Thoughts on the random - page 24
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
The point is that if you take a 1 point increment, with only M5, you cannot know how many Renko bars a single M5 candle will produce, nor their configuration (hence, the sign change statistics as well). It is the ticks that are needed.
In other words, in the model by opening prices, the average modulus of difference between opening prices of neighboring bars can be 20-30 points, while the average error of reanco generation is exactly proportional to the ratio of the average_size_of_the_watch/renko_increment
Let me explain, Alexey. For example, after 7 bars I may get +15 points at the next open price with increment threshold of 10 points. Suppose the price is equal to 1.2348. A new data point is formed accordingly. Then I say, the next data point is more likely to be formed at the level not higher than 1.2338, than at 1.2358. And I wait for a n-number of bars until this happens. Further, the price may go as low as -20, but on average it will go modulo 10 pips.
Everything seems to be without error.
For the sake of experiment, try running the same script on M1. If the graph is exactly the same, then you are right. If there are differences, then you can immediately see from which increments the plot is wrong.
I will do it. I have just a million minutes for the last 3 years.
Will do. I've got like a million minutes in the last three years.
Okay, we're waiting.
Ok, waiting for it.
I'll post it tonight. I have all the materials at home and I'm not at home right now.
For aesthetes I can run it on all ticks later, real ones, I have several hectares from Ducascopy, but I think minutes will be enough. Also - I haven't learned how to hook data from an external file yet, and Excel doesn't hold more than a million.
I'll also post the script. It's a mini achievement of mine. Works fast, posted report counts in 30 seconds.
I will post it in the evening. I have all the materials at home and I'm not at home right now.
For aesthetes I can run it on all ticks later, real ones, I have several hectares from Ducascopy, but I think minutes will be enough. Also - I haven't learned how to hook data from an external file yet, and Excel doesn't hold more than a million.
I'll also post the script. It's a mini achievement of mine. Works fast, posted report counts in 30 seconds.
Maybe it would be better to switch to MQL5. There are only RAM limitations there. The investigation process will be more productive. And it's easier to study MQL5 than VBA. (You don't have to worry about it. ))
I've got interesting results for about the same scheme. But I've been doing some rough tests so far. At present I'm working on the engine for deep and voluminous tests. Perhaps I will finish it in half a year (I still need to study) and after that I will show the results with detailed description of what I have been doing.
Let me explain, Alexey. For example, after 7 bars I may have got +15 pips at the next opening price, with an increment threshold of 10 pips. Suppose the price is equal to 1.2348. A new data point is formed accordingly. Then I say, the next data point is more likely to be formed at the level not higher than 1.2338, than at 1.2358. And I wait for a n-number of bars until this happens. Further, the price may go as low as -20, but on average it will go modulo 10 pips.
Everything seems to be without error.
Here is the picture:
rank=10 pips. On the opening prices: on the first bar it went up from 0 to 10. You write down the "+". On the second bar, down from 10 to 0. You write "-". Total: "+-".
And in reality (as seen from the shadows) ++--.
I.e. calculation by opening prices will underestimate progression and overestimate return at small values of the rank in comparison with volatility of timeframe, on which it is calculated
here's a picture:
rank=10 pips. On opening prices: the first bar has risen from 0 to 10. You will record a "+". On the second bar, down from 10 to 0. You write "-". Total: "+-".
And in reality (as seen from the shadows) ++--.
I.e. calculation by opening prices will underestimate progression and overestimate return at small values of the rank in comparison with volatility of timeframe, on which it is calculated
Yes, I get it now. Extremes are not taken into account. But if my scheme is somehow applied to trade at open prices, the patterns found will not be lost.
Yeah, got it now. Extremes don't count. But if my scheme is somehow applied to trading exactly at opening prices, then the patterns found will not be lost.
it is possible to write an advisor. 12 years ago the return for small moves was stronger, but the spreads were also much larger.
Based on your calculations, the spread would be less than 2 pips, so that at a rank of 10 pips it would be a plus. This spread is only there for the last 3 years. Calculate the return for the last 3 years and you will see that it is not as attractive. I.e. these low values of 0.44-0.46 were obtained due to 1999-2006 somewhere.
In general, you can trade returns, but knowing when/under what conditions. Head-on - like all the time, griders, etc. won't work.
If you trade 10 years ago with today's spreads, grail deteted))
it is possible to write an advisor. 12 years ago the return for small moves was stronger, but the spreads were also much larger.
Based on your calculations, the spread would be less than 2 pips, so that at a rank of 10 pips it would be a plus. This spread is only there for the last 3 years. Calculate the return for the last 3 years and you will see that it is not as attractive. I.e. these low values of 0.44-0.46 were obtained due to 1999-2006 somewhere.
In general, you can trade returns, but knowing when/under what conditions. Head-on - like all the time, griders, etc. won't work.
If you trade 10 years ago with today's spreads, grail deteted))
))) I will also analyze situations when the price closed in one direction twice in a row, or 3 times in a row. Perhaps the picture will be more interesting.
))) I'll still analyse situations where price closed in the same direction twice in a row, maybe 3 times in a row. Perhaps the picture will be more interesting.