A stock prints a golden cross and the headlines write themselves. Bullish signal. Trend confirmed. Almost none of those headlines mention the 1992 study that made economists take this seriously, or what happened when other researchers tested it on data that came after.
Brock, Lakonishok, and LeBaron tested 26 trading rules against 90 years of Dow Jones data and found every single one beat buy-and-hold. Moving-average rules were among the strongest. Published in the Journal of Finance. Not a forum theory. An academic finding that made technical analysis worth a second look from people paid to be skeptical of it.
Then, in 2014, three researchers took the exact same rules and tested them on 25 years of fresh data the original study never touched. No evidence the rules predicted anything at all. Same rules. Different decades. Two opposite answers, and this report walks through exactly why.
It goes further than any of the source papers do. It builds two original simulations from scratch, thousands of runs each, to answer the one question none of the academic literature prices out directly: what happens to an account that mechanically trades every single golden cross and death cross, for twenty years?
The strategy beat buy-and-hold in 8.8 percent of the runs. Its median result finished more than $15,000 behind doing nothing at all, on a simulated $10,000 start. A documented, formally tested signal, and a losing strategy anyway, once the lag and the cost of acting on it got added back in. The full math, and every source behind it, is below.
A Trading Habits Report
The Golden Cross Report
What the 1992 study that started the debate actually found, what a true out-of-sample test found decades later, and two original simulations testing what trading every golden and death cross costs against holding a position and doing nothing.
- Length 20 pages, with 8 original charts and two from-scratch simulations
- Author TradingHabits.com
- Format PDF, delivered as an instant download right after checkout
- Covers the 1992 study, the 1999 Reality Check, the 2014 out-of-sample test, a separate chart-pattern research strand, one dated case study, and two original simulations
Exactly one stock is named anywhere in this report, in a single dated historical case study. No stock list, no screen, and no recommendation appears anywhere in it.
What's Inside
20 Things This Report Actually Says
- 01The plain arithmetic behind a golden cross and a death cross, no mystery, no magic.Page 4
- 02Why financial media covers every cross the same way, and the side effect that coverage cycle creates.Page 5
- 03Brock, Lakonishok, and LeBaron's 1992 finding: all 26 rules tested beat buy-and-hold over 90 years.Page 6
- 04The caveat the authors attached to their own paper that almost never gets repeated with it.Page 7
- 05The 1999 Reality Check: a formal bootstrap test for data snooping, and what it found on fresh data.Page 8
- 06The 2014 true out-of-sample test: the same rules, 25 years of data the original study never saw.Page 9
- 07Two competing explanations for why a documented edge can stop working, and what the evidence favors.Page 10
- 08A separate 2000 study asking a different question: does a chart shape carry information at all.Page 11
- 09One single stock, two death crosses, three years apart, with two completely different outcomes.Page 12
- 10An original simulation: how many golden and death crosses appear with zero trend behind them.Page 13
- 11A second original simulation: trading every cross for 20 years against a plain buy-and-hold baseline.Page 14
- 12The result: an 8.8 percent win rate and a median shortfall north of $15,000 on a $10,000 start.Page 14
- 13Why the shortfall is mostly about lag, not trading fees, once the numbers get broken down.Page 15
- 14The psychological reason a lagging signal still feels predictive after a trend is already visible.Page 16
- 15Why the identical signal behaves completely differently in a trending market versus a choppy one.Page 17
- 16Eight questions to run before trading any moving-average crossover, not only golden and death crosses.Page 18
- 17Four sources, from the original 1992 paper to the 2014 out-of-sample test, listed with exactly what each backs up.Page 19
- 18A seven-term glossary defining every concept in the report the first time it appears.Page 3
- 19The full stated assumptions behind both original simulations, nothing hidden in the method.Pages 13-15
- 20Why "some measurable information" and "a profitable trading edge" are two different findings, not one.Page 11
HABITS
★
Certificate of Guarantee
60-Day, No-Questions-Asked
If The Golden Cross Report doesn't earn its place on your desk, email us any time within 60 days of purchase for a full refund. No form to fill out. No reason required.
