Random wandering - page 52

[[Deleted]]

Clearly illustrated here and below.

[[Deleted]]

secret #:
Rename the branch to "Sewerage" already)

Another lunatic.

On the pole!

[[Deleted]]

Forum on trading, automated trading systems and trading strategy testing

Random Wandering :

Oleg avtomat, 2018.11.04 08:29

.

And here is the SB with Laplace distribution.

The graphs show : SB, method, phase portrait, ACF, result, MO, variance.

But it is unlikely that the screamers here understand at least half of what is presented in these graphs.

[Maxim Kuznetsov]

Autumn, the rooks have gone, the woodpeckers have arrived :-)

Now they'll be chattering about the beaten coin until the rains come down.

[vladavd]

Олег avtomat #:

On the other hand, I can see that you, too, are stuck on false dogma. And you flat out deny even the possibility. And this fully indicates your incompetence in the matter.

No, I'm not. My practice has confirmed the theory. I understand from the theory why it is impossible to earn on SB and, in principle, one could stop and not dig further, but I purposely checked it on different random series, what I got results on the market. The result is logical: what works on the market, does not show anything at all on OOS. So far I haven't encountered any convincing arguments or empiricism (just not curvafitting, as A_K has, of course) in favor of SB earning possibilities, including in your thread. By the way, I wouldn't be surprised if your gcc produces not such a random series and that is what misleads you and gives the impression of predictive capability, are you sure that matcad has seriously considered the stability of its generator, have you checked it? This is a serious question, otherwise you may not be dealing with what you think you are dealing with. Generally speaking, what you claim is nonsense:the CGSPF is required to be impossible to predict its values, this requirement is satisfied (such algorithms exist), but you say that prediction is possible, so you are dealing with not quite random series, otherwise it is incorrect to say that robust generators exist. But if you look into the real world, it turns out that the data is safe, and no one has demonstrated earnings on SB yet.

[[Deleted]]

vladavd #:

No, my practice has confirmed the theory. I understand from the theory why it is impossible to make money on SB, and in principle I could stop and not dig further, but I purposely checked on different random series what my results show on the market. The result is logical: what works on the market, does not show anything at all on OOS. So far I haven't encountered any convincing arguments or empiricism (just not curvafitting, as A_K has, of course) in favor of SB earning possibilities, including in your thread. By the way, I wouldn't be surprised if your gcc produces not such a random series and that is what misleads you and creates an impression of predictive capability, are you sure Matcad has seriously considered the stability of its oscillator, have you checked it? This is a serious question, otherwise you may not be dealing with what you think you are dealing with. Generally speaking, what you claim is nonsense:the SGSPF is required to be impossible to predict its values, this requirement is satisfied (such algorithms exist), but you say that it is possible to predict, so you are dealing with not quite random series, otherwise it is incorrect to say that robust generators exist. But if you look into the real world, it turns out that the data is safe, and no one has demonstrated earnings on SB yet.

Your mistake is that you don't see the difference between GSF and SB. And these are very different entities.

[[Deleted]]

Maxim Kuznetsov #:

Autumn, the rooks have gone, the woodpeckers have arrived :-)

Now they'll be chomping at the bit until the rains come down.

Another parrot has flown in.

Onthe pole!

[Maxim Kuznetsov]

apr73 #:

OK, show a graph on which for 10,000 (ten thousand ) throws the line has gone further than 250 on the Y-axis.

You don't need to predict the SB graph, you need to use statistical properties of the sequence.

While the woodpecker is chiseling away at the coin, I wrote a short program, this is actually what you wanted :

The CSV you wanted is in the trailer, the program is here (you can't attach the C-sheet, so it's like this):

#include <stdio.h>
#include <stdlib.h>

#define  TOTAL 10000
#define  MAXDEV 250

long Pass(long *result,int total)
{
        long p=0;
        long dev=0;
        result[0]=p;
        for(int i=1;i<total;i++) {
                if (rand()>RAND_MAX/2.0) {
                        p++;
                } else {
                        p--;
                }
                result[i]=p;
                if (p>0 && p>dev) dev=p;
                else if (p<0 && p<-dev) dev=-p;
        }
        return dev;
}

int main(int argc, char **argv)
{
        long *data=(long *)malloc(sizeof(long)*TOTAL);
        unsigned int pass=1;
        while(1) {
                long dev=Pass(data,TOTAL);
                if (dev>MAXDEV*2 || dev<=-MAXDEV*2) printf("pass %d, max deviance=%ld, last=%ld\n",pass,dev,data[TOTAL-1]);
                if (data[TOTAL-1]>=MAXDEV || data[TOTAL-1]<=-MAXDEV) {
                        printf("Final data[last]=%ld",data[TOTAL-1]);
                        break;
                }
                pass++;
        }
        FILE *f=fopen("randomwalk_10000_250.csv","w");
        if (f!=NULL) {
                for(int i=0;i<TOTAL;i++) {
                        fprintf(f,"%d ; %ld\n",i,data[i]);
                }
                fclose(f);
        }
        return 0;
}

[Uladzimir Izerski]

Check. No one to do business with))

[pribludilsa]

Олег avtomat #:
Who shouted about being prepared to pay $50,000 for a demonstration of making money on SB?

Willing to pay for a prediction, not a grid, you need to read carefully.