Which Internet Investing Advice Actually Survives Scrutiny? A Quant’s Three-Question Test
🧮 Which Internet Investing Advice Actually Survives Scrutiny? A Quant’s Three-Question Test
Welcome back, everyone! 📊 I’m your Quant Analyst, filtering out market noise using data, statistical modeling, and systematic insights. 👩💻✨
The internet is full of investing techniques, and almost all of the discussion about them is people asserting that they work. I want to do something narrower and more useful: apply one consistent test, show the arithmetic, and let some famous advice fail. To cut straight to the chase: the techniques that survive are the boring ones that need no forecast — and even for those, the reason usually quoted on the internet is weaker than the reason that actually holds up.
▲ Almost all investing advice reaches you with the author’s track record hidden. The arithmetic is the part you can check yourself
📌 First, the actual numbers
Three primary sources carry this post. SPIVA (S&P Dow Jones Indices, data as of Dec 31, 2025): 89.93% of US large-cap active funds underperformed the S&P 500 over 15 years, and 78.78% over one year. Morningstar’s Mind the Gap 2025 (published Aug 13, 2025; 10 years to Dec 31, 2024): the average dollar earned 7.0% a year against an 8.2% fund total return — a 1.2 percentage point gap. And the Financial Analysts Journal (Fulkerson, Jordan, Riley & Yan, Vol 82 Issue 3, May 12, 2026) re-ran that same sample and put the timing cost at 0.10% a year instead — about twelve times smaller.
Sources: S&P DJI SPIVA; Morningstar Mind the Gap 2025; Financial Analysts Journal, 2026.
That third citation is the interesting one, and I will come back to it in section 3 — because it forced me to stop using a statistic I had repeated myself.
1. The three questions I ask before trusting a technique
I don’t grade advice by whether it sounds wise. I ask three things, in this order:
- Is there a mechanism, or only a backtest? A mechanism is a reason the technique must work, expressible as arithmetic. A backtest is a claim that it did work, on one sample, chosen after the fact. Mechanisms survive regime changes; backtests frequently do not.
- Does it require me to predict something? Techniques whose payoff depends on a forecast inherit the accuracy of that forecast. Techniques that work regardless of what the market does are structurally more reliable, even when their payoff is smaller.
- Has anyone serious tried to break the evidence? A statistic that has never been attacked is not the same as a statistic that has survived attack. This is the question almost nobody asks, and it is the one that changed my mind while writing this post.
2. Four techniques that pass
a) Minimise costs. This is the only input in investing whose sign is guaranteed. You cannot know next year’s return; you know the fee exactly, in advance, and it compounds against you every year. Take $10,000 at a 7% gross return over 30 years:
| Annual fee | Value after 30 years | Lost to fees |
|---|---|---|
| 0.03% | $75,485 | — |
| 0.50% | $66,144 | $9,341 |
| 1.00% | $57,435 | $18,050 |
| 1.50% | $49,840 | $25,645 |
A 1.47-point fee difference removes 34% of the final balance. No forecast is involved anywhere in that calculation, which is exactly why it is the most reliable advice on the internet — and the least exciting.
b) Default to the index unless you have a specific reason not to. The evidence here is unusually clean because S&P DJI publishes it against their own benchmarks:
Active Funds That Failed to Beat Their Benchmark
*SPIVA U.S. Scorecard, all domestic large-cap funds vs the S&P 500, data as of Dec 31 2025. Note the 3-year bar: the failure rate is not monotonic in horizon.
And here is the part the internet version leaves out. Look at the 3-year bar — 66.84%, dramatically lower than the 5-year figure of 88.96%. If “active management fails” were a law of nature, that number would rise smoothly with horizon. It doesn’t. What it actually shows is that the failure rate is regime-dependent: in a market led by a handful of very large stocks, a diversified active manager is nearly guaranteed to lag, and the 15-year figure is dominated by exactly that kind of decade. The honest version of this advice is “indexing is the right default and the burden of proof is on deviating,” not “it is mathematically impossible to beat the market.”
c) Trade less. The strongest form of this evidence is not the headline number, it is the ordering. Morningstar sorted funds into quintiles by how much investors moved money in and out, and found the gap widening across the range:
The More Investors Traded, the Less They Kept
*Morningstar, Mind the Gap 2025 — annual investor return gap by cash-flow-volatility quintile, 10 years to Dec 31 2024. The absolute level of these bars is disputed (section 3); the ordering is the durable part.
d) Size positions so that being wrong is survivable. This one is pure arithmetic and it is the least discussed of the four. Losses and gains are not symmetric, because a loss shrinks the base the recovery has to work on:
| Drawdown | Gain needed just to get back to even |
|---|---|
| −20% | +25.0% |
| −30% | +42.9% |
| −50% | +100.0% |
| −70% | +233.3% |
| −90% | +900.0% |
Nobody needs to forecast anything to use this. It says only that the cost of a bad outcome grows faster than its size, so the position that matters is the one you would still be holding after it halves.
3. The statistic I had to stop using
Here is where question three earns its place. “The average investor underperforms their own funds by about 1.2 points a year, roughly 15% of their returns” is probably the single most-quoted statistic in personal finance. It comes from Morningstar’s Mind the Gap, it is a real study by serious people, and I have used it myself.
In May 2026 the Financial Analysts Journal published a paper by Fulkerson, Jordan, Riley and Yan with a title that leaves little room for interpretation: “Bad Timing Does Not Cost Investors 15% of Their Funds’ Returns.” Running the same sample, they put the cost of poor timing at 0.10% a year — about a twelfth of the headline figure.
What I now think, and why the chart above still stands
The disputed quantity is the level — how many points bad timing costs in absolute terms. That number should not be repeated as fact. What the critique does not overturn is the relative ordering across quintiles: funds whose investors traded most had wider gaps than funds whose investors traded least, and a comparison between quintiles does not depend on the contested baseline. So “trade less” survives; “trading costs you 15% of your returns” does not. That is a smaller claim, and it is the one I can defend.
I’m including this partly because it is a good example of the third question working. If I had only asked “is there a mechanism?” and “is there evidence?”, this statistic passes both. It took asking whether anyone had attacked it.
4. What fails the test
Leveraged ETFs as a long-term hold. This is the most confidently repeated bad advice on the internet, and it fails on mechanism, which makes it a rare reliable negative. A daily-reset k-times fund compounds at exp(k·μ − k²σ²/2), so the drag term scales with the square of leverage. Assume the index compounds at 8% and the fund charges 0.95%:
| Index volatility | What “3x” sounds like | Actual long-run CAGR |
|---|---|---|
| 15% | +24% | +16.6% |
| 25% | +24% | +3.5% |
| 35% | +24% | −13.6% |
| 45% | +24% | −32.0% |
The index rises in every row. The fund still loses money in two of them. That is not a warning about risk tolerance — it is a statement about what the instrument is built to do, which is track one day at a time. I’ve worked through this in more depth in whether leverage is inherently dangerous and in the SOXL fair-value model.
The whole family of “pattern plus confidence, no base rate.” Chart formations, indicator crossovers, and “this setup has never failed” claims share one defect: they tell you what happened when the pattern worked and are silent about how often it appeared and did nothing. A pattern without a denominator is not a strategy. The fix is cheap — count every occurrence, not just the memorable ones — and almost nobody does it, which is itself informative.
“Buy the dip” with no definition of a dip. As stated it is unfalsifiable: any decline is a dip until it isn’t. It becomes a real technique the moment you attach a threshold, a sizing rule and an invalidation level, and at that point it is just a plan — which is the point.
Quick FAQ
Q. Isn’t “keep costs low, index, trade less” the most generic advice imaginable?
Yes, and that is the finding, not a failure of the analysis. When you apply a consistent test, the surviving set is small and boring. The value I can add isn’t a more exciting list — it’s showing you the arithmetic so you can check it, and telling you which of the usual supporting statistics does not hold up.
Q. So dollar-cost averaging beats investing a lump sum?
Not on expected return — a market that rises more often than it falls means waiting usually costs you. DCA passes my test for a different reason: it reduces the variance of your entry point and makes a bad first year survivable. Judge it as a risk-control tool, and it is defensible; judge it as a return-enhancing trick, and it isn’t.
Q. Does any of this mean active management is pointless?
No, and the 3-year SPIVA bar is why I won’t say that. It means the base rate is against it, so the burden of proof sits with the deviation. “I have a specific, stated reason and I’ve defined what would prove me wrong” is a legitimate answer. “I think I can pick better” is not.
💡 Quant Strategy & Takeaways
Trust the techniques whose payoff is arithmetic rather than forecast — low costs, indexing as a default, trading less, and sizing for survival. Then check whether the statistic being used to sell it has ever been attacked; the most famous one in personal finance was just cut by a factor of twelve in a peer-reviewed journal.
When market volatility spikes, remove emotion and focus strictly on the numbers! 🤖
Which piece of investing advice do you follow that you’ve never actually checked the arithmetic on? Let me know in the comments! 📈✨
Disclaimer: This article evaluates publicly discussed investing methods for educational purposes and is not financial or investment advice. All projections shown are arithmetic illustrations under stated assumptions, not forecasts of any actual return. Always do your own research or consult a licensed financial advisor before investing.
Comments
Post a Comment