From theory to practice - page 872
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 your system works better when it is only trending or flat - it does not and will not work a priori. the probability will be less than 50% initially, and you will only have to increase it at the expense of risk.
no
I'm just testing it with 1 pip increments and 0.2 lot at 50 quid depo... )))
well if the increments are 10 or 100 it's pretty cool
one order is also possible.
I didn't do it that way but in principle it is possible... with a virtual second order most probably
If you look at the whole picture, from the start of a trading session to the profit taking, you get a system that works for both trends and a flat at the same time.
I don't.
I was just testing it with 1 pip increments and 0.2 lot with 50 quid deposit... )))
well if the increments are 10 or 100 then that would be pretty cool
What worries me now is this... there's Bayes' theorem.
Let's take the MA for example.
let's take a signal of MA for [i-1] and Close[i] and check if the forecast corresponds to the result.
i.e. use Bayesian theorem to check how justified the forecast based on MA, for example, "If the MA goes up, the price should go up".
Of course, we do understand that most often it will be 50/50. Take 194 MAs from period 6 to period 200, check the degree of their predictability in TF M1 and deduce which MAs at the current time provide the best predictive ability. And if there is a "cascade", i.e. several optimal ones for prediction, then try to trade.
Do you think it's just a test of Bayes' theorem or not...
No.
I was just testing it in a tight frame with 1 pip increments and 0.2 lot at 50 quid depo... )))
well if the increments are 10 or 100 then that would be pretty cool
Look what else I noticed in my strategy:
you see the drawdowns marked in red?
So I have a tricky coefficient that regulates the ratio (profit/loss)
The essence of it is to minimize the presence of the order in the market with minimal losses.
It turns out that the time an order is in the market directly affects the losses and the risks.
This means that the price is mostly random, as a random process only depends on the time.
Look what else funny thing I noticed in my strategy:
you see the drawdowns marked in red?
So I have a tricky coefficient that regulates the ratio (profit/loss)
his essence with minimal losses to minimize the presence of the order in the market.
It turns out that the time an order is in the market directly affects the losses and the risks.
This means that the price is mostly random, as a random process depends only on the time.
the price is random for a trader with a sufficiently risky market order for no more than 8 minutes
with less risk, it is possible that the price is random at all times
What worries me now is this... there's Bayes' theorem.
Let's take the MA for example.
let's take a signal of MA for [i-1] and Close[i] and check if the forecast corresponds to the result.
i.e. use Bayesian theorem to check how justified the MA prediction is, for example, "If the MA goes up, the price should go up".
Of course, we do understand that most often it will be 50/50. Take 194 MAs from period 6 to period 200, check the degree of their predictability in TF M1 and deduce which MAs at the current time provide the best predictive ability. And if there is a "cascade", i.e. several optimal ones for prediction, try to trade.
Do you think it's as a test of Bayes' theorem whether it really works or not...
I used to write the same kind of code...
but it suddenly dawned on me - a person cannot make an adaptive quote that intricate and start searching for a basic calculation
one of the 2 articles above discusses the market approach to deciphering price movements from one level to another (or vice versa), but it's too complicated
everything is simpler, I mean the decoding of pseudo-volatility artificially introduced by a price-quote.
by the way, if you remove this fiction at once, you get a bounce, i.e. a linear trend
the other article is about how a triangle is quoted, so trading more than one side of it is highly discouraged, a hassleI remember my robot in . I wrote 15000 lines of code and it took me more than three months to write - imagine how excited I was XD
MA is a spread printing machine, the best assistant to a trader, but not to a quitter.
No no you misunderstood me, Bayes theorem has "a priori" and "posterior" probability and the point is that his formula gives a relationship between one and the other. so, I wanted to check this relationship on MA and see if anything comes out). Well not MA so I have a model of a flat trend ... in general to this mechanism can be screwed everything you want).
The main thing is that "past-present-future" relation will work and, the most important thing, there will be an adequate probability of assessing the real situation on its basis.No no you misunderstood me, Bayes theorem has "a priori" and "posterior" probability and the point is that his formula gives a relationship between one and the other. so, I wanted to check this relationship on MA and see if anything comes out). Well not MA so I have a model of a flat trend... in general, to this mechanism can be screwed everything you want).
The main thing is that the "past-present-future" connection will work and the most important thing is that it will give an adequate probability to estimate the real state of affairs on its basis.Yeah, but not MA
In fact, if such a connection is found, then why bother with the computer?
I do not want to trade ;)
the relationship exists, but again it has two options - up or down, the probability of prediction just increasesyes, but not the MA
Actually, if such a connection is found, why else torture the computer?
I don't want to trade ;)
ahah crafty))
i think so too but the point is the correct estimation of this relationship should give 50/50 most of the time. right? by the way did you check your theory about the order and 8 minutes?
why i'm asking, because my statistical analysis showed that the higher the TF, the more random the price is within the movements on that TF.
I just know it's true. But the spread of a random move on DAY is e.g. EURUSD about 2000 pips while on M1 it is 25-50 pips.