Instant Download
$9One-time payment. No subscription.
Order Now20-PAGE PDF · DELIVERED IMMEDIATELY AFTER CHECKOUT · 60-DAY GUARANTEE
Behind The Report
Every Rule Beat The Market In 1992. Then Researchers Tested It Again.
The Study That Started It
In 1992, Brock, Lakonishok, and LeBaron tested 26 technical trading rules, including moving-average crossovers, against 90 years of Dow Jones data. Every rule beat a passive buy-and-hold benchmark. The moving-average rules were among the strongest performers in the entire study.
The authors attached their own caveat: testing many rules on one historical sample raises the odds some of them look good purely by chance. That caveat turned out to matter more than almost anyone citing the paper realized.
What Fresh Data Found
In 2014, researchers Fang, Jacobsen, and Qin took the exact same rules and tested them against 1987 through 2011, 25 years the original study never saw. No evidence the rules predicted anything.
So this report built its own simulation on top of that finding: 8,000 runs, 20 simulated years each, pricing out what trading every cross actually costs. Exactly one stock is named anywhere in this report, in a single dated illustration.
The Same Rules, Two Different Eras
A simplified rendering of the research discussed on pages 6 and 9 of the report. Background only. The report itself covers the full sourcing, the 1999 Reality Check test in between, and every stated assumption.
Try It: A Whipsaw-Cost Calculator
A simplified version of the cost math behind the report's own Section 10 and 11 simulations, using the report's stated 6.65 percent buy-and-hold growth assumption and a fixed illustrative lag-cost figure per switch. Not a forecast, not tied to any actual broker's fee schedule, and not a recommendation. The full 8,000-run simulation, with its complete assumptions, is in the report.
Background only. Nothing on this page or in this report is a recommendation to buy, sell, or hold any security, or to trade around any golden cross, death cross, or other moving-average signal.
Common Questions
Is the golden cross actually backed by research, or is it trading folklore?
It has a published research base behind it, more than most technical signals get credit for. Brock, Lakonishok, and LeBaron's 1992 paper in the Journal of Finance found moving-average rules beating buy-and-hold across 90 years of Dow Jones data. That does not settle whether the edge holds up today, which is exactly what Sections 4 and 5 test directly.
Does the golden cross still work today?
The strongest available evidence says the original edge did not carry forward cleanly. A 2014 study applying the same 1992 rules to 1987 through 2011 data found no evidence they predicted returns. The report covers two competing explanations for why, and what a separate, more nuanced 2000 study found about chart patterns carrying some information without necessarily producing a profitable trading edge, on page 9 and page 11.
Does this report tell me how to trade golden and death crosses?
No, and it isn't trying to. It lays out the published research, an original noise-baseline simulation, one dated illustration from an actual stock's history, and a second simulation pricing out what trading every cross costs against holding a position and doing nothing, then hands you an eight-question checklist on page 18. Exactly one stock is named anywhere in it.
What did the cross-trading simulation actually find?
Eight thousand simulated 20-year runs, comparing a strategy that mechanically trades every golden and death cross against a simple buy-and-hold position with the identical underlying market assumptions. The switching strategy's median result landed near $21,200 from a $10,000 start. Buy-and-hold's median result landed near $36,200. The switching strategy only beat buy-and-hold in about 8.8 percent of the runs, once the lag and a stated illustrative switching cost were applied. The full method and every assumption are on pages 14 and 15.
Sources & Further Reading
-
Brock, W., Lakonishok, J. & LeBaron, B. (1992). “Simple Technical Trading Rules and the Stochastic Properties of Stock Returns.” The Journal of Finance, 47(5), 1731-1764.
Tested 26 rules, including moving-average crossovers, against 90 years of Dow Jones data. Every rule beat buy-and-hold in this original sample.
-
Fang, J., Jacobsen, B. & Qin, Y. (2014). Published in Review of Financial Economics.
Re-ran the same 26 rules against 1987 through 2011, 25 years the original study never saw, and found no evidence they predicted anything.