What is a market timing game?

A game where you choose when to be in or out of real S&P 500 history and are scored against an opponent that buys once and never sells. Across the whole ledger that opponent, a couch, wins about 62% of games.

BEAT THE COUCH · THE ARCHIVE

Where Timing Worked, and Where It Died

Q: Does market timing work in a crash?

In this dataset players beat buy-and-hold 72.4% of the time in windows starting in 1929, but only 51.4% in windows starting at the 2007 peak. Only 7.0% of 2008 players passed a statistical skill audit.

Q: What was the worst time to try to time the market?

Windows starting in January 1995: players beat a buy-and-hold couch only 3.3% of the time over the following two years. Windows starting in 2012 are second worst at 4.8%.

TEN WINDOWS · 120,800 GAMES ON THE LEDGER · REAL S&P 500 DAILY DATA

Ten famous two-year windows from the last century of the S&P 500. For each one: what a couch made by buying once and never selling, and how often real players beat it. Sorted by how well humans did. The pattern is not subtle.

The window	Couch	Worst drop	Humans beat it	With skill
The Bubble, 1999	+7.0%	-17%	75.3%	15.7%
The Great Crash, 1929	-53.9%	-58%	72.4%	16.0%
The Top, 2000	-23.5%	-36%	70.1%	9.8%
The Long Grind, 1973	-34.0%	-44%	61.5%	7.8%
The Relapse, 1937	-24.0%	-52%	54.6%	13.7%
The Crisis, 2008	-28.4%	-55%	51.4%	7.0%
Black Monday, 1987	+21.9%	-33%	39.5%	14.6%
The Bottom, 1932	+126.0%	-39%	34.8%	14.2%
The Melt-Up, 2012	+49.1%	-10%	4.8%	4.3%
The Worst Year To Be Clever, 1995	+71.3%	-7%	3.3%	3.0%

Window figures come from the run export of August 28, 2026 and cover every game whose window started in that year. Across these ten windows players beat the couch 48.6% of the time; these are the famous years, which is the point. Across the whole ledger the couch wins 62% of 120,800 games, with 12% of all games ending in a win that passes the skill audit.

The pattern

Every window at the top of that table is a disaster: 1929, 1999 into the top, the 2000 crash, the 1973 grind. Every window at the bottom is a calm market where nothing went wrong: 1995, 2012, 1932's recovery, 1987 after the crash was over.

Humans beat buy-and-hold when the market falls, and they get taken apart when it rises. Across the whole ledger, players won 65% of games where the market fell over the window and 26% where it rose. That is not skill appearing and disappearing. That is the same behaviour, sitting in cash, being right by accident in one regime and expensive in the other.

The catch is the one thing no page can fix: from inside the window, 1995 and 1999 look identical. You find out which one you were in afterwards.

Take a window you have not seenfree, no signup. Two secret years, one button, and a couch.

Frequently asked

Does market timing work in a crash?

Better than at any other time, and less than you would think. In windows starting in 1929 players beat buy-and-hold 72.4% of the time, and in 2007 to 2008 only 51.4%, roughly a coin flip, in the most anticipated crash of the century.

What was the worst time to try to time the market?

January 1995. Players beat the couch 3.3% of the time over the following two years. The 2012 melt-up is second worst at 4.8%. Neither window contains a crash.

What is the couch?

The opponent in a free market timing game. It buys the index on day one, never sells, collects every dividend and makes no decisions. Across 120,800 games on the ledger it wins 62% of the time.

BEAT THE COUCH · the game · the archive · the evidence · the data · Privacy · Terms

Not financial advice. The couch is not licensed for anything.