Time-dependence of Excel Rand function | Forex Factory

Options

Permalinks
Print Thread

Similar Threads

MT4 to Excel through Excel RTD (Finally!) 612 replies

required to function call function 2 replies

stop loss function and edit time function for Breakout EA 1 reply

I need help with a particular Excel Function and Formula 16 replies

Need a little help of a couple simple function in Excel 13 replies

Time-dependence of Excel Rand function

  • #1
  • First Post: Edited 8:48am Apr 10, 2010 7:47am | Edited 8:48am
I have generated price chart using Excel Rand function. Nevertheless, for the charts generated I have no way to believe it is random. Even though Microsoft claims it is random, I am not convinced. I guess they say it is truly random only in the sense that the prob of the value generated is evenly within the range of 0 to 1. Nevertheless, I believe that the random numbers are somehow time-dependent.

Any wise method for testing whether the data are time-dependent?

http://www.forexfactory.com/showpost...8&postcount=48
http://www.forexfactory.com/showpost...7&postcount=50
http://www.forexfactory.com/showpost...7&postcount=30

Many thanks.

http://www.forexfactory.com/attachme...6&d=1269613820
http://www.forexfactory.com/attachme...7&d=1269613820
  • #2
  • Apr 10, 2010 7:50am Apr 10, 2010 7:50am
Alternatively, if you have any method to generate truly random price chart. Please let me know.

Many thanks
  • #3
  • Apr 10, 2010 8:36am Apr 10, 2010 8:36am
I used to try 2 techniques to test the RNG.

1st: I am not sure about other excel version but the RNG for my old excel version 2002(10.2614.2625) fails to follow normal distribution above 2 standard deviation. I have tested with some random number generated from a webpage(I forgot which one) and it is completely normally distributed.

2nd: I think Hurst exponent is a better way to verify it. For my old excel version the hurst exponent of the RNG is fluctuated around 0.52-0.57. I get this number by doing a random walk model in excel.

Btw i didnt use rescaled range analysis to calculate the hurst exponent because it's too complicated. I only use a simple method:

Position after certain step = Std Deviation of every step X Number of step ^ Hurst

p/s: I think this kind of discussion shouldnt consider as programming discussion.
  • #4
  • Apr 10, 2010 8:45am Apr 10, 2010 8:45am
Thanks for this. Cheers you are the right person.

Nevertheless, can these methods test time-dependence? I means there are cases in which the data set can be completely random and normal distributed, but the sequence of the data series is clearly not random.

P.S. Programming forum may not be the best place, but also it may be too technical to put it in general trading discussion.

Quoting ForexQuant
I used to try 2 techniques to test the RNG.

1st: I am not sure about other excel version but the RNG for my old excel version 2002(10.2614.2625) fails to follow normal distribution above 2 standard deviation. I have tested with some random number generated from a webpage(I forgot which one) and it is completely normally distributed.

2nd: I think Hurst exponent is a better way to verify it. For my old excel version the hurst exponent of the RNG is fluctuated around 0.52-0.57. I get this number by doing a random walk model in excel.

Btw i didnt...
  • #5
  • Apr 10, 2010 8:53am Apr 10, 2010 8:53am
Quoting kk007
Thanks for this. Cheers you are the right person.

Nevertheless, can these methods test time-dependence? I means there are cases in which the data set can be completely random and normal distributed, but the sequence of the data series is clearly not random.

P.S. Programming forum may not be the best place, but also it may be too technical to put it in general trading discussion.
If i'm not mistaken, you mean the random data has memory right? Hurst exponent is the right tools, IMHO.
  • #6
  • Apr 10, 2010 9:00am Apr 10, 2010 9:00am
Quoting ForexQuant
If i'm not mistaken, you mean the random data has memory right? Hurst exponent is the right tools, IMHO.
OK. Let me try to study this. Thanks a lot.
http://en.wikipedia.org/wiki/Hurst_exponent
  • #7
  • Apr 10, 2010 9:04am Apr 10, 2010 9:04am
Quoting ForexQuant
If i'm not mistaken, you mean the random data has memory right? Hurst exponent is the right tools, IMHO.
Hi ForexQuant,

If the Hurst exponent is that useful, what happen if I take a real EURUSD daily chart for testing? Would it tell me whether the price movement is random??????

cheers,
kk
  • #8
  • Apr 10, 2010 9:16am Apr 10, 2010 9:16am
Quoting kk007
Hi ForexQuant,

If the Hurst exponent is that useful, what happen if I take a real EURUSD daily chart for testing? Would it tell me whether the price movement is random??????

cheers,
kk
i believe so.
  • #9
  • Apr 10, 2010 9:24am Apr 10, 2010 9:24am
Quoting ForexQuant
i believe so.
It will be great. Many people on the forum may argue that there is no evidence to prove price movement is non-random.

Thanks. I haven't studied the testing method yet. Let me check it out to see. cheers.
  • #10
  • Apr 10, 2010 9:50am Apr 10, 2010 9:50am
Quoting kk007
It will be great. Many people on the forum may argue that there is no evidence to prove price movement is non-random.

Thanks. I haven't studied the testing method yet. Let me check it out to see. cheers.
There are so many proof that market is not random. However, it does not mean that we can predict the market with great accuracy if it is not random, in fact market itself is a non-linear, dynamic, non-deterministic chaotic system.
  • #11
  • Apr 10, 2010 9:58am Apr 10, 2010 9:58am
Quoting ForexQuant
There are so many proof that market is not random. However, it does not mean that we can predict the market with great accuracy if it is not random, in fact market itself is a non-linear, dynamic, non-deterministic chaotic system.
Cool! Can you send me some readings/journal, etc.? Thanks.
  • #12
  • Apr 10, 2010 10:14am Apr 10, 2010 10:14am
IMHO, The (Mis)Behaviour of Market, Fooled by Randomness and The Black Swan is a good start.
  • #13
  • Apr 10, 2010 10:16am Apr 10, 2010 10:16am
Quoting ForexQuant
IMHO, The (Mis)Behaviour of Market, Fooled by Randomness and The Black Swan is a good start.
Isn't "Fooled by Randomness" a book claiming that everything is random?

Cheers,
  • #14
  • Last Post: Apr 10, 2010 10:53am Apr 10, 2010 10:53am
Personally I prefer to use 'non-deterministic chaotic' instead of 'random'. They looks similar but not the same.

Btw i'm still learning all these things and i maybe 101% wrong!
0 traders viewing now