Most strategies we build or try never go on sale. That isn't modesty; it's the job. A backtest can make almost anything look good, so before anything is offered it has to clear the same five tests, whoever built it.
We have scored 21 strategies so far: 2 are on sale, 1 passed but isn't offered (it's someone else's product, and even one micro contract risks too much for a typical prop account), and 18 were rejected. Failing any one of the five tests means rejection, however good the rest looks. Every one has its own page on the Tested list, with its results and the reasons.
One thing to say first. The bar was set in October 2026, after we had seen the results of the strategies we offer, and then applied unchanged to every strategy, including those two. A bar set after the fact can be shaped to fit, which is one more reason the forward record matters more than any backtest.
The five tests
R is the amount risked on one trade. All figures are after commission and slippage.
- Edge after costs. At least +0.10R per trade on average. Below that, a small change in fills or costs can wipe out the whole edge.
- Enough trades. At least 100, so the average means something. Ten good trades in a row is a story, not a result.
- Holds over time. Profitable in both the first and the second half of the test period. A strategy that made all its money in one market phase may only work in that phase.
- Not one lucky streak. Still profitable with its 10 best trades removed. If a handful of outliers carry the result, the next few years may not contain them.
- Fits a prop account. A deepest drawdown of 25R or less, so a $2,000 trailing limit holds at up to $80 risk per trade.
Passing all five isn't the end. A strategy is only offered if it also beats what we already offer on the same market, measured by the Sharpe ratio of its daily results over the same period.
What failed, and how
Most of the 18 rejections weren't near misses: 14 of them failed four of the five tests. Counting every failed test across the 18:
- 16 failed edge after costs
- 16 failed fits a prop account
- 15 failed not one lucky streak
- 14 failed holds over time
- only 1 failed enough trades
Having enough trades is the easy part. Having an edge that survives costs, time and the loss of its best days is where almost everything falls down.
Four examples show the tests doing different jobs:
- LIQMES averaged +0.13R per trade over 584 trades, above the edge test, and was profitable in both halves. Remove its 10 best trades and the average drops to -0.04R: the profit lived in a few outliers. Its deepest drawdown of -39.7R also failed the prop account test.
- LIQMNQ was profitable, passed the first four tests and averaged +0.18R per trade, but its deepest drawdown reached -66.2R. Keeping that inside a $2,000 trailing limit would mean risking about $30 a trade, too little to be practical, so it was rejected.
- VWAP pullback, gold passed four tests, with a drawdown of only -14.5R. It failed one: an average of +0.05R per trade, half the bar. Close isn't enough when costs and fills can move.
- Inside day breakout, gold had just 37 trades in its test period, the only rejection for too few trades, and with its 10 best trades removed it averaged -0.91R.
CFTC Rule 4.41 statement. These results are based on simulated or hypothetical performance results that have certain inherent limitations. Unlike the results shown in an actual performance record, these results do not represent actual trading. Also, because these trades have not actually been executed, these results may have under-or over-compensated for the impact, if any, of certain market factors, such as lack of liquidity. Simulated or hypothetical trading programs in general are also subject to the fact that they are designed with the benefit of hindsight. No representation is being made that any account will or is likely to achieve profits or losses similar to these being shown.
What passing means, and what it doesn't
Passing doesn't make a strategy safe, and it doesn't predict future results. It means the backtest didn't depend on a handful of trades, one lucky period, or a drawdown no prop account could survive. Those are the most common ways a backtest misleads, but not the only ones: Five silent traps in a TradingView backtest covers how the backtest itself can be wrong before any test is applied.
The real test is the one no backtest can run: what happens next. Both strategies we offer have been forward-tracked since 1 October 2026, and every trade is on the forward record as it closes, losing trades included.