S/R and TL on Randomized Price Movement | Page 5 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Price movement as a function of Supply and Demand 43 replies

Spot Fx Trades based on futures volume and price movement 22 replies

How to differentiate pro and amateur price movement? 6 replies

Random Price Movement 177 replies

GBP/USD daily pip movement vs. EUR/USD daily pip movement 3 replies

Attachments: S/R and TL on Randomized Price Movement
Exit Attachments

S/R and TL on Randomized Price Movement

  • #81
  • Apr 10, 2010 10:02am Apr 10, 2010 10:02am
Hi kk,

I didnt notice this interesting thread until you pm me. Reading from 1st page now...
  • #82
  • Apr 10, 2010 10:06am Apr 10, 2010 10:06am
Quoting ForexQuant
Hi kk,

I didnt notice this interesting thread until you pm me. Reading from 1st page now...
Enjoy! Thