I don't believe in best timeframe, I go into it with idea of which timeframe I want, cause testing over number of timeframes goes back to finding most money unless you have ten-fifteen years of tick data you are back testing and back test in three year increments and each side of "best" is making slightly less profits. Profits should look like a bell curve distribution regarding best parameters in indicators and timeframes, when it does not, you not going to have a smooth parameter and most likely will fail.
Have lost tons on "making" strategy profitable, I have developed like "units" one unit is patterns that look over "content" of so much time like in day trading, so like in one minute bars, 90% of the unit is only concerned of last 120 bars of different patterns that would negate a good signal. This "unit" has increased through the years and is used in all systems and all timeframes, cause the look back is 120 bars regardless. So each area I use units, like risk unit, open trades unit, target unit, etc...I try to keep units as close as I can in percentages if possible to they can be used in other systems, where I can't cause the underlying system might have strange entries then those systems will have different units.
If my sample size is not 3,000 or more, drawdowns and other expectations are useless and just a fantasy to me. 6k can be low or high depending on account size. Are you trading corn? Cause if it $300 min on what size account, sounds like if all this worth the effort, if $300 is on crude oil, it very low whereas $300 on Eurodollars is pretty good amount as margins in Eurodollars are very low overnight-usually like $350/400. Matter of fact, if what you doing is not working out, check Eurodollars as it is heaviest market on the planet for volume. For a number of years, always used to back test a system on Corn for day trading or long term, if it makes money in Corn, often times it make money in almost anything else for ME.