The concept of an ideal indicator - page 6

 
YurijM писал (а) >>

My attitude to those indicators based on historical prices is as follows: you look out the window, see the sun and say "It's not going to rain". But there's a cloud on the other side of the house and it's pouring in two minutes. Once again, you CANNOT PREVENT HISTORY (IMHO). Or have predictor indices already been invented? If so, tell me how they work and I may start to make predictions relying on coffee grounds. :)

The thing is that the indicators themselves never predict anything. They are not designed for that. They are used to look for regularities on the market.

 
Korey писал (а) >>

An example of an indicator that is rejected because it cannot predict the future =ZigZag, as well as many other methods with the edge effect.
The mental attitude "I am looking for the predictions", which was formed during the first steps of the acquaintance with the market, - prevents you from using the whole arsenal of TA.

The TS should "predict" the future. Indicator Does not predict. An indicator Predicts.

LeoV wrote (a) >>

The thing is that indicators themselves never predict anything. They are not designed for that. They are intended for searching for regularities on the market and based on these regularities that are detected through these indicators, you can try to forecast the market.

+1 I can't help but support

 
LeoV, you're absolutely right. The turkey is not secret, it's in the codebase. DEMA by DiNapoli. Naturally, it's not perfect. No one is diminishing the capabilities of neural networks. They're the future. Maybe we will see the support of neural networks in MT6 or MT7. Although it is not certain. As long as special software rules.
 
LeoV писал (а) >>

The point is that indicators themselves never predict anything. They are not designed to do that. They are designed to find patterns in the market and based on those patterns, which are detected by these indicators, you can try to predict the market.

That's right. I'm talking about creating an indicator that allows making accurate forecasts. The discussion shows that this indicator is rather a system integration of many market characterizing indicators. It may look like this - in the main chart window - the channel of linear regression and the lines of supports - resistances and targets and, perhaps, entry points (as derivatives of the reversal oscillator), vertical news release lines (may be of different colours and thickness - depending on the importance). Additionally, in the basement - the oscillator-histogram of market conditions. For example, 6 - levels, 3 - up and 3 - down. 1 - calm, 2 - flat, 3 - trend. If 2 and 3 levels are divided into a couple of sublevels, it is possible to show the "degree" of the phase (it is possible to make a traffic light in the chart corner instead of a histogram). 1 min before the news release, when the price reaches the support/resistance line or target, and when the indicator changes its state - audible signals explaining possible developments. The decision is made by a trader or an Expert Advisor. Approximately.

 
We come to a very interesting point.
The requirements for a perfect indicator can be met... But only on H4, D1.
 
IMHO histograms are not very reliable. Korey, you could also use H1 for accurate inputs.
 
For input accuracy you need a different indicator ideal, i.e. Ilya, we have discussed the indicator-observer
now we can discuss indicator-operator == requirements for an indicator which can be trusted with entry/exit points.
 
To FION.
+1 Clean, nicely done, but why put it out there?
 

to FION

By the way, about the channel - I was experimenting with building a channel on the MA hump, it may be like a zigzag, the steps are smaller, the EA is calmer, but the channel becomes wider and the analysis is bigger.

 

кстати насчет канала

You can use NonLagMA_v5 . The channel holds well, it's just a question of fitting the parameters.

I don't know who builds saport-resistances, there are many variants at the moment, I really like the following method by fractals, since we are talking about them. Although formalising it head-on will lead to deplorable results.

As for entries and exits, here is the case. One part of the issue is identification of the trend in its embryonic stage, the other part is the exit at its peak and not in the rollback position when a part of profit is lost. There is a lot of literature devoted to this issue. Basically, my opinion is the levels. Either Fib or again just resistance-support. Although here the question is ambiguous. Let us take the Eurobucks as an example. The resistances can be drawn in a purely hypothetical way. The Euro goes where it hasn't been before. Here the so-called psychological levels and the fundamental analysis are relevant.

Therefore the question of the identification of the exit is undoubtedly difficult. The price seems to exit well by the DMA, but again, on the history. There are plenty of those pseudo exits during the trend, and increasing the period, we miss even more profits.

Further on, according to the waviness of the trend. Many people are familiar with the works of Elliott, Williams and others. I remembered them by the question of entering on the pullback waves. Here IMHO indicators with a fixed range are unambiguous. The choice is purely personal.

Reason: