Even the gaps are there - they're simply covered up (masked) as a result of consolidating data (moving to a larger TF) RN
Disagree, some particularly in equities would have nothing, but that seems to be the common denominator in this thread.
Compare how? Eyeballing it can be deceptive. What else did you have in mind? [/QUOTE] Definitetely not eyeballing. Statistical probabilities of turns working vs failing vs risk rewards of the various timeframes.
That sounds like just assuming some random strategy as a standard. I'd use a choppiness indicator. When the choppiness value is low, you can make money with trend following. When the choppiness value is high, you can make money with rtm strategies. So the iffy part is at the halfway mark, which either means equal parts trend and chop or more likely randomness, which is untradeable. So take the choppiness reading of your 5000 1-minute data and the choppiness reading of your 5000 eod data. Whichever one is farther from the halfway point is the more tradeable timeframe. You should probably use something better than the Driess Choppiness Index, which is pretty kludgey IMO.
I think you deviated from my original statement. All I'm saying is that taking risk and reward into consideration a turn in a higher timeframe is not more reliable than the ones in the smaller ones.
Just simple bar turns, new bar taking previous high or previous low or in the case of outside bars, both.