We take the trading strategies going viral and run them on real price data, with real spreads and costs, across three independent data feeds. Then we publish the result. Most of them fail. We publish those too.
One email when a new test is published. Nothing else.
Every result below comes from our own backtests. Where a test ran on fewer than three feeds, we say so.
Time-of-day drift, liquidity sweeps and stop hunts, momentum, mean reversion, session breakouts, calendar effects, volatility regimes.
beat random entries once costs were included
Each idea was built on 2024–25 data, then checked once on 2026, against hundreds of random-entry versions of itself. The best breakout variant from 2024–25 flipped to losing in 2026.
A hedge with no stop loss: every closed trade is a winner, because a trade without a stop can only close at its target.
account breached its floor while the balance sat at an all-time high
The losers never showed up in the trade history. They were still open. On day 60 the balance read $260,929; real equity was $177,969, below the $180,000 limit. The open loss later touched −$13.95 million on a $200,000 account.
Double the position after every loss until the cycle closes in profit.
in open positions needed to survive the 72nd cycle
It banked $1.19 million and closed 71 of 72 cycles green. The one it didn't needed 83,886 lots open at once, 487 times what a $200,000 account can carry, and sat at −$2.1 million.
If a trader keeps losing, just do the opposite of everything they do.
for the mirror, against −$33,808 for the original
The losing trader used wide stops and tiny targets, so they won 81% of trades but needed 90.9% to break even. The mirror inherits a stop ten times tighter and pays the same spread on every trade.
$500 into each of the 20 largest cryptocurrencies, sell any coin at 2×, repeat.
from $10,000, depending only on the start date
Starting January 2024, 13 of 20 coins doubled and the plan made $15.0k. Starting January 2025, one coin doubled and every version ended between $6.7k and $7.6k. The rule didn't decide the outcome. The entry date did.
An indicator's bullish signal candles on the 4-hour chart, bought with a 1:1 take-profit.
per trade, against +0.052R for buying every bullish candle
It lost in 2024 on both feeds that cover it and won in 2025–26 on all three. Buying any bullish 4-hour candle with the same rules did almost as well. The profit was gold's two-year rally, not the signal.
The same rules apply whether a strategy comes from a famous account or a reader's message.
Prices differ between brokers. A result counts when it holds on three independent sources. When it doesn't, we show the disagreement.
Spreads, commissions and the actual bid/ask, not the mid-price a chart shows you. Plenty of strategies only work for free.
Rules are fixed on one period and tested on a later one we never tuned on. Many ideas that look great in the first period fail in the second.
If entering at random with the same stops and targets does as well, the signal isn't doing anything. We run hundreds of random versions to check.
A strategy that holds its losers can show a perfect trade history while the account is underwater. We measure what the account was actually worth, minute by minute.
We report every year in the test, including the bad ones, and say what the market was doing at the time.
Every test in full: the exact rules, trade lists, equity curves on each feed, the year-by-year breakdown, and what would have to be true for the strategy to work. Members also send us strategies to test next.
Opening soon. Join the list and you'll get the founding price.