Stereo Neuro Net - 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
When engaging in this dialogue, we, subconsciously, solve different optimization problems (in the global sense). About what approach you have chosen I can only guess. About mine I can say that at this stage of research I have enough computational power at my disposal not to limit myself by parameter "complexity of NS training". Obviously, there is no harm in retraining (additional training) of NS at each step. Thus, I can concentrate my attention on other interesting aspects of AI by lowering the dimensionality of the parameter space in the investigated domain by one. I think, in that sense, I am doing optimally.
I looked through the cartoon again, and a profound question arose:
why doesn't the NS define class 3 - FLET ?
Oops, budimir, does such a class really exist? Give me its definition, please.
Oops, budimir, does such a class really exist? Give me its definition, please.
I'll give you the definition of the FLET class: since the classes BUY and SELL exist, so does the FLET class !
I define the FLET class: if the classes BUY and SELL exist, so does the FLET class !
why not?
I define the FLET class: if there are BUY and SELL classes, then there is also a FLET class !
That's a clever way of defining it. As I understand it, we are talking about an "out of the market" position?
Yes, but there is nothing tricky here, for example:
1.Many traders invent "digital" muwings (quite successfully), where the trend is separated from the flat state
2.Members of CHAMPA-2008 invent their own EAs for the trend or for the flat, and they themselves write about it in their interviews.
да,но хитрого тут ничего нет,например:
Oops, budimir, does such a class really exist? Give me its definition, please.
Hello, Alexey!
I support you in your request.
Obviously, the category "trend-flat" is strictly limited to a particular TS, it is a purely subjective evaluation and cannot in any way characterize the process or the market status. Indeed, look at Fig.
For TS working out targets up to 10 pips, this time series is a trend one (we open a position in the direction of price movement and hold it until the trend changes). But for TS with 50 pips and above, this is a pure "flat" and the strategy is obvious - trade within the corridor. So, the "finesse" has its place here.
After all said above, budimir, formulate your question once again, please.
Hello, Alexei!
I'll support you in your request.
It is obvious that the category "trend-flat" is rigidly tied to a particular TS, being a purely subjective assessment and cannot in any way characterize the process or the market condition. Indeed, look at Fig.
For TS working out targets up to 10 pips, this time series is a trend one (we open a position in the direction of price movement and hold it until the trend changes). But for TS with 50 pips and above, this is a pure "flat" and the strategy is obvious - trade within the corridor. So, the "finesse" has its place here.
After all of this, budimir, formulate your question once again, please.
You have contradicted yourself:
For a TS working up to 10 pips, this timeframe is a trend one
What kind of trend is this - a 10-point one ? i.e. the trend = 4...5 spreads(!?),
If the market bites, the trend will disappear, just like the depo will disappear!
Your image is beautiful, but it is a pure sine wave,
since when did the market become stationary?
Maybe you are right.
Then the correct way would be: let the TS handle market movements with a characteristic amplitude of about 10 pips. Then, for such an TS, the given market dynamics will be a trend. And vice versa, let the TS process the market movements with a characteristic amplitude of about 100 points. Then the resulting market dynamics for this TS will be pullback (flat).
Now is it correct?
If we speak about categories (trend-flat), the correspondence or nonconformity of the model (sin()) to the real market process in minor details is not essential, which facilitates understanding of the subject.
Yes, but there is nothing tricky here, for example:
1.Many traders invent "digital" muwings (quite successfully), where the trend is separated from the flat state
2.The members of CHAMPA-2008 specifically invent their EAs for the trend or for the flat, and they themselves write about it in their interviews.
They all like to write that there is nothing tricky, the simple truth of life...
1. For some reason none of the most cunning digital filters has ever learned to effectively separate states in which it works well from those in which it works poorly. Of course, it's not the filter itself that works, but the strategy based on it, and we should keep that in mind.
2. It's not a big deal what they write. Some participants are just very lucky that the market is so rabid right now and is actually good for both pipsetting and trend strategies at the same time. However, I still haven't seen a sensible criterion that effectively (= profitably) separates its "profitable" states from its "unprofitable" states even in a given TS.