First, the sample size is much, much too small. Personally, I don't look too far into anything with a PF under 1.5 but that's just my personal criterion, and not everyone will agree with it, and it doesn't necessarily mean the systems aren't sensible ones anyway - just that they wouldn't interest me. I also don't like the average loss being bigger than the average profit (and significantly so, in one case), but against that doesn't necessarily invalidate the thing. My general sense, given the frequent differences between backtesting results and live results, is that these certainly wouldn't be for me, anyway. For myself, I'd look no further into them, as they stand, anyway. Just my perspective, and probably not what you wanted to hear - sorry.
I have automated trend lines and make them thick, means to give a variance of 1 to 1.5 ticks, life is not rigid nor is trading. But for twenty years I did everything by hand as when I started in 1978, there was no home PCs. Am glad I didn't have home computers then cause brain to hand, you learn better. Yes, look on 60 minute charts for entries as they are considered stronger by the masses. I was saying using pivot lows/highs on the five minute and putting stops like 4 ticks below them or above, not too close otherwise noise trips them. I don't trade by using factors, I get a signal and it is done, "factors" is like an illusion, or saying when you are going to die, we never know, so software sees a "touch" and takes a position. The factors or percentages are back tested and real time to give clear idea of improving method as far as risk management which includes rules on when not to take some trendline signals. Like I use certain degreed angles I won't take as they show higher percentage of losses am willing to accept-often they still profitable overall but if I don't make so much a trade, not worth the risk. Similar to making minimuim wage or $150 an hour. But it all based on backtesting, like the 60 minute system I have backtested it 25 years of 60 minute bars, so I know what to expect, and each month it backtested again looking for either improvements if overall yearly changes which has it's own unique rules I have discovered through the years, ie 2008, I don't normally take trendlines when severely angled down as they often reversing when coming back to a trendline, but knowing from other free falling time periods, I found I can take those signals, so the more you can identify in terms of year to year and percentages of normal activity, you can adjust settings or code your automation to where it will make changes based on previous years.