Canvas is cool! - page 4

New comment
 
George Merts:

The demonstration of possibilities is very impressive indeed. But where to apply them is unclear to me.

And the author doesn't use it in that way in his own place. In the very beginning he explained:"In order to make people start using Canvas for useful things you have to start from demonstrating useless things. :))"
 

As the saying goes: Feel the difference with a real example:

  • Visualisation is particularly important during the design and strategy creation phase, especially when complex algorithms are involved. The language of numbers is incomprehensible to humans. With a proper visualization of the processes, it is easier for the programmer to find errors in the algorithm and to come up with the right ideas for improving the algorithm.
  • An example of a 4-dimensional space, in which the price moves: Time, price, ratio of bulls and bears, the real volumes. To simplify visualization, it is enough to simplify this model into a three-dimensional space Price - bull-bear-ratio - volumes. Each tick forms a point in such a space, and a line is drawn, consistently connecting all the points (ticks), and this sequence will essentially be the 4th dimension of Time. Clouds in the form of channels and three-dimensional spirals will be formed in this way. The informative and predictive power of such a visualisation is far superior to the classic 2-dimensional Price-Time visualisation. This is just one of an infinite number of examples that your imagination can provide.
  • It is a mistake to think that it takes an N-dimensional array to visualise price movement in N-dimensional space. Any dimension requires only a 2-dimensional array, the first dimension being the number of ticks or bars. And the second is the number of measurements (coordinates). I.e. for each tick (bar) there are N values (characteristics). In the example above, each tick will contain 4 parameters: time, price, current ratio of bulls to bears and real total volumes, forming therefore a 2-dimensional array Tick[n][4], where n is the number of ticks (bars). Or we may do with four one-dimensional arrays. I.e., if we compare price movement in 4-dimensionality vs. classic 2-dimensionality, we need only twice as much memory, not a cube of times.
 
If I move in space, it's 4 dimensions. Time and position in space.
If I consider this displacement relative to another process, there are more axes. But this "dimensionality" refers to the system in question, it is a characteristic of it, not a dimensionality in the general sense, not a space-time characteristic. The process existed within space-time and did not go beyond it. I think it is the dynamic N-dimensional arrays that are more suitable for the accuracy of the statement. Although I agree that an array is only suitable for a single process too. If we consider the notion of indexes, they too are constructed in a multidimensional system, it's a bit different, still the multidimensionality is not spatial. That's it, brains are broken - goal achieved. Capitulir.

I wonder who thinks about this.
 
Evgeniy Zhdan:
That's a cool rainbow schedule! I'd also like to make the candles dance to the music and twirl around each other!
Ahaha.
It's going to be a disco.
 
George Merts:

For the AI alone, all these beautiful things are not worth a damn.

In addition, there is a diametrically opposite point of view, that all complex sophisticated systems are very unstable, and it is impossible to make money with them. At the same time the simplest TS of crossing the price and sliding bar type, works quite well and has a better stability.

I have tested it. Without any manual adjustments they work.

 
Nikolai Semko:

As they say: Feel the difference with a real example:

  • Example of a 4-dimensional space in which price moves: Time, price, bull-bear ratio, real volumes. To simplify visualization, it is enough to simplify this model into a three-dimensional space Price - bull to bear ratio - volumes. Each tick forms a point in such a space, and a line is drawn, consistently connecting all the points (ticks), and this sequence will essentially be the 4th dimension of Time. Clouds in the form of channels and three-dimensional spirals will be formed in this way. The informative and predictive power of such visualisation is far superior to the classical 2-dimensional Price-Time visualisation. This is just one of an infinite number of examples that your imagination can provide.

The problem is that these are superfluous entities that have little to do with the imperfections of the market that are used in trading.

Indeed, we can come up with ten "dimensions" in which price moves - but will it bring us any closer to greater profits ?

The picture with coloured channels, directions also looks very impressive, but again - what is the point? What new information can a person get from looking at these beautiful colours?

It reminds me something of a neuronet, which is easily capable of extracting regularities but has no common sense, as a result of which found regularities can both really exist and be a "statistical artefact".

 
Alexey Volchanskiy:

Alas this is not the case, I have checked. Without manual adjustments they pour.

Strange, but my results are different. There are those who are pouring and those who are earning. And any TS has periods of profit and periods of loss.

Now I have more than 200 TS, and the only question is how to select the most stable.

The more stable TS is, the less degrees of freedom it has. Therefore, all these extra entities - "ten-dimensional price movement spaces" - work against us, reducing the stability of the TS.

 
George Merts:

The problem is that these are superfluous entities that have little to do with the imperfections of the market that are used in trading.

Indeed, we can come up with ten "dimensions" in which price moves - but will it bring us any closer to greater profits?

The picture with coloured channels, directions also looks very impressive, but again - what is the point? What new information can a person get from looking at these beautiful colors?

It reminds me something of a neuronet, which is easily capable of extracting regularities, but at the same time has no common sense, as a result of which the found regularities can both really exist and be a "statistical artefact".

What can one do...

Some people see only mud in a puddle, while others see reflection of stars. But it is understandable, because seeing the outer world is only a reflection of the inner one.

It brings to mind the cartoon"Oh and Ah".

What a load of pessimism, Georg. It's quite clear where the extra entities you see are prescribed. Personally, I don't notice them.
 
George Merts:

Strangely, my results are different. There are those that pour in, there are those that make money. And any TS has periods of profit and periods of loss.

Now I have cut more than 200 TS, and the question is only in the method of selecting the most stable ones.

And the more stable TS is, the fewer degrees of freedom it has. Therefore all these extra entities - "ten-dimensional spaces of price movement" - work against us, reducing stability of TS.

Congratulations, you've finally come to the realization that it's just a matter of methodology to select primitives!

Even a primitive strategy on one or two wagons will be super effective if every, minute, hour, day or even every tick,correctly change the period of this (or those) wagons.

It remains a small thing - to develop this correct algorithm (methodology) of selection. :)))

And that, Georg, is artificial intelligence.

 
ILNUR777:
If I am moving in space, there are 4 dimensions. Time and position in space.
If I consider this displacement relative to another process, there are more axes. But this "dimensionality" refers to the system in question, it is a characteristic of it, not a dimensionality in the general sense, not a space-time characteristic. The process existed within space-time and did not go beyond it. I think it is the dynamic N-dimensional arrays that are more suitable for the accuracy of the statement. Although I agree that an array is only suitable for a single process too. If we consider the notion of indexes, they too are constructed in a multidimensional system, it's a bit different, still the multidimensionality is not spatial. That's it, brains are broken - goal achieved. Capitulir.

I wonder who thinks about it.

There's nothing to think about. It's a piece of cake.

Any point in N-dimensional space has N coordinates. And if you have X such points then you need N*X values.

Any set of points in N-dimensional space can be projected onto a lesser dimensional space, including a flat two-dimensional screen. Our brain is programmed to interpret 4-dimensional space (the space-time continuum), so it would be reasonable to use 3D monitors, but even if we don't have a 3D monitor, we can simply add some rotation (a slight rocking may be) and let our brain build a 3D picture by itself due to different angular velocity of objects at different distances.

There is no need to use N-dimensional arrays in N-dimensional spaces. In fact, even the Creator of our World, which we perceive as 4D, does not use multidimensional arrays. Just look at the observer effect in quantum physics.

Even Einstein wondered, "Does the moon exist when I am not looking at it?" That is, the point is, what is the point of displaying information on a screen if there is no observer.

1234567891011...95
New comment
Reason: