Whether there is a process whose analysis of one part does not allow predicting the next part. - page 2
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
Air temperature graph.
1. I didn't write about this - I wrote about the opposite, that if you know a priori that the series is not a representation of some process, then it is illogical to predict it.
2. And what the author wants is a GSF with an arbitrary variable distribution.
1. What is the point of looking at a series that has nothing to do with the process? - We are talking about studying and predicting a hypothetical process and a numerical series belonging to it.
2. RNG is predictable - knowing the previous segment one can predict the value at the end of the next segment with a certain accuracy (numbers cannot stay long enough near the same value - otherwise it won't be an NF series anymore).
I didn't write about that - I wrote about the opposite, that if you know a priori that a series is not a representation of some process, then it is illogical to predict it. And what the author wants is a RNG with an arbitrary variable distribution.
Any series is a representation of some process, at least the process of creating that series.
Hi.
I suggest that the esteemed community come up with a process that cannot be predicted (so that no money can be made on this prediction). At the same time, the process should not have stationary stat-characteristics over time.
We toss a coin, heads tick up, tails tick down... Instead of a coin MathRand()%2.
We toss a coin, heads tick up, tails tick down... Instead of a coin, MathRand()%2.
Tried. Used a grid with one hidden layer. I managed to achieve positive MO - I predicted the direction of the next increment. So your variant is no good.
->
2. The LFO, as well as the MF, is predictive - knowing the previous segment you can predict the value at the end of the next segment with a certain degree of accuracy (the numbers cannot stagnate long enough around one value - otherwise it will no longer be adjacent to the MF).
Tried. Used a grid with one hidden layer. Was able to achieve a positive MO - predicted the direction of the next increment. So your version is no good.
->
My point is that it is possible to predict any process with positive MO.
But still, I suspect there is only one situation where prediction is impossible - when the process itself "knows" about the observer and its previous predictions, i.e. these are forward and backward-coupled systems. In this case it is impossible to predict without adjusting the predictor over time, and that is if the lag in the adjustment is sufficiently small (and there is always a lag - the discreteness in the observations can't be removed).
If the process knows that the observer knows that the process knows?
If the observer knows that the process knows that the observer knows that the process knows that the observer knows that the process knows that the observer knows that the process knows that the observer knows that the process knows... Who wins?
If the observer knows that the process knows about the observer?
If the process knows that the observer knows that the process knows?
If the observer knows that the process knows that the observer knows that the process knows that the observer knows that the process knows that the observer knows that the process knows that the observer knows that the process knows... Who wins?