Still working. The data server sleeps when idle, so the first load after a quiet spell can take up to about 40 seconds. This page will fill in on its own.
Still working. The data server sleeps when idle, so the first load after a quiet spell can take up to about 40 seconds. This page will fill in on its own.
Skill vs luck
A portfolio that outran the index tells you less than it feels like. Two things make almost any track record look good in hindsight, and neither is skill.
If you picked individual stocks, you’re now looking at the ones that survived. The companies that looked just as reasonable and went to zero aren’t in your portfolio, and they’re not in the backtest. A regression on the survivors always shows some outperformance — that’s the selection, not your judgment.
The same allocation started in different years can return wildly different amounts. Begin in early 2009 and almost anything looks brilliant; begin in 2000 and the same mix struggled for a decade. If your headline number leans on one lucky entry point, it won’t repeat.
Regress your monthly returns on the known return drivers — market, size, value, profitability, investment, momentum. This is the standard academic decomposition. Two things to read:
For a basket of hand-picked stocks, a large positive alpha with a weak t-stat is the signature of hindsight. For a passive index mix, a small alpha is usually a rebalancing bonus or a data artifact — not an edge you can count on going forward.
Run the regression on your portfolio
Portfolio Lab regresses your holdings on the Fama–French five factors plus momentum, reports the alpha and its t-stat, and shows a start-date sensitivity heatmap so you can see how much your entry year flattered the record.
Analyze a portfolio →Not necessarily. A 10-year outperformance can easily come from a single style tilt (say, toward large-cap growth) that happened to be in favour, plus the luck of when you started. The test is whether the excess return survives once you account for those known exposures.
You regress your portfolio's monthly returns on a set of well-documented return drivers — the overall market, company size, value, profitability, investment, and momentum. Whatever return is left unexplained is 'alpha'. If its t-statistic is below about 2, it's statistically indistinguishable from zero — i.e. noise.
Run the same allocation from every possible starting year and the annualized return often spans 8+ percentage points. A single headline CAGR hides that. If your good number depends heavily on having started in, say, 2009, that's timing luck, not a repeatable edge.
Educational, not advice.