Probability. - page 10
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
If we rely on the calculated points, these are rather zigzag knees or essentially waves of some order.
I.e. it is essentially a zig-zag in percentage terms, and these movements are somehow filtered, you can get a "cloud" of data on M-1.
If you try to combine these technics, you may use the Boryspolz channels strategy, there's analysis at H-4, H-6, the standard deviations channel is not very informative (applicable) in trading, look at the USD-JPY pair flat as an example (about 4), the other point is to predict (analyze) with linear tools
Non-linear formations even in terms of probability, I think this is not the right tool to use. Try to describe a snowflake with a ruler, the nature of its formation by temperature + condensation and material (dusting as a reference point) of snowflake formation (growth).
Go for it comrade!
This is the correct way to do it.
It's a pity the first posts and the last ones are on the same page. Nobody understands anything.
I've got an indicator somewhere, lying loose in the forecast branch.
I think it's from, like, '14.
Wizard's been sniping at it, but it's there.
I remember saying at the time that it wasn't tied to a TF.
I'll try to remember the formula now....
Something like:
probability = (high + low)/(2*current price).
it's roughly... some of the logic...
It is similar to buy and sell probabilities, i.e. we get 2 formulas
Sell probability = current price / high
Buy probability = Hai / currentprice
Multiplied by 100, it's a percentage.
The end result is cool systems, of course, but not so much...I.e. it's essentially a zig-zag in percentage terms, and these movements are somehow filtered out, you can get "tons" of data on M-1.
If you try to combine these technics, you may use the Boryspolz channels strategy, there's analysis on H-4, H-6, the standard deviations channel D-1 is not very informative (applicable) in trading, look at the USD-JPY pair flat as an example (about 4), it's another moment to forecast (analyze) with linear tools
Non-linear formations even in terms of probability, I think this is not the right tool to use. Try to describe a snowflake with a ruler, the nature of its formation by temperature + condensation and material (dusting as a reference point) of snowflake formation (growth).
Greetings, gentlemen traders and programmers. I have an opportunity to continue our conversation.
I did not understand anything of theVeniamin Skrepkov's post:
Let's start from the beginning.
Let's assume that one ZZ knee is a wave.
Will it be convenient for you?
From the previous picture we can draw the following conclusion:
At the moment there is a corrective wave upwards.
There is only a 28% probability of further upside. At the moment (not unimportant).
It may be clearer that way.
The probability should be calculated based on statistics and not based on some assumptions, algorithms. Let us take, for example, a certain set-up (a candlestick pattern) and check it on the history. If it is encountered 10,000 times and out of them it can bring profit 6,125 times, the probability of its occurrence will be 61.25%. Another example: GEP with the market opening on Monday. Again, if the GEP was closed 6,473 times out of 10,000, we will see a 64.73% probability of its occurrence.
I.e. to calculate the probability, we should have a sample and its history (statistics). This does not mean that the homemade approach will not work, it is quite possible it will. But it has nothing to do with probability.
The probability should be calculated on the basis of statistics, not on the basis of some assumptions or algorithms. Let's take a set-up (a candlestick pattern) and check it on the history. If it is encountered 10,000 times and out of them it can bring profit 6,125 times, the probability of its occurrence will be 61.25%. Another example: GEP with the market opening on Monday. Again, if the GEP was closed 6,473 times out of 10,000, we will see a 64.73% probability of its occurrence.
I.e. to calculate the probability, we should have a sample and its history (statistics). This does not mean that the homemade approach will not work, it is quite possible it will. But it has nothing to do with probability.
Yes, statistics is needed, if you go into it thoroughly.
But here, no matter what statistics have been collected, it will not work much, because it was made on the trading history. But the history is dead...
More or less neural network. Fitting trade to history in short, analog of stat-analysis. And it's still nonsense.
Trailer.
Yes, statistics are necessary if you get into it thoroughly.
But here, whatever statistics are collected, they will not work particularly attractively, because they are made on trading history. But the history is dead...
More or less neural network. Fitting trade to history in short, analog of stat-analysis. And it's nonsense anyway.
In the trailer.
See you in the morning. Let's do it in the morning. Tired today)))
You can just congratulate me on the birth of my granddaughter.
See you in the morning. Let's do it in the morning. Tired today)))
You can just congratulate your granddaughter.
See you in the morning. Let's do it in the morning. Tired today)))
You can just congratulate me on the birth of my granddaughter.
The probability should be calculated on the basis of statistics, not on the basis of some assumptions or algorithms. Let's take a set-up (a candlestick pattern) and check it on the history. If it is encountered 10,000 times and out of them it can bring profit 6,125 times, the probability of its occurrence will be 61.25%. Another example: GEP with the market opening on Monday. Again, if the GEP was closed 6,473 times out of 10,000, we will see a 64.73% probability of its occurrence.
I.e. to calculate the probability, we should have a sample and its history (statistics). This does not mean that the homemade approach will not work, it is quite possible it will. But it has nothing to do with probability.
Although you can not completely ignore the history, but such statistics - statistics are mostly about nothing, and most likely it will not work. The same as systems after optimization work successfully only on the time-series on which they were optimized.
Working stats need to be collected and evaluated as the play progresses, similar to how probability of winning in poker and other games is evaluated. Past game stats are little or no information that anyone needs anymore. The basic statistics are the here and now.
Greetings to all gentlemen traders and programmers. There is an opportunity to continue the discussion.
I don't really understand anything ofVeniamin Skrepkov's post:
Let's start from the beginning.
Let's assume that one ZZ knee is a wave.
Will it be convenient for you?
I agree that it is a wave. The question may be about filtering waves of 5-7 pips - is this noise regulated in some way ? As for the channels, there was an exchange of views and I decided to take part too.
The snowflake as an image of the model, i.e. its constituents are the components of market movements - 1) volume (condensate) 2) change of market participants about price (temperature) 3) formation point . As soon as a process is defined, an understanding emerges.