Portfolio stress test — the sleeve that switched sides
There is no such thing as a defensive asset. Only a defensive year.
The sleeve that saved you in one crisis led the losses in the next. Stress your own mix, then
see why the usual comparison fails.
+28%
Mar 2000 to Oct 2002, while US stocks fell 49%.
−68%
Oct 2007 to Mar 2009. Worse than the S&P 500’s 57%.
How far would your portfolio have fallen?
Build your portfolio, then replay it through the crashes that actually happened. This tool cannot tell you what the next crash holds — nobody can. It shows you how a mix like yours fell in 2008, in the COVID crash, in the dot-com bust, how deep the hole went and how long the market took to climb back out. The question it leaves you with is the only one that matters in a crash: could you have held?
The free version replays three headline crashes — 2008, the COVID crash and the dot-com bust — in full, with the dip-buying readout and a shareable link. Pro replays all nine crashes back to 1929, compares portfolios side by side, adds a custom-crash builder, and unlocks image and CSV export.
The shareable link is free, for everyone, always. Pro is a separate one-time purchase — if you already subscribe to DCA Simulator Pro, this is not included in it.
Educational content only — not financial advice. Every figure here is a rounded, peak-to-trough index proxy from an event that actually happened, in US dollars, with no fees, tax or slippage. The portfolio drawdown assumes the sleeves bottomed together, which is close for an equity-heavy book and errs toward overstating the loss; recovery months are how long the US market took, not your specific mix. This is history replayed, not a guess about what comes next.
There is no such thing as a defensive asset. Only a defensive year.
Portfolios are usually stress-tested against an idea — a bad year, a standard deviation, a forecast. The trouble is that real crises do not resemble each other, and the asset that saved you in one is frequently the asset that ruined you in the next.
+28%
Mar 2000 to Oct 2002. US stocks fell 49% and REITs rose. This is the decade that taught a generation that property was the diversifier.
−68%
Oct 2007 to Mar 2009. The worst-hit sleeve in the whole crisis — worse than the S&P 500’s 57%. Nothing about the asset changed. The crisis did.
That pair is not a curiosity, it is the pattern. Gold, the permanent hedge, fell 45% in the 1980–82 recession and did nothing at all — 0% — through the 2022 inflation shock, the precise event it is supposed to be insurance against. Bitcoin fell 50% in the COVID crash while the S&P 500 fell 34%: it did not hedge the crash, it amplified it.
The most expensive assumption in the list is the quiet one. In 2022 the aggregate bond index fell 13% and long treasuries fell 31%, in the same year equities fell. The half of the portfolio that exists to cushion the other half went down with it. And in 2008, international stocks fell 57% — exactly what US stocks did. Being diversified across equities turned out not to be diversification at all.
So this tool does not model a crash. It replays nine of them, from 1929 to 2022, using peak-to-trough figures from what actually happened, and shows what your allocation would have done in each. Where an asset did not exist or did not trade freely — bitcoin in 1929, gold under the gold standard — the sleeve is excluded from that scenario rather than quietly scored as a flat zero, because pretending an asset held its value when it simply was not there is how a stress test flatters you.
Nothing here is hidden. Check every line.
A classic 60/40 — sixty per cent US stocks, forty per cent aggregate bonds — holding $100,000, replayed through four real crashes. Peak to trough, nominal, in the dollars of the day.
| $100,000 in a 60/40 | At the trough | Drawdown | Gain needed to recover |
|---|---|---|---|
| 2000–02 Dot-com bust | $80,600 | −19.4% | +24.1% |
| 2020 COVID crash | $79,600 | −20.4% | +25.6% |
| 2022 Inflation shock | $79,800 | −20.2% | +25.3% |
| 2007–09 Financial crisis | $67,800 | −32.2% | +47.5% |
Read the 2022 row against the two above it
The inflation shock cost this portfolio almost exactly what the COVID crash cost it, and slightly more than the dot-com bust, which is not what the headlines said at the time. The mechanism is the difference: in 2000 and in 2020 the bond sleeve held or rose and offset part of the equity fall. In 2022 bonds fell 13% alongside stocks, so nothing offset anything. Same portfolio, same weights, a completely different job done by the same forty per cent.
The last column is the one people underestimate, and it is pure arithmetic rather than commentary. A fall and the recovery from it are not the same size. Losing 32.2% requires a 47.5% gain to get back to level, because the gain is calculated on the smaller number. 1 / (1 - 0.322) - 1 = 0.4749. That asymmetry is why avoiding the deepest drawdowns matters more than capturing the best years.
Change the mix and the ordering changes with it. A portfolio holding 20% bitcoin alongside 55% stocks and 25% bonds lost 30.0% in 2022, where the plain 60/40 lost 20.2% — and needed 42.9% to recover rather than 25.3%. The tool runs this for whatever weights you actually hold, which is the only version of the question that is about you.
What the stress test actually replays.
Four inputs, no account, no email. Your allocation is sent to this site to be calculated and the answer comes straight back — nothing you enter is written to a database, kept after the response, or passed to anyone else.
Say what you hold
Eight sleeves — US and international stocks, aggregate bonds, long treasuries, gold, bitcoin, property and cash. The weights do not need to add up to a hundred; whatever you enter is normalised for you.
Pick the crashes
Three are free and they are the ones most people lived through: 2008, the COVID crash and the dot-com bust. Pro opens the other six and runs back to 1929, which is where the assumptions get tested hardest.
Read the damage
What the portfolio was worth at the trough, how far it fell, and the gain required to get back to level — which is always larger than the fall, and is the figure people consistently underestimate.
Add a monthly amount
Optional. If you were still contributing through the fall, the tool shows what steady buying did to your average cost. It is labelled stylised, and the section below explains exactly what that word is doing.
It replays, it never forecasts — and three readouts need their limits stated
Every number in this tool is a peak-to-trough figure from an event that actually happened: index-proxy total returns over the event window, rounded, in US dollars, nominal. Not one cell is modelled, simulated or projected. That is the design constraint the whole thing is built around, and it is why the tool can say something about 1973 without pretending to know anything about next year.
Where an asset did not exist or did not trade freely, the sleeve is excluded from that scenario rather than scored as a flat zero. Worth seeing what that is worth: a four-way split of stocks, long treasuries, gold and cash, run through 1929, loses 23.7% — gold was fixed at $20.67 an ounce under the gold standard, so it is dropped and the remaining three are reweighted. Score that missing gold as a comfortable 0% instead and the same portfolio prints 17.8%. Nearly six points of false comfort, from one absent asset in one scenario.
Two other figures mean less than they appear to. The recovery time is the S&P 500’s — months to a new nominal high from its trough — not your portfolio’s; it is a benchmark for scale, not a promise about your mix. And the dip-buying readout is stylised: a simple straight-line path down and back, not a real monthly series, because per-month data for nine events does not exist here and inventing it would be worse than admitting the limit. It supports exactly one claim — that buying through the decline lowered your average cost. It is not evidence about lump sums, and this site will not use it as any.
Most complaints about this tool are correct.
A stress test built on nine historical events has exactly the weaknesses you would expect it to have. Here are the four that actually land.
“The next crash will not look like any of these.”
Almost certainly correct — and it is the point
It will not. That is the finding, not the flaw. The nine events in here do not resemble each other, which is precisely why any single one of them makes a bad template. Gold saved you in 1973 and cost you 45% in 1981; bonds cushioned 2000 and made 2022 worse.
So this tool does not claim history repeats. It claims a replay is checkable and a forecast is not. What you get is a range of ways a portfolio has actually been hurt, which is a better preparation than a single projected bad year — and considerably better than the assumption that whatever protected you last time will do it again.
“Nobody holds a fixed allocation through a crash.”
Correct, and the tool cannot answer it
True. This applies one peak-to-trough figure per asset to the weights you entered, held frozen. It does not rebalance into the fall and it does not model anyone selling at the bottom. Real portfolios do both.
It cuts both ways and we will not pretend to know the net. Someone rebalancing into the decline would probably have come out ahead of this figure; someone who capitulated at the trough did far worse than it. The engine has no month-by-month paths, so it cannot price either behaviour, and inventing one to make the answer feel complete would be the wrong trade.
“One number per asset class is far too crude.”
Correct — it is a sketch, at sketch resolution
Each sleeve is a broad index proxy over the event window. “US stocks” is not your fund, and sector tilts, factor tilts, individual holdings and the specific bonds you own are simply not represented. If your portfolio is concentrated, your real experience diverged from this, possibly by a lot.
What survives that crudeness is the shape of the answer: which sleeves moved together when it mattered, and by roughly how much. That is the question the tool is for. Anyone who needs holding-level precision needs their actual holdings, and no eight-slider tool is going to supply it.
“Cash looks great here. Should I just hold cash?”
No — and this is where our own numbers mislead
Every figure in this tool is nominal, and cash is the sleeve that distorts most under that choice. It shows +25% through the 1980–82 recession and +12% through the 1973–74 shock. Both were periods of severe inflation, so the real return on that cash was negative in each case while the number on the screen was green.
It runs the other way too. In 1929 cash reads as a flat zero, and through the deflation that followed it gained substantially in real terms — the tool understates it just as badly. Read cash’s row as a nominal figure with an inflation story attached that this tool is not telling you.
Built for the calm before, not the panic during.
One question, answered properly, once. If what you actually need is something else, the honest answer is that this will not give it to you.
It fits if
- You hold a mix you have never actually stress-tested and want to see the size of the hole before a real one arrives rather than during it.
- You believe one of your holdings is the defensive part of the portfolio and are willing to check that belief against nine crashes instead of the one you remember.
- You are deciding how much of something volatile to add, and want the drawdown cost of that decision priced before you make it.
- You keep hearing that a 60/40 is broken or that gold always helps, and would rather look at what each actually did in specific years.
It does not fit if
- Markets are falling right now and you are looking for a reason to sell. A table of historical drawdowns will not settle you, and a decision made at a trough is not one this tool can help with.
- You want a view on the next crash. This replays what happened and predicts nothing whatsoever, deliberately and permanently.
- You want to be told which allocation to hold. It measures damage; it does not recommend anything, and any mix that looks good across all nine events is a mix chosen by hindsight.
- Your portfolio is concentrated in individual holdings. Eight broad sleeves will not describe it, and the answer would be confidently wrong.
What actually happened, and what it did to a portfolio.
Answered against what the tool actually does, not against what would be convenient to claim. All returns are peak-to-trough, nominal, in US dollars.
How much did a 60/40 portfolio lose in 2022?
About 20.2% peak to trough — $100,000 down to $79,800 — and the reason is the part worth understanding. US stocks fell 25%, which is unremarkable. The aggregate bond index fell 13% in the same window, and long treasuries fell 31%.
So the sleeve whose entire purpose is to cushion the other one fell alongside it. That is why the 2022 shock cost a 60/40 nearly the same as the COVID crash (20.4%) and slightly more than the dot-com bust (19.4%), despite equities falling far less than in either. A 60/40 is not a hedge against every kind of bad year — it is a hedge against the kind where bonds rally.
Does gold protect against a stock market crash?
Sometimes, spectacularly, and sometimes not at all. In the 1973–74 stagflation gold roughly doubled while US stocks fell 48% — the case everyone quotes. In the 1980–82 recession gold fell 45%, crashing from its January 1980 peak while stocks were also falling.
And in the 2022 inflation shock — the exact scenario gold is marketed as insurance against — it returned zero. Not a loss, not a hedge, nothing. “Gold always helps” is a claim these three events falsify between them, which is the case for replaying more than one crash.
Is bitcoin a hedge against a market crash?
On the evidence available, no — it has amplified crashes rather than cushioning them. In the COVID crash bitcoin fell 50% while US stocks fell 34%: it went down with equities and it went down harder. In the 2022 rate shock it fell 65%.
The honest caveat is the sample. Bitcoin has only lived through two of the nine events here; in every earlier crash there was no investable market, so the tool excludes it from those scenarios entirely rather than scoring it as flat. Two crises is thin evidence for a hedge, and both of them point the same unhelpful way.
How much did portfolios lose in the 2008 financial crisis?
A 60/40 fell 32.2% — $100,000 to $67,800 — and needed a 47.5% gain to get back to level, because a recovery is always larger than the fall that caused it. 1 / (1 - 0.322) - 1 = 0.475
The lesson underneath is what happened to the things meant to spread the risk. International stocks fell 57%, exactly what US stocks fell — correlations went to one, and being diversified across equities was not diversification. Property was worse than either: REITs fell 68%, the deepest loss in the whole crisis, having risen 28% through the dot-com bust seven years earlier.
How long does it take to recover from a market crash?
Anything from months to decades, and the spread is the answer. From the 2020 trough it took about five months; from the 2009 trough, 49 months; from the 1932 trough, 266 months — twenty-two years. Planning around the recent ones is planning around the easy end of the distribution.
Two limits on that figure, both worth holding onto. It tracks the S&P 500 back to a new high, not your portfolio, so it is a benchmark for scale rather than a statement about your mix. And it is measured in nominal terms — a nominal high reached after twenty-two years of price changes is not the same as getting your purchasing power back.
Is my data saved or sent anywhere?
No. There is no account, no email gate and no sign-up. Your allocation is sent to this site to be calculated, and the answer comes straight back — nothing you enter is written to a database, kept after the response, or passed to any third party. The only thing this page stores in your browser is whether you chose dark mode.
The free version replays the dot-com bust, the 2008 crisis and the COVID crash in full, with the dip-buying readout and a shareable link. Nothing in the free answer is clamped or truncated within those three events.
Everything above stays free. Pro goes further.
Pro replays all nine crashes back to 1929, compares portfolios side by side, adds a custom-crash builder, and unlocks image and CSV export.
Paid once. Not a subscription, not a bundle.
The free version replays 2008, the COVID crash and the dot-com bust in full, with the dip-buying readout and a shareable link. If those three already told you what you needed about this mix, you do not need this. Educational content only — not financial advice.