"if this thing happens then this other must be true"
Sort of a long post -- sorry. One of my buttons got pushed...
@4fxonly -- of course -- it is good to take note of observations and deduce consequences. When dealing with Black Boxes this is what helps set us humans apart from other species. Some humans are better at it than others. Some humans are too lazy to even try. You are definitely one of the smart ones.
Indicators are just like any other program -- they are a chunk of code that get executed. That code may be complex or simple -- whatever. If the code never or rarely gets executed, then its complexity doesn't matter when it comes to CPU usage. Only when the code executes will it take up CPU cycles. For most indicators, for better or worse, their code gets executed every tick. For a particular symbol -- all time frames receive the same number of ticks (with one minor exception). However, different symbols receive different number of ticks depending on your broker's data feed. If you attach a chart to EURUSD you may receive several ticks per second. If you attach an indicator to an exotic symbol, perhaps once an hour -- I don't know.
My point was this -- you don't need to understand the inner workings of MT4 or even how to program. After advocating that I take this indicator down a particular path, you suddenly decide its not possible and indicate you are no longer interested -- all within a few minutes. I don't need help programming, but as I've mentioned a few times previously, I never fully understood @George's methodology and never took the time to mark-up charts -- I just didn't have the patience. So, I would have expected some help in validating the new indicator did the right thing.
It seems to me that when given guidelines people fall into one of two categories -- those who take things seriously
and those who take things literally
. I fall into the former. If @George suggests using 5000 bars I take that to mean "a lot" of bars -- i.e. somewhere between 500 and 5000. Someone who takes things literally thinks it must be exactly 5000 bars or the methodology won't work. I doubt that @George tried 1000's of combination and after 20 years discovered 5000 bars works and any other number doesn't -- perhaps that is the case, I don't know. Furthermore, if you were to take 100 people and ask them to mark-up M15, H1 and H4 charts (with 5000 bars) by drawing lines along TOPS and BOTTOMS and then subsequently removing them, how many of the resulting charts do you think would look exactly like George's? I'm guessing not many. But they would probably be good enough
to be used with @George's methodology.
Whenever I encounter a problem that on the surface seems intractable -- e.g. 5000 bars is too slow -- rather than throw up my hands in despair I'll look for a simplifying assumption that allows me to move forward. That is, I'll take the guidance seriously
and understand I need quite a few bars, though not necessarily exactly 5000. Others take the guidance literally
, and if it can't have exactly 5000 bars then they figure its hopeless.
As a side note -- some applications lend themselves to be taken seriously, others to be taken literally. For example, in the theoretical
world, such as physics, it's probably a good idea to take things literally. The laws of physics tend to fail if you take short cuts. On the other hand, in the empirical
world it often is okay to figure things out by trial-and-error since you might find a solution that is "good enough" to get you to the next level.
It's also interesting to me that as one door closes another opens. At the same moment the door closes on 5000 bars, then @SwingMan comes along and offers some insight into how using higher level fractals to pick out TOPS and BOTTOMS might actually yield better results (than combining multiple time frames). A coincidence, perhaps. An opportunity -- definitely. If I get nothing else out of this thread -- learning about higher-level fractals from @SwingMan's posts made it more than worthwhile.