The moment you try to fit and tweak data and thereby generate some buy/ sell signals, you are wasted. I don't understand why you think that tweaking your data will make you money.
Certainly bars make sense intraday, what a strange opinion. Why do you say that? However the posting was targeted at Abattia. But I would like to hear your reason for my current question
I mean grouping candlesticks by 1, 5, 15 min etc. is arbitrary. The meaning of OHLC makes sense for Day, Week , Month but not make sense intra-day.
And why is this so? Why doesn't a 5 min bar make sense? Or a VolumeBar, TickBar or RangeBar. I don't understand
Really doesn't matter if it's a candlestick or a OHLC, just an accumulation/ grouping over some interval and type.
OK, if OHLC is equivalent of Open, High, Low, Close, then they have a "natural" meaning for Daily, Weekly, Monthly, etc. intervals - you have Open for the Day, Close for the Day, High for the Day and Low for the Day. Now in intra-day interval of 1 minute or other minutes OHLC do not have this "natural" meaning. What Open for 1 minute means? Or Close for 1 minute? Now if you change the interval from 1 minute to 2, or 3 or whatever you will get a "phase shift" - if you know what I mean. If bars are represented visually as candlesticks you get different coloring if you change the minute intervals - it is not "natural" and means nothing because intervals lengths are arbitrary.
How does a day, week or month make more sense the something that is a multiple of a second? Not all trading days are the same length, neither are weeks or months. And, while their length may vary, they are all aggregates of a unit of time that is defined by a fixed number of oscillations of a certain atomic particle. Besides, what happens if the stock market were to merge into a truly global 24/7 trading environment? Wouldn't the daily begin and end times become just as arbitrary as you say intraday intervals are? IMHO, I think you're putting too much emphasis on all this. All charts, bar, candle, P&F, etc. are nothing more than visual analogs for a stream of trade data. What really matters is that the algorithm used to construct them not only produces them "correctly" but, that it can also graciously handle delayed openings, early closes and thinly traded instruments.