Predictive indicators - page 5

 
AlexeyFX:

And you must be forced by someone to trade on the short term with an uncomfortable pair.

The USDJPY was just an example, I was trading NZDJPY at the time and it was even worse at certain times. I try to use all possibilities and the shorter term does not cause me any problems. Moreover, while I take 100-200 pips off on the long term, on the short term it takes me several hours. That's not the point, it's about elephantiasis.
 

Yusuf, here is a link to a program that will turn the MT4 tester into a trainer for manual trading and you will be able to test your indicator and strategies built on it in an accelerated mode.

A demo version will be enough to work in the tester.

http://www.expforex.com/load/exp_virtualtradepad/exp_virtualtradepad/5-1-0-49

 
yosuf:

.... I want to find out how bad my indicator is, since it was said to be "drawing" and where is "non-drawing"?

I've written to you 10 times now. I'm finally getting it. I don't need another indicator, just yours. Your forecast and the fact that it worked. I even gave you a link to the charts to see what it would look like if the indicator could make predictions. You just dismissed them as nonsense that they asked me during my university graduation. You build this graph and you will see.

It takes 5 minutes to have an indicator. You've already wasted 1000 times as much time on the flubber.

 
yosuf:


My aim was to find out whether there are, in principle, other predictive indicators or indications of preferential price directions. Instead they suggest predicting force majeure situations, which are earthquakes.


Got it. You are writing your doctoral dissertation. You need references. And we're talking practice, profit...
 
gpwr:



Nikkei lost almost 20% after the Japanese earthquake. Do you think this Japanese market movement between March 11th and March 15th could have been predicted by looking at past prices?


Yes.
 
joo:

The price takes into account everything. This statement is correct but only partially.

The correct statement is "The price has already taken everything into account". The current price (now) contains information about the past movement (historical data) of the instrument + the external influence (the totality of all factors affecting the instrument).

Blue is always available. It can be easily analyzed by the simplest indicators and other tools.

The red is always unavailable. Because of this, what the indicators "say" at the moment is incomplete information. In this case, the incomplete information can be considered as inconsistent for practical purposes. The ideal indicator can consider only the blue, in reality, even the blue has not been fully considered.


Above we were talking about the indicators without adaptive, self-learning mechanisms. Indicators with learning are theoretically able to predict what red will be on the basis of the analyzed blue. Of course, the efficiency here can never reach 100% and in practice loss-making trades will occur. But the results will be better only for indicators with learning mechanisms.


It would be more accurate to put it this way.

Price = intrinsic market fluctuations under the influence of the past external influences + reaction to the current external influences.

The blue is indeed analyzable by the simplest indicators, but only in the past. This is interesting for the analyst and absolutely unnecessary for the trader. The trader needs a prediction of blue in the future, that is, an extrapolation of blue.

The red cannot be predicted by any indicators including training, and it is not needed, the blue extrapolation is enough for stable profits.

So, the task comes down to two actions:

1. Separate blue and red.

None of the publicly available indicators I know of is able to provide any acceptable quality of such a separation.

2. extrapolate the blue.

Let's not even extrapolate the price right away, let's try something simpler, let's say MACD. Try to plot the 1st and 2nd derivative of the MACD that you need for the quadratic extrapolation, you will see for yourself.

 
yosuf:


Now under the pressure of general criticism, I'll ask a direct question to everybody: Can you confront my indicator with any indicator that can do badly or well as it does, while guessing the price direction, the moment of reversal, the price value, where the price moves when the trend is over? If not, then don't criticise without any words, I think, but help me to improve the existing indicator

You're asking if someone can contradict your indicator with any indicator that is able to guess the price direction, the moment of reversal, price value, where the price tends to end the trend, even if it's good or bad.

In August 2008 before the Championship I made a prediction for 3 months ahead, for GBPUSD as an example and posted the picture on this forum.

Then the strongest anomaly happened - the currency crisis of 2008. But in spite of that the first month was predicted quite accurately and with pivot points and even levels. One person even saved it https://www.mql5.com/ru/forum/110698/page4 and compared it a month later to the real picture.

It's not a contradiction, because with polynomial regressions, I've had enough of them before and have given up on them. In particular, here is my early realization of polynomial regression https://www.mql5.com/en/code/8237 that many people rewrite and put into indicators or Expert Advisors.

Of course your idea of a symmetrical polynomial redrawing inversely is interesting, but as I wrote earlier, you shouldn't do that, because trying to use the rollover moment for including the expert will lead to serious losses, because it shows one way and then another.

The Expert Advisor needs a specific unambiguous signal. And only after the price moves away from this level, you can think about other actions. Therefore it is better to use channels around the pointer or divergence bands, in order to guarantee a safe distance.

 

Chaos theory has recently been one of the trendiest approaches to market research. Unfortunately, a precise mathematical definition of chaos does not yet exist. Chaos is now often defined as the extreme unpredictability of constant non-linear and irregular complex motion occurring in a dynamic system.

A striking example of chaotic behaviour is the movement of a billiard ball. If you've ever played billiards, you know that the final result depends on the initial accuracy of the shot, its force, the position of the cue relative to the ball, the assessment of the location of the ball to be struck, and the location of other balls on the table. The slightest inaccuracy in any of these factors leads to the most unpredictable consequences.

Another major property of chaos is exponential error accumulation. According to quantum mechanics the initial conditions are always uncertain, but according to chaos theory these uncertainties will rapidly accrue and exceed the admissible limits of predictability.

The second conclusion of chaos theory is that the reliability of predictions decreases rapidly with time.

This conclusion is a significant limitation for the applicability of fundamental analysis, which usually operates with long-term categories.

 

It's fun to read all this, especially when you know that the market is 100% predictable at certain times for me personally. And at all other moments - it is unclear only because of my personal reluctance, laziness to finish what I started (well, I don't have the patience to stare at charts for 12 hours a day, first in the visual testing, then in real trading for 2-3 weeks ;))

 

This Yusuf doesn't seem to want anything. I asked the question twice yesterday. No reply or goodbye. And he takes offense that no one takes him seriously. He should first learn to be nice to people, and then maybe people would be nice to him too. Like "do unto others as you would have them do unto you.

If he really wants the truth and not to swim in illusions, he has only one option. Put at least a couple thousand quid into a real account and start real trading. And he may leave his ambitious theorizing in an aggressive manner for those who are interested in getting smart, not making money.