Update, 24 October 2026: these tests were made for versions 1.00 and 1.01. Version 1.02 adds a third strategy and was tested on 30 contracts back to June 2019; the product page has its numbers.
Most product pages show the version that worked. This one lists the ideas that looked good and did not survive the test. All of them were measured the same way: real ticks, contract by contract, the full period.
A quieter-market filter: three thresholds, one kept
Strategy 1 skips entries when the market is too quiet. The threshold matters (backtest after costs, 11 contracts):
- lower threshold: $26,843;
- middle threshold: $27,406, kept;
- higher threshold: $25,555.
The higher threshold blocked more bad trades, but it blocked more good ones too.
The next three come from an earlier research round on an earlier version of Strategy 1 (April 2024 to September 2026, before costs), so compare them with each other, not with the product page.
Moving the stop to breakeven
Moving the stop to the entry price after a first gain sounds safe. In the test it cost money: between $878 and $2,067 less, depending on the trigger. Trades that went back to the entry and then reached the target were closed at zero.
Stops based on volatility instead of fixed stops
Stops set as a multiple of recent volatility (two settings tested) earned $1,187 to $1,707 less than the fixed stops. Wider stops in fast markets made the losses bigger exactly when they happened.
Skipping the "bad" hour
A table of profit by hour showed one weak hour, so we tried skipping it. The result got worse ($247 less): the blocked setups simply fired an hour later. Profit-by-hour tables mislead when the same signal can appear again later.
Opening range breakout: the variants
For Strategy 2, in an earlier test (April 2024 to September 2026, before costs):
- entering at the market instead of waiting for the breakout: $6,044 instead of $14,067;
- no profit target: $9,604;
- a 15-minute opening range instead of 5 minutes: more profit ($17,008) but a drawdown of $4,827 instead of $2,381 and a worst single trade of $615, so it was rejected;
- a stop based on volatility instead of the other side of the range: $7,443.
Why this matters
A rule set can be tuned until the backtest looks perfect. Each extra rule fitted to the past makes it less likely to hold in the future. We kept the version that was more profitable after costs without making the losses deeper, and the live test is there to show whether it holds.
Backtest results. Past results do not predict future results. This is software, not investment advice.
