Honest, reproducible backtests of the strategies, indicators and prop-firm claims everyone shares. One fixed protocol. Every strategy is compared with the same trades placed at random, to see whether it beats luck. Every version we try is published, the losers too.
The two that passed every gate: candidates, not proven, now tested on new data
The same rules for every test: rules dated before the data, costs on every trade, and development, validation and holdout periods. Published once, linked everywhere.
The same exits and sizing on entries moved to random other times. If a strategy can't clearly beat them, no edge has been demonstrated.
No cherry-picking. We say how many variants were tried and how many survived.
Every strategy gets a fair trial and a proper burial.
Six years of buying dips. The luck was real; the edge never was proven.
Green paid. Red never did.
Looked great on the chart. Not on the ledger.
Crossed about 5,000 times. Proved nothing.
Passed five gates on the hourly chart. Validation said no.
Four cents before costs. Thirty cents of costs.
The five-day average turned. The price did not follow.
It waited for the pullback. The pullback was random.
New market, same rules, same result.
Seven trades in ten won. The other three cost more.
Price came back to the VWAP. The spread came with it.
The right hour was not enough.
Not 80 per cent. Not quite enough, either.
The character changed 678 times. The odds did not.
Zero before costs. Twenty-three losing years after.
Five gates on the Dow. Two at most on gold.
Five gates of six. Not enough evidence yet.
The lines looked clean. The results did not.
The quiet day stayed quiet.
Three small gaps in four filled. The trade still lost.
Simple to trade. Not simple to profit from.
Green crossed red six thousand times. The spread kept the score.
Built for stocks. Beaten by holding them, and by chance.
It followed the money. The money did not follow back.
It stopped, it reversed, it paid the spread. Every eleven hours.
The pattern beat chance. The spread beat the pattern.
Five gates of six. A little weaker every period.
It worked on the Nasdaq. The S&P did not follow.
Better than chance. Not by enough to pass.
It worked on the day in the video. Not on the other 1,500.
The gap filled. The stop came first.
It saw the trend. It could not tell which way to bet.
The pause was real. The breakout meant nothing.
It beat chance. It did not beat the spread.
Gold rose at night. Mostly because the night is long.
It dodged the worst crashes. It never beat holding.
Half its trades won, half lost. The costs took the rest.
The paper's numbers came back. The edge did not.
Five of six gates. Its first nine years could not prove it.
Bought oversold, sold overbought: worse than chance.
Every 30 minutes or every 15: nothing before costs.
Five of six gates. Gold's lean years stopped it.
Bought the dip at 61.8%. The dip kept dipping.
One trade in three wins. Only the daily cross came close.
Everyone watches the line. Following it did not pay.
Forty per cent, every time. It needs more.
Six crossings a day. Costs ate the rest.
Wins two trades in three. Loses the money anyway.
Nothing before costs. Costs did the rest.
Beat the control on fresh data. Nine years gave too few trades.
Passed every gate in New York. Now it has to do it again on data nobody has seen.
Close in New York, nothing in Asia, worse in London.
Beat random days, earned nothing after costs.
The first half-hour said nothing about the last.
The first half-hour said nothing about the last.
The gold result did not travel.
Profitable in every period. Still not an edge we can prove.
A precise level. No proven edge behind it.
The 15-minute ribbon beat the control. It did not clear the bar.
Paid in gross. Costs took most of it.
A hint of an edge. Not enough to bank on.
Chased every breakout. Bought the top.
Profitable. The divergence's share is unproven.
The lines crossed. The edge didn't show.
96 tries on four pairs. No survivors.
Reversing flips the result, not the costs: 1 in 4 ended positive, the same as leaving them alone.
The stop did not make money. It made the worst day smaller.
Neither style won after costs. Breakouts came closest to an edge.
A good result on one market predicted nothing about the next.
Profitable before costs, losing after: 137 of 416 versions.
The target sets the win rate. Only the entry sets the edge.
Three papers, four markets. One rule passed, in one session, on one market.
109 versions. One passed every gate, until the data check.
A challenge rents you size. It cannot rent you an edge.
The spread is paid on every trade. The edge, rarely.
The rules reward variance. Only an edge gets you paid.
The challenge rewards variance. Skill is optional.
3 trades on ES, 15 on gold. Too few to bury it, too few to believe it.
Passed every gate on gold. Then the data check found thin ground.
The gold survivor did not travel. 12 tries on EUR/USD, 12 funerals.
The curves on the tombstones are real net equity curves from our tests (protocol v1.0), usually of the best-known version of each strategy. The bars on Prop-Firm Evaluation Odds show simulated pass rates.