We Simulated 19,591 Market Timers Across 150 Years. Nearly Three in Four Lost.
We built a population of behavioral market-timers, panic sellers, drawdown dodgers, coin flippers, and momentum chasers, and ran every one of them across every 30-year window since 1871. Only 28.4% beat an investor who just bought every month and never sold.
We wanted to know what actually happens to people who try to time the market, so we simulated a whole population of them. Not one clever strategy, but 19,591 individual market-timers, each following a simple behavioral rule, each turned loose on real US market history. Then we counted how many beat the most boring plan imaginable: invest every month and never sell.
The answer: 28.4% of them. Nearly three in four lost. This is the experiment, the numbers, and the honest caveats.
If you want the strategy-by-strategy version, our companion piece Can You Time the Market? tests named tactics one at a time (missing the best months, buying the dip, a hindsight-perfect crystal ball). This article does something different. It builds a crowd of imperfect humans, runs them all, and reports the shape of the outcomes.
What we simulated
Every timer starts the same way: $10,000 invested, plus $500 added every month. While a timer is “in”, that money rides the market and the paycheck keeps buying. While a timer is “out”, the whole position sits in cash earning nothing and the paycheck piles up as cash. Exiting moves everything to cash; re-entering deploys all of it back in. The only thing that varies is the rule a timer uses to decide when to jump out and when to jump back in.
We modeled four archetypes, each a caricature of a real investor instinct:
- The panic seller bails after any single month that drops a lot (we tested thresholds of 5%, 8%, and 10%), then waits for a short streak of positive months (2, 3, or 4) before buying back in. That is nine variations of the same fear response.
- The drawdown dodger watches the running high and sells the moment the market sits 15% below it, then refuses to buy back until the market fully reclaims that old high. The “I will get back in once it recovers” investor.
- The coin flipper has no thesis at all. Each month there is a small random chance (1% or 2%) of flipping between in and cash. This is our control group: pure noise, no skill, no fear.
- The momentum chaser sells after three straight down months and buys after three straight up months. Trend-following on a hair trigger.
Nine panic variants, one dodger, two coin flippers, one momentum chaser is 13 timer types. We ran each one across every rolling 360-month (30-year) window in the data, and there are 1,507 of those since 1871. That is 13 times 1,507, or 19,591 timer-window runs in total. Every run is compared against a steady investor facing the exact same cash flows in the exact same window, who never once sells.
The results
Here is how each archetype did against steady monthly investing. “Beat steady” is the share of that archetype’s runs that finished ahead. “Median shortfall” is how far the typical run finished behind (a negative shortfall would mean the typical run won). “Worst run” is the single unluckiest outcome in that group.
| Archetype | Runs | Beat steady | Median shortfall | Worst run |
|---|---|---|---|---|
| Panic seller | 13,563 | 37.6% | 6.1% | -60.2% |
| Momentum chaser | 1,507 | 13.1% | 14.1% | -54.1% |
| Coin flipper | 3,014 | 8.5% | 43.3% | -86.7% |
| Drawdown dodger | 1,507 | 1.3% | 46.9% | -63.0% |
| All timers pooled | 19,591 | 28.4% | 11.8% | -86.7% |
All figures computed from the Shiller monthly real total-return series across every 30-year window since 1871, using the reproducible script described in Methods. Abstract tax-free portfolios, so the comparison isolates the timing rule.
Pooled across all 19,591 runs, only 28.4% beat the boring line and the median timer finished 11.8% behind. That is the one-sentence version: a crowd of behavioral timers loses to plain monthly investing about 72% of the time.
But the average hides the most interesting result, which is how differently the archetypes failed.
The dodger got destroyed by its own patience
The drawdown dodger sounds like the responsible one. Sell when things get bad, wait for the recovery, get back in. It beat steady investing in just 1.3% of windows and trailed by a median of 46.9%, cutting the typical final balance nearly in half.
The reason is the re-entry rule, not the exit. Selling at a 15% drawdown is not what wrecked it. Refusing to return until the market reclaimed its old high is. Big recoveries do most of their work early, well before the old peak is back, so a rule that waits for full recovery sits out exactly the stretch that mattered. The dodger kept discipline all the way down and then stayed disciplined through the whole way up.
Panic selling was the least bad, because panic sellers came back fast
This surprised us. The panic seller, the archetype that sounds the most reckless, was the least damaging of the four. It beat steady investing 37.6% of the time and trailed by a median of only 6.1%.
The difference is the re-entry rule again. Our panic sellers bought back in after just two, three, or four positive months, so they were rarely out for long. A fast return kept them close to the boring line even though they sold at bad moments. The lesson the data keeps repeating is that time out of the market, not the act of selling, is what does the damage. The dodger stayed out for years; the panic seller stayed out for months.
Even pure noise made the point. The coin flipper had no strategy whatsoever and still beat steady investing only 8.5% of the time, trailing by a median of 43.3%, because randomly sitting in cash means missing compounding you never get back.
The tails, and why the winners do not rescue the story
Timing did produce some huge winners. The single best run in the whole cohort, a coin flipper that happened to be out during the worst months of its window, finished 252.8% ahead of steady investing. The best panic seller finished 229.0% ahead.
Those exist. They are also the point, not the exception to it. When a strategy’s rare wins are enormous and its typical result is a loss, that is the signature of luck, not skill, the same shape a lottery ticket has. The worst run in the cohort finished 86.7% behind steady investing. You do not get to pick which tail you land in ahead of time, and nearly three in four landed on the losing side.
Methods
We want this fully reproducible, so here is exactly what we did.
- Data. The Shiller monthly US stock series, in real (inflation-adjusted) total-return terms, including dividends, starting February 1871. Bundled as static reference data, not a live feed.
- Windows. Every rolling 360-month window, 1,507 of them. These windows overlap and share history, so they are not statistically independent samples. Treat the counts as a description of the historical record, not as 1,507 independent coin flips.
- Money model. $10,000 starting balance plus $500 per month. Invested money compounds at the monthly market return; cash earns 0% real. Exiting moves the entire position to cash, re-entering deploys all cash back in.
- Randomness and seed. Only the coin flipper uses randomness. It draws from an inline mulberry32 pseudo-random generator re-seeded per run from a fixed base seed (
0x5eed1871), so the entire study prints identical numbers every time it runs. The other archetypes are fully deterministic. - No taxes, no fees. These are abstract portfolios. Removing taxes and trading costs is deliberate: it isolates the timing rule so the result cannot be blamed on friction. Real timing would fare worse, not better, once taxes on every realized sale are added.
That last point is also the key difference from our timing game. The game charges a 15% tax on gains you realize each time you cash out, because selling in a taxable account is a real event. The steady ghost that never sells never pays it. So the game is stricter than this study, and a real-world timer would sit somewhere below these already-losing numbers. The full script lives in the repository as scripts/timer-cohort-study.mjs; the numbers above are pasted from its output verbatim.
Think you would have timed it better?
Play through 30 hidden years of real market history, one month at a time, with the dates concealed. Cash out when it feels scary, buy back when it feels safe, and see where you land against the steady line and a set of ghost strategies at the end.
The honest takeaway
None of this proves the market only goes up, and none of it is advice. It is a description of what a large crowd of rule-following timers would have done against 150 years of real returns. The pattern is consistent across every archetype: the cost was almost always time spent out of the market, and the strategies that stayed out longest lost the most. A few timers got rich on luck, and you cannot know in advance whether you are one of them.
The boring line, invest what you can every month and leave it alone, won 71.6% of these matchups without trying to be clever once. That is the whole finding.
Frequently asked
Do market timers usually beat buy-and-hold investors?
In our simulation, no. We ran 19,591 behavioral market-timers across every 30-year window of US market history since 1871, and only 28.4% of them beat an investor who simply invested every month and never sold. Nearly three in four finished behind, and the median timer trailed steady investing by 11.8% in final wealth.
Which market-timing strategy did the worst?
The drawdown dodger, a timer that sells when the market falls 15% off its high and waits to buy back until the market fully recovers that high. It beat steady investing in only 1.3% of windows and trailed by a median of 46.9%, because waiting for a full recovery kept it out of the market through the entire rebound, which is where a large share of the gains happen.
Is panic selling as bad as it sounds?
It was the least bad archetype we tested, but it still lost most of the time. Panic sellers who bought back in after just two or three positive months beat steady investing in 37.6% of runs and trailed by a median of only 6.1%. Getting back in quickly limited the damage. The timers that stayed out waiting for a clear all-clear did far worse.