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
Where did the postulate that the price process is non-Markovian come from in the first place? It is the condition of a Markovian process that the future is independent of the past and the present is fixed...
well from the heavy tails, like not random but with memory
well from the heavy tails, like not random but with memory
But VisSim will not be able to get it out now.)
Well, there are VisSim to C code converters... ))
A Markovian process may have memory and internal states, so it shouldn't be a problem...
Well, there are just states and transition probabilities, in each new state only the current probability will influence the transition, and the memory of the previous step is erased
And I don't understand much about the description of non-Markovian, it's rather complicated there)
Well, there are just states and transition probabilities, in each new state only the current probability will affect the transition, and the memory of the previous step is erased
That's right, the memory is put into the current state, so the previous state is no longer needed, although it is always there if needed for calculations....
Regarding future transition of course it is all the same probability, but even if not, it can be taken into account in transition metrics...
That's right, memory is put into the current state, so the previous state is no longer needed, although it's always there if you need it for calculations....
Regarding future transition of course everything is equally probable, but even if not, it can be taken into account in transition metrics...
well, they do all sorts of temporal difference and stochastic models, but i'm not very good at it and am just beginning to learn
For example, you can look at q-learning in machine learning, there are both stationary and non-stationary models on the temporal difference, t-tn, about which Alexander wrote, but the approach from the other side. And the hardest part is to apply them to continuous processes like markets, with discrete ones everything is more or less clear.
For example, you can look at q-learning in machine learning, there are both stationary and non-stationary models on the time difference, t-tn, which Alexander wrote about, but the approach from the other side
I think discrete hidden Markov models and algorithms are more relevant to the problem because you don't need to know the model itself, which makes it similar to neural networks...
The equation of diffusion and Brownian motion looks very far-fetched... The market is obviously far from Brownian))
I think discrete hidden Markov models and algorithms are more relevant to the problem because you don't need to know the model itself, which makes it similar to neural networks...
The equation of diffusion and Brownian motion seems like a bit of a stretch... The market is obviously far from Brownian))
Well in general it takes a lot of work and... think :) it's mine, so to speak.
You'll have to forgive me. But don't you think you're overdoing it here. It seems there's a competition going on to see who's smarter than the other. Here's the easiest tip. You go in with a lot that suits you, you make 10 points in a deal, close half of it, and the rest goes in the bank. And you will be happy. And no headaches)))
You'll have to forgive me. But don't you think you're overdoing it here. It seems there's a competition going on to see who's smarter than the other. Here's the easiest tip. You go in with a lot that suits you, you make 10 points in a deal, close half of it, and the rest goes in the bank. And you will be happy. And no headaches)))
The creative process, you know...