forex-scalping.com

Crucial Adjustments for a Rigorous 10-Minute Test

If you implement the 10-minute hold validation loop described in your testing framework, enforce these hard constraints in your code:

The Crossing Spread Trap: When simulating a market order entry, your code must fill long positions at the current Ask price and short positions at the current Bid price. Exits must do the exact reverse (close long at Bid, close short at Ask). Failing to account for this round-trip crossing of the spread will skew short-term 10-minute results into a false positive state.

Time-Based Exit Integrity: If your backtest uses a hard 10-minute clock, ensure the exit price is pulled from the historical feed precisely
T+10 minutes, and verify that a stop-loss or take-profit barrier hasn't been breached inside that 10-minute window via higher-resolution data (e.g., M1 bars or tick data). Using only close prices on an M15 or H1 chart for a 10-minute target introduces severe lookahead distortion.

Thanks for taking the time to go through it properly, PTScalper. I agree with you that a meter on its own is not an edge. When I ran my own tally, I logged about 150 meter picks over three weeks of London and New York, and the pair it flagged was already roughly a third of the day's range into its move at the time I looked. That is the lag you describe showing up in plain numbers.

Where I would push back a little is on dismissing it completely. I stopped using it as a trigger and now only use it to cut my watchlist from the usual dozen pairs to three or four, then I trade off the raw chart. That made me less scattered, but I cannot honestly say it raised my win rate.

Your point on costs is the one I would underline. In my log, the picks that looked fine gross turned flat once I took off the spread on the weaker crosses, which the meter loves to rank at the extremes. I now only count a meter pick if the pair is one I would trade anyway.