From theory to practice - page 856

[Evgeniy Chumakov]

I'm tormenting all sorts of things on the minutes at the moment.

Could be a coincidence, but the most losing trade in the test period was liquidated using the Speed < Kurtosis condition, where Speed = (Absolute sum of increments)/t and Kurtosis = t * (M4/MathPow(M2,2))

This is where the non-parametric kurtosis failed.

[Alexander_K]

Renat Akhtyamov:

In addition, testing shows that the doll is virtually harmless from H4 and above

Agreed. Working with a time window lower than H4 is a tantrum for the nervous ones or even geniuses, as practice shows.

But I would not recommend to climb to the daily chart either - it would be boring, and the trader needs the cash here and now.

There should be an optimal balance between the number of received ticks and the time window - this is also one of the keys to the Grail.

[Alexander_K]

Evgeniy Chumakov:

I'm tormenting all sorts of things on the minutes at the moment.

Could be a coincidence, but the most losing trade in the test period was liquidated using the Speed < Kurtosis condition, where Speed = (Absolute sum of increments)/t and Kurtosis = t * (M4/MathPow(M2,2))

It's where the non-parametric excesses failed.

Are the dimensions the same? Is there any literature on the formula for kurtosis?

If not, we have to rely on God's gift. This is for creative people only. Well, you understand.

[Evgeniy Chumakov]

Alexander_K:

Is there any literature on the formula for kurtosis?

https://www.macroption.com/kurtosis-formula/

Alexander_K:

Are the dimensions the same?

They are the same.

Another thing I noticed from the test is that if the current value of the non-parametric kurtosis > the value at the beginning of the sliding window (shift), then there is more profitability.

[Evgeniy Chumakov]

I would start by calculating the average excess per pair (as mentioned above) to see if there is any correlation.

[Renat Akhtyamov]

Alexander_K:

I agree. I agree. Working with a time window below H4 is for the nervous ones or even geniuses, as practice shows.

But I would not advise to buy on the daily market either - it would be boring, and the trader needs the cash here and now.

There must be an optimal balance between the number of received ticks and the time window - this is also one of the keys to the Grail.

You're missing the clarification: " ... on a demo account"

because it is impossible to gain profit with ticks on the real account.

So the demo graal, or rather the demo TS, is a kind of pilot project.

[Evgeniy Chumakov]

Renat Akhtyamov:

there is a clear lack of clarification here: "... on a demo account"

For it would be impossible to work in the black on a real account with ticks.

So the demo graal, or rather demo TS, is something.

So Sasha takes the data from the real account, and tests it on the demo.

[Renat Akhtyamov]

Evgeniy Chumakov:

So Sasha takes the data from the real account, but tests it on the demo.

testing on a demo, extreme fantasy...

Now let him try it on the real account and then tell us where it really hurts and where it does not.

Let him get a taste of Puppet's tics, so to speak.

[Alexander_K]

Evgeniy Chumakov:

So Sasha takes the data from the real account and tests it on the demo account.

Exactly. I have 2 terminals running on my computer. One as a server with a real NDD account from which I take quotes, and the other is a demo account where I just execute open/close commands and that's it. The difference with the real one I cannot see.

[Evgeniy Chumakov]

Alexander_K:

Exactly. I have 2 terminals running on my computer. One as a server with a real NDD account from which I take quotes, and the other is a demo account where I just execute open/close commands and that's it. I do not see any difference with the real one.

The difference is that the difference may appear as soon as the order is opened on the real account.