who has series of random generated price data? | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Is random really random? 231 replies

How to remove indicator generated after backtest? 1 reply

MT4 - No signal being generated 0 replies

no ticks generated in strategy tester 2 replies

Price action pattern, and random thoughts 6 replies

Attachments: who has series of random generated price data?
Exit Attachments

who has series of random generated price data?

  • #1
  • First Post: Feb 19, 2011 8:11am Feb 19, 2011 8:11am
I just wanna get some fake price data(H1 data, or M5 data...) which is generated by computer and follows the random walk, so no strategy will work with that data.
We can compare it to the real price data, and get the difference, and develop a strategy.
  • #2
  • Feb 19, 2011 8:43am Feb 19, 2011 8:43am
Quoting bobowang
I just wanna get some fake price data(H1 data, or M5 data...) which is generated by computer and follows the random walk, so no strategy will work with that data.
We can compare it to the real price data, and get the difference, and develop a strategy.
Generating the data is very easy, but I wonder how you are going to use that to develop and test strategies?

If you are serious about this, you might benefit from reading David Aronson's Evidence-Based Technical Analysis.

http://www.amazon.co.uk/Evidence-Bas...8126160&sr=8-1

And your assumption that no strategies can make money from a random market is wrong. Claude Shannon, the information theory pioneer, discovered what has later been called volatility pumping, which is a strategy to make money from a random market. Developments of the idea are also in use for real life trading strategies. See the following link for a taster.

http://www.castrader.com/2006/11/universal_portf.html
  • #3
  • Feb 19, 2011 11:14am Feb 19, 2011 11:14am
thanks so much,
I know how to generate the random price, but when I try to generate 10 years h1 data, my matlab is always "out of memory". I can't sovle it. So I just wander if I can download some random price generated by others.

use the random data, you can see a lot of price patterns showed in Jim16 thread also happened very often. if you can see big difference between the real price data and random generated price data when testing a price pattern, then this price pattern may work!

and the "volatility pumping" sounds interesting. Does that mean we can earn money from tossing a coin using some special strategy? it's just against the math.let me check it.


Quoting medici
Generating the data is very easy, but I wonder how you are going to use that to develop and test strategies?

If you are serious about this, you might benefit from reading David Aronson's Evidence-Based Technical Analysis.

http://www.amazon.co.uk/Evidence-Bas...8126160&sr=8-1

And your assumption that no strategies can make money from a random market is wrong. Claude Shannon, the information theory pioneer, discovered what has later been called...
  • #4
  • Feb 19, 2011 11:40am Feb 19, 2011 11:40am
now i know who is Claude Shannon, i am very familiar with his chinese name "申农"...ha.

his strategy doesn't work. believe me. definitely you can't earn money from a random walk market!
  • #5
  • Edited 12:32pm Feb 19, 2011 11:54am | Edited 12:32pm
Quoting bobowang
now i know who is Claude Shannon, i am very familiar with his chinese name "申农"...ha.

his strategy doesn't work. believe me. definitely you can't earn money from a random walk market!
What do you mean? See attached spreadsheet modelling flips of a fair coin where this strategy wins.
Attached File(s)
File Type: xls Volatility Pumping.xls   512 KB | 473 downloads
  • #6
  • Edited 12:51pm Feb 19, 2011 11:57am | Edited 12:51pm
Quoting bobowang

use the random data, you can see a lot of price patterns showed in Jim16 thread also happened very often. if you can see big difference between the real price data and random generated price data when testing a price pattern, then this price pattern may work!
That may wishful thinking unless you really know how to test this statistically. Here's a quote from the book I mentioned earlier:

From the standpoint of EBTA (Evidence Based Technical Analysis) subjective methods are the most problematic. They are essentially meaningless claims that give the illusion of conveying cognitive content. Because the methods do not specify how they are to be applied, different analysts applying it to the same set of market data can reach different conclusions. This makes it impossible to determine if the method provides useful predictions. Classical chart pattern analysis, hand-drawn trend lines, Elliot Wave Principle, Gann patterns, Magic T's and numerous other subjective methods fall into this category. Subjective TA is religion - it is based on faith. No amount of cherry-picked examples showing where the method succeeded can cure this deficiency.

So what is your proposed testing method? Kernel regression, bootstrapping, or something else?
  • #7
  • Feb 19, 2011 12:45pm Feb 19, 2011 12:45pm
Quoting medici
What do you mean? See attached spreadsheet modelling flips of a far coin where this strategy wins.
Then either the spread sheet is wrong (not a random walk) or there is some other well hidden flaw like in any other paradoxon too. This would be the perpetuum mobile of trading.

This strategy simply cannot work since it can be formally proven that you cannot profit from trading a random walk.

In the very same way I can make a general proof that a perpetuum mobile cannot exist and then never ever have to look at any concrete "implementation" again to prove it wrong, the above mentioned spread sheet can be proven wrong by simply referring to the proof that such a strategy is generally not possible. Its not needed to find where exactly the error is, its enough to know that there must be an error.
  • #8
  • Feb 19, 2011 12:49pm Feb 19, 2011 12:49pm
Quoting 7bit
Then either the spread sheet is wrong (not a random walk) or there is some other well hidden flaw like in any other paradoxon too. This would be the perpetuum mobile of trading.

This strategy simply cannot work since it can be formally proven that you cannot profit from trading a random walk.

In the very same way I can make a general proof that a perpetuum mobile cannot exist and then never ever have to look at any concrete "implementation" again to prove it wrong, the above mentioned spread sheet can be proven wrong by simply referring...
The spreadsheet is correct.

This example is taught in information theory classes in universities all over the world.

Could you please show me the "proof" you are referring to?
  • #9
  • Feb 19, 2011 12:56pm Feb 19, 2011 12:56pm
Quoting 7bit
Then either the spread sheet is wrong (not a random walk) or there is some other well hidden flaw like in any other paradoxon too. This would be the perpetuum mobile of trading.

This strategy simply cannot work since it can be formally proven that you cannot profit from trading a random walk.

In the very same way I can make a general proof that a perpetuum mobile cannot exist and then never ever have to look at any concrete "implementation" again to prove it wrong, the above mentioned spread sheet can be proven wrong by simply referring...
I checked the file, and the edge comes from the doubling-halfing rule of the coin flip game. This edge isn't present in trading, as profits and losses are one-for-one.
The nail that sticks out gets hammered back in
  • #10
  • Feb 19, 2011 1:02pm Feb 19, 2011 1:02pm
Quoting Noloqy
I checked the file, and the edge comes from the doubling-halfing rule of the coin flip game. This edge isn't present in trading, as profits and losses are one-for-one.
The edge is to some extent present in real markets. For a starter see Chapter 16 (Information Theory and Portfolio Theory) in Thomas Cover's Elements Of Information Theory. Or see attached pdf.

There are even hedge funds that base their strategies on this principle...
Attached File(s)
File Type: pdf COV NSF 53.pdf   575 KB | 586 downloads
  • #11
  • Feb 19, 2011 1:06pm Feb 19, 2011 1:06pm
i just lost my internet connection,it 3am in beijing. i am using my cell phone to reply. i will post my detail comments tomorrow.
  • #12
  • Feb 19, 2011 1:07pm Feb 19, 2011 1:07pm
Quoting medici
The spreadsheet is correct.

This example is taught in information theory classes in universities all over the world.

Could you please show me the "proof" you are referring to?
You will surely find it when you do some research.

The fact that something is "taught" does not mean it is correct. It has been taught that the earth is flat for a long time and eventually it was proven wrong. Or maybe (more likely) they simply do not even claim to trade random walks (see my next argument) and you and many others and the creator of this spread sheet simply understood it wrong.

The example with rebalancing the value that is held in a portfolio of stocks which is used in the practical example is not trading random walks at all: Effectively with this method you are trading some kind of grid system on the (multidimensional) spread between all the stocks and it can be reasonably assumed that this spread is not a random walk at all, it is more likely that this spread is almost stationary! And trading a grid (or similar) system on a stationary instrument is not trading a random walk at all!
  • #13
  • Feb 19, 2011 1:07pm Feb 19, 2011 1:07pm
Quoting Noloqy
This edge isn't present in trading, as profits and losses are one-for-one.
Perhaps true for specific strategies, but not true if we are discussing the set of all possible strategies.

Then the model chosen for asset price dynamics would be the arbiter on profits versus losses. And, in fact, most models used are based on some form of linearity in the log-prices - which would leave the edge in the spreadsheet intact.
  • #14
  • Feb 19, 2011 1:09pm Feb 19, 2011 1:09pm
Quoting 7bit
You will surely find it when you do some research.

The fact that something is "taught" does not mean it is correct. It has been taught that the earth is flat for a long time and eventually it was proven wrong. Or maybe (more likely) they simply do not even claim to trade random walks (see my next argument) and you and many others and the creator of this spread sheet simply understood it wrong.

The example with rebalancing the value that is held in a portfolio of stocks which is used in the practical example is not trading random walks...
In this case you are the flat-lander....
  • #15
  • Feb 19, 2011 1:15pm Feb 19, 2011 1:15pm
Quoting medici
In this case you are the flat-lander....
I'm sorry to inform you that you did not understand the "volatility pumping" strategy, it is not trading random walks. And the spread sheet is wrong as the other poster already explained (and you already admitted (the log problem and not trading a true random walk)).

Volatility pumping is some kind of brute force stat arb and works because enough of them are at least to some degree cointegrated well enough in the long run. It would not work with all independent random walks.
  • #16
  • Feb 19, 2011 1:19pm Feb 19, 2011 1:19pm
Quoting 7bit
I'm sorry to inform you that you did not understand the "volatility pumping" strategy, it is not trading random walks. And the spread sheet is wrong as the other poster already explained (and you already admitted (the log problem and not trading the random walk)).
No, I am not the one misunderstanding things here...
  • #17
  • Feb 19, 2011 1:21pm Feb 19, 2011 1:21pm
Quoting bobowang
i just lost my internet connection,it 3am in beijing. i am using my cell phone to reply. i will post my detail comments tomorrow.
I'm sorry to hear. As this thread has been invaded by half-educated dimwits, I think I am wasting my time, and I shall leave you with a suggested method for generating random data and testing the chart patterns (see pdf).

Goodbye and Good Luck!
Attached File(s)
File Type: pdf Marshall Young Rose 2006.pdf   199 KB | 7,363 downloads
  • #18
  • Feb 19, 2011 1:28pm Feb 19, 2011 1:28pm
Quoting medici
No, I am not the one misunderstanding things here...
Yes you are.

Obviously. Otherwise you would understand my arguments and would be able to come up with a meaningful reply instead of insults. I won't continue my posting here in this thread for today to give you some time to think about it, maybe you still have a chance to understand it if you try to think about it a little longer and a little harder.
  • #19
  • Feb 19, 2011 1:32pm Feb 19, 2011 1:32pm
Quoting 7bit
Yes you are.

Obviously. Otherwise you would understand my arguments and would be able to come up with a meaningful reply which you didn't. I won't continue my posting here in this thread for today to give you some time to think about it, maybe you still have a chance to understand it if you try to think about it a little longer and a little harder.
Very interesting. You jump into a thread and critise without proper argument. For example, I asked you to present the proof you claimed you had, and you couldn't.

Whereas every piece of information I have presented is correct and has been provided with the intention of assisting bobowang.

And now you accuse me of not providing meaningful replies!

Goodbye.
  • #20
  • Feb 19, 2011 1:44pm Feb 19, 2011 1:44pm
Quoting medici
Very interesting. You jump into a thread and critise without proper argument.
You invaded this thread with the bold claim that one could trade a random walk.

Extraordinary claims like this need extraordinary proof.

Instead you come up with the "volatility pumping" example which has nothing to do with trading random walks at all, it is about trading a portfolio of stocks which are not random walks at all, and a flawed excel sheet which also does not implement a random walk as it has become clear after analyzing it.

All arguments were already given in this thread. Simply scroll back and try to understand them.
0 traders viewing now