Testing real-time forecasting systems - page 61
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
There is a subtle philosophy here. Entropy minima cannot be chosen as a criterion, simply that zero entropy is an area with zero probability of price in that area, in other words, it makes no sense to expect price where it cannot be, one must wait where its probability is highest. There's another subtlety, I think you're looking at informational entropy as physical entropy, and these are slightly different things. :о)
I agree that the minimum should not be taken. But the maximum is not very suitable either, imho. In order not to confuse the concepts of entropy I propose to consider it always as a result of box-counting (moreover, most people here seem to consider it so), i.e. a measure of dispersion of a random variable in space (space dimension is a constant for the same model, within which we form several realizations - each has a different entropy). In this formulation (correct me if I'm wrong), I don't see much logic in using entropy as an indicator for selecting one implementation from the whole heap.
Is it to be understood that the formula defining the forecast horizon by entropy is inapplicable in our case? Why do pundits cite it exactly in the context of predicting chaotic processes?
I expect a precipitous collapse of the pound, by at least 600 pips, the nearest support level is 1.6
Alex, you have your own thread where you were going to demonstrate your method in online trading. Why don't you post your predictions there as you have done so far? Too many opponents? You should be used to it by now.
Please, use your own thread for your predictions and discussion of your trades. Not because your ups and downs predictions are not coming true, but because all your opponents will follow you here and this (and any other) thread will turn into a mess just like the other one. And like the one that was torn down. And like all the others that came before. Don't start all this here again.
I understood that Yuri asked for a picture of another forecast (I was playing thimbles there :o))
I asked for a picture of the forecast you posted just before you left.
I don't have all of them, just the main trajectories and it still hangs in the terminal (I was about to delete it). Here is the picture. It's clear enough. It may be enough.
I didn't understand what was the forecast on your chart - grasn or yours ?
I asked for a picture of the forecast you posted just before you left.
I did not understand only whose forecast it was on your chart, grasn 'or yours ?
Yes, grasn, as requested. I've written an induke to output the predicted time-series to the terminal. Earlier in this thread, you can download both the inductor and the csv file for it from grasn.
to marketeer, Yurixx
Согласен, что минимум брать нельзя. Но и максимум не очень подходит, имхо.
That's the subtlety of it, a true realisation will tend towards maximum entropy, but not necessarily that it will always be maximum. I hinted earlier "It's simple, a deterministic signal carries no new information, only stochasticity carries new information. The question is how much "new" to expect." I have a strong suspicion that the level of entropy to which the actual realization will "stick" is somehow correlated with the "cyclicality" and "quantity" (but not quality) of expected news, but this is only a guess. (imho) Anyone have any thoughts on this thought? :о))))
to marketeer
Not to confuse the concepts of entropy I propose to consider it always obtained as a result of box-counting (especially since most people here seem to think so),
It's easy and I don't think that the method of calculation plays any principal role. One only has to remember that we are talking about informational entropy, not thermodynamic or any other entropy.
i.e. a measure of scatter of a random variable over space (dimension of space is a constant for the same model within which we form several realizations - each has a different entropy).
it has nothing to do with the dimensionality of space. Besides, dimensionality is just an input parameter to the model.
In this formulation (correct me if I'm wrong) I don't see any special logic in using entropy as an indicator for choosing one realization from the whole heap.
I don't think it's right or I'm missing something.
Is it to be understood that the formula of determining the forecast horizon by entropy is not applicable in our case? Why do men of science bring it exactly in a context of forecasting of chaotic processes?
Why would you get it wrong? This formula will give the lowest possible prediction horizon for a very complex system, of large dimensionality. Entropy itself is the sum p*ln(p) of independent random events, the value of which will probably depend on the number of summed components. I think so.
I have a strong suspicion that the level of entropy to which the actual implementation will "stick" is somehow correlated with the "cyclicality" and "quantity" (but not quality) of expected news, but that's just a guess. (imho) Who has any thoughts on this thought? :о))))
I agree, the greater importance is the significance of news, but to introduce the significance as a parameter, I think, we shouldn't. The significant news is played by the whole market, the quotes behaviour in the run-up to the significant news changes, and of course we can't miss it on the indicator, that gives something new at least. If you don't notice it (imho), then what is the value of the method?
Good to see colleague :o) As for the avatar and the quote - it's my favourite cartoon ever. I, for example, also like this one:
-Professor, where were you on the 15th evening?
- (rubbing the back of his head) - And where am I now?
PS: Well, since you're the first to start, ... just curious, why do you respectable bucket on his head?
Where do you think the leaky bucket should be? I think Futurama is the most masterful cartoon ever made.
I agree that the greater importance is the significance of news, but I think we should not introduce the significance as a parameter, because the market reacts to the news, the quotes behavior in the run-up to the significant news changes beforehand and of course it is impossible not to notice it in an indicator, which at least gives something new. If it is invisible (imho), then what's the point of the method?
It's a little bit different. It's about choosing different levels of entropy at different times as a likelihood criterion, rather than the maximum value all the time. These levels can be found by experience (or by the same way to understand that they do not exist), from the assumption of correlation of news frequencies (perhaps additionally introducing clustering by news types/categories)
PS: I don't use indicators for a few years now.
Where do you think the hole bucket should be? I think Futurama's the most masterful cartoon, similarly.
How are we supposed to argue about this? You're right, it's the best place for a bucket. Wear it, plus it makes you look more masculine...
:о)))