Any tool might be used in a wrong way.
My current WFO set-up with 4w IS/1w OOS to find params is quite robust as it covers a repeating "economic reports" cycle.

RE using Ask prices:
- Until some time ago, I also did not really believe that with currently tiny spreads in FX, this would make any difference. And so, used Ask in all calculations.
- However, just by chance, I decided to test the system with mid-point and the results proved somewhat different/ worse. I then tried to understand why and found that for the better part of the test period the market was going up. So, using Ask prices did overstate the results! Not by a lot, but still! Since then, all my testing is based on mid-point.

- As for Ask-based execution: any significant widening of the spread on the ask side(glitch in a system, liquidity,etc) will lead to the premature triggering of simulated buy stop orders.
It will also lead to bigger slippage/losses in a falling market for sell stop orders (when bid runs away faster; and/or ask might disappear altoghether).