That's interesting about the multitester, I've really only been using it to see if a strategy is applicable to other pairs, how are you applying it (like other timeframes etc?)

I agree with Blaiserboy, you'll be passing up a lot of potential strategies if you keep things that tight.  I generally run them 15-25% dd and 25-50d stagnation for a first pass.  Most of the time you can pretty easily look at the equity curve after the fact and filter out things that are awful but sometimes there are outliers that can be amazing with a small tweak after generation (for instance if the sl/tp the generator came up with allowed a big drawdown but the rest of the strategy was good.)

The 1m tf is tricky, not sure if it's bad luck or not but I've been though a ton of 1m strategies that looked amazing on fsb, alright (but not as good) in demo and performed like crap live.  Still have a handful that have held up but it's easier to get a solid 5m than 1m by far (I'm sure there's probably people that say that 4h and 1d are even better but those take more patience to forward test. )

One other thing that might help, I always run at 40% oos during generation and optimization.  It makes it way more obvious which strategies are over optimized when you compare them in the repository collection.