borrows as much, and its longest stretch under water was 42 months. Its worst decade against that control was −1.15 points a year, for seventeen years, starting in 2009.

And now the fact that reframes everything above.

This portfolio drifts so far from the thing you would compare it against that thirty years of holding it cannot establish whether it worked. Drift is how far your returns wander from that comparison in a typical year, and here it is 6.0%. The arithmetic is exact and there is no way around it: the time you need to demonstrate an edge grows with the square of the drift divided by the edge. Against a plain cheap index fund holding the same mix, thirty years of data would let us spot an edge of 93 basis points a year. The edge we think is there is 92 basis points, which puts the wait at 31 years.

You will not find out. Not in thirty years, not from your own account. That is not a reason to hold nothing, and it is a reason to stop checking. If you set a performance review on the trend fund you will delete it for doing its job: at a 30% weight it trails in 43.8% of simulated ten-year histories even when the extra return it is supposed to earn is really there.

The one discipline that matters: whatever you compare this against must have the same borrowing in it. Comparing a portfolio carrying 132% of exposure against a plain index fund credits the extra exposure to the strategy on the way up and blames the strategy for it on the way down.

What would change it

What would change this

Your own numbers, first and hardest. How much of your tax-deferred third is a rollover IRA rather than a workplace plan, which moves the placement result from −2 to +7 basis points a year. The deepest fall you could sit through without selling, which is the single most useful thing we do not know about you and is not a research question. Your tax bracket.

Then the split between stocks and everything else, which this portfolio sets at all stocks by omission. Moving from 60/40 to 90/10 was worth +1.27% a year against 4.85% of drift, larger than every lean in this portfolio put together. This portfolio holds no bonds. Decide that on purpose.

Then four measurements with dates on them.

• How much trend RSST actually delivers, currently +0.681, plausibly +0.41 to +0.96 on 31 months of filings, which is roughly one kind of market. We look again at 48 months, around September 2027.
• The borrowing cost inside RSST, which no issuer discloses because it is buried in the price of a futures contract rather than printed on a statement. It decides whether the trend part helps you at all.
• RSST's next December distribution, which settles how much tax that fund is really storing up.
• Whether your broker charges $100 a trade on it.

One more, for later. RSIT launched on 6 May 2026 and does for international stocks what RSST does for US ones, which is exactly where this portfolio is thin. It is three months old with no holdings filings, so there is nothing to measure yet. Ask again in two years.

Nothing here is regulated advice. Every number is a function of stated inputs, and a different investor should restate them. US federal taxes only; state tax is excluded and makes every gap in the placement table smaller without reordering it.