What techniques and methods can be used in multiframe indicators to prevent getting a pretty picture due to peeking into the future on higher TFs ? - page 8
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Wrong indicator. It's peeking. Close[0] price is the same on all TFs. And it should dangle on the unfinished bar of the higher TF as well as it dangles on the lower one. Therefore it is not correct to show it permanently at the level of the future close price.
It is not Close[0], it shows a fragment of the indicator on the history and explains why the indicators from the higher timeframe are misleading.
It's not Close[0], it shows a fragment of the indicator on the history and explains why the indicators from the higher timeframe are misleading.
Yes, you're right, it's the history that turns out to be misleading. But I wonder if it is impossible to specify that the step to the bar closing price occurs at the moment of bar closing. Or is it done deliberately to mislead?
Not intentionally, the candle already exists in the history with formed OHLC. It is impossible to request the state of the candle at a certain time of its formation. This data is on smaller timeframes.
The time discretization is 1 candle, the same time throughout the candle.
Yes, you're right, it's the history that is deceptive. But I wonder if the step to the bar close price can't be made at the moment when the bar is closed. Or is it done deliberately to mislead?
We cannot do it deliberately. That is how it is done using the simplest approach - iBarsShift().
And how to make the step happen at the moment of closing? For a bar being formed, it would move all the way up and down.
I wrote here, there is an option - for each bar of the main timeframe to calculate the value of the indicator of the older timeframe,
Then the indicator looks like a normal indicator and not in steps.
How much is the older timeframe different from the current one? Prescribe multipliers for each timeframe and you have a non-drawing multitimesframe, if the current one does not draw on the current one.
ATR = 14 on M15 and ATR = 14*4 (H1)
It does not work with all indicators
What kind of indicators it does not work with? It should work with all of the averages and regressions.
In general, probably there is no more reliable solution than just increasing the resolution by mapping to a younger timeframe. Why should I calculate using a higher one and spend too much time trying to get rid of jumps and draws to save 0.01% of the CPU?
Which ones, for example, don't work? Any averages and regressions for sure.
I don't know about all of them, but RSI definitely makes a big difference.
I don't know about all of them, with RSI there is definitely a big difference.
Differences in absolute values, but not in form, which is the most important thing.
Differences in absolute values, but not in form, which is the most important thing.
Strong differences.
Which ones, for example, don't work? With all averages and regressions it should.
In general, there is probably no more reliable method than simply increasing the resolution by converting to a lower timeframe. Why should we calculate using a higher one and try to get rid of jumps and skewness in order to save 0.01% of the CPU?
"It doesn't work". It is not about TF as everything is great in theory. But in practice everything is not so happy. For example, we need to construct МА from D1 on М1. Theoretically, we take 1440 candles of M1 and use them for the calculation. ButM1 bars in D1 are not always 1440 in practice.It is good if it is 1434, as shown in the picture, or maybe 1200.
But in practice, things are not so rosy.