For a few years at the end of the 1990s, buying almost any company with ".com" in its name felt like genius. The Nasdaq Composite climbed from under 1,000 in early 1995 to a peak of 5,048.62 on March 10, 2000 — more than a fivefold rise. Fortunes were made on companies that had never earned a dollar of profit. And then the tide went out — not in a day, but over two and a half years.
The build-up: when "up" felt like skill
The mania had a long runway. Netscape's August 1995 IPO showed that a company with no profits could double on day one. In December 1996, Fed chairman Alan Greenspan wondered aloud about "irrational exuberance" — with the Nasdaq around 1,300. The market's answer was to nearly quadruple after the warning. In 1999 alone the Nasdaq rose 85.6%, the biggest single-year gain a major U.S. index had ever posted. Profitless companies were valued on "eyeballs" and "mindshare," and because everything went up, every portfolio looked brilliant. That's the trap this article is really about: a rising market makes picking feel skillful, when most of the gain is simply the tide — the part of a return we'd call beta.
Here's the whole arc in one chart, using the broad market (the S&P 500, which "only" fell by half):
Two and a half years of down
Unlike 1987 — one catastrophic Monday — the dot-com crash was a slow bleed punctuated by false dawns. The worst single week came early: in the five sessions ending April 14, 2000, the Nasdaq lost about a quarter of its value. But the decline kept unfolding for years, through the 2001 recession and one failed rally after another:
| Date | What happened | Level |
|---|---|---|
| Mar 10, 2000 | Nasdaq Composite peaks | 5,048.62 |
| Mar 24, 2000 | S&P 500 closing peak | 1,527.46 |
| Apr 14, 2000 | Nasdaq's worst week ever: roughly −25% in five sessions | — |
| Nov 2000 | Pets.com liquidates, about nine months after its IPO | — |
| Mar 2001 | U.S. recession begins (it ends in November 2001) | — |
| Oct 9, 2002 | The closing lows: Nasdaq −78%, S&P 500 −49% from peak | 1,114.11 / 776.76 |
| May 30, 2007 | S&P 500 finally closes above its 2000 peak — 7 years later | 1,530.23 |
| Apr 23, 2015 | Nasdaq finally closes above its 2000 peak — 15 years later | 5,056.06 |
Look again at the middle of the chart above. From the post-9/11 low in September 2001, the S&P 500 rallied about +21% over six months — a move big enough to convince many investors the bear market was over. It wasn't. The market rolled over and fell to new lows for another seven months. That's the cruelest feature of a valuation unwind: the tide comes back in just long enough to pull people back to the beach.
Not all crashes are the same species
Put the dot-com crash next to other famous declines and a pattern emerges. Structural crashes — a crowded trade or a piece of market plumbing breaking, like 1987's portfolio insurance or 2024's yen carry unwind — have tended to heal fast, because nothing was fundamentally wrong with the underlying businesses. Valuation crashes — where prices had simply lost contact with earnings — take years, because there's no snap-back mechanism: the market has to wait for reality to catch up with the old prices.
Why it's a lesson and not just a story. During the boom, a rising market made nearly everyone look brilliant. The crash exposed how little genuine, market-beating judgment had actually been at work. Survival and quality diverged sharply: some richly-valued names never came back, while a handful of durable businesses fell just as hard yet went on to compound for decades. And the people who fared best were rarely the ones who called the exact top — they were the ones who had been asking, all along, a quieter question: how much of this return is the market, and how much is me?
How we try to carry the lesson. That question is exactly why we publish a market-relative number, not just an absolute one, next to every result — and why we log our reasoning and our regime read each day, in the good stretches as much as the bad ones. A record kept only when the tide is rising isn't a record; it's a highlight reel. The dot-com era is a reminder that "up" and "skilled" are not the same word, and that the only way to tell them apart is to keep honest score across a full cycle — including the years when the tide runs out.
1. A bull market hides the difference between skill and tide — the Nasdaq's 85.6% gain in 1999 made everyone a genius until the −78% that followed sorted them out.
2. Valuation unwinds are slow and deceptive: the 2000–2002 decline took 2.5 years and contained a +21% rally that failed. Structural crashes (1987, 2024) round-tripped in months or days; this one took the S&P 500 seven years to fully retrace.
3. The durable defense wasn't calling the top — it was measuring returns against the market the whole way. A pattern from history, not a prediction.
This is informational and entertainment content about market history, not investment advice. The S&P 500 chart is built from Tiingo data (SPY adjusted close); index levels and dates are widely documented historical records. Past results do not guarantee future outcomes.