Anyone who wanted to see charts without missing bars - here =) - page 10

 
nikkei:
komposter I'm having a problem here, can you tell me what to fix? All the details are here http://forum.kimiv.ru/viewtopic.php?t=177
As far as I'm concerned, it's all solved now, isn't it? ;)
 
komposter, thanks for the tweaked expert. I'll test it and let you know the results!

About the H12, the problem has already been solved. I did some research and found that what I suggested you to do introduces some systematic component to the variance. I examined different variants and combinations of samples and came to a conclusion that the Close price is the best one for making channels. I can get much better results based on it. Therefore, I do not need to modify it at all!
Thanks for your willingness to help! :o)
 
solandr:
komposter, thanks for the tweaked expert. I'll test it and let you know the results!
The Expert Advisor fixes only messages
2006 12.09 03:26:29 29 WithoutSunday_4: invalid handle -1 in FileWriteDouble
2006.12.09 03:26:29 29 WithoutSunday_4: invalid handle -1 in FileWriteDouble
2006.12.09 03:26:29 29 WithoutSunday_4: invalid handle -1 in FileWriteInteger
2006.12.09 03:26:29 29 WithoutSunday_4: invalid handle -1 in FileSeek

It doesn't solve the problem, it just makes fewer errors and is generally correct ;)
 
solandr:
I have investigated different options and combinations of samples and have come to the conclusion that the Close price is the most optimal one for plotting channels. It produces much better results.
It is not in vain that everyone says that Close can be used for analysis. For example, analysis by the price (High+Low)/2 is more predictable, but less practical, because you cannot make a decision in the middle of bar development that we already have the price (High+Low)/2 with 100% confidence, while you can get Close at the moment of a new bar appearance.
 
I think that using schemes like (H+L)/2 or as in my case before (O+H+L+C)/4 is similar to working by muving, that is working with filtered quotes, as a result of which the variance becomes less - it is simply "filtered" by averaging. In general, Close rules - now I'm sure of it!!! ;o)))
 
https://www.mql5.com/ru/forum/50458 solandr 17.01.07 10:41

Thanks very much for the links! It was very interesting to read it! The author of the site suggested a representation of price in different coordinate systems, and compared different methods of representation.
It would be cool if an article like this appears somewhere on www.mql4.com. It would be interesting for both beginners and experienced traders who have not yet found this information. If someone (for example komposter) could write a script (expert) that could display charts in different coordinates in MT4 (in offline mode of course), then such a script (expert) would be a hit of downloads in CodeBase section at once!!! :o)
 
solandr:
It would be great if an article like this could appear somewhere on www.mql4.com
I'll keep that in mind ;)
 
komposter писал (а):
solandr wrote (a):
It would be great if an article like this could appear somewhere on https://www.mql4.com//
I'll take note ;)

I suppose the initial data should be M1, and from it the required re-calculated candlesticks will be formed according to different formulas (re-calculation parameters must be set by a user - first of all volume, delta, and similar parameters, according to the author's idea). In this case, of course, since MT4 has a rigid binding to time, the obtained candlesticks will non-linearly coincide with the MT4 time grid, but it does not matter in new coordinate systems where time is not present. When writing recalculated candlesticks into an autonomous file, the open time of bars can be set by the time of the first candlestick of the M1 period, from which the recalculation bar was started, and new candlesticks can be written into the autonomous file of the M1 period. But it will be absolutely unimportant for further analysis of quotes in new coordinates and will be just a technical detail of idea realization in terms of existing MT4. Judging by the results shown in the links, it can significantly change the market representation. For example, it may change the appearance of the existing quote distribution in time-price coordinates to one of standard distributions in new coordinates, which will allow to make more accurate forecasting of regression channel boundaries.
 
solandr:
I am assuming that M1 should be taken as input data.
I don't think I'll be implementing it any time soon...
I've put this idea on my to-do list, but I don't know when I'll get around to it.

So if there are volunteers, go for it ;)
 
komposter писал (а):
solandr wrote (a):
I'm assuming that M1 should be taken as the raw data.
It's unlikely I'll get around to implementing it anytime soon...
I put this idea on my to-do list, but when I get to it, I do not know.

So if there are volunteers, go for it ;)

It's been a long time in the making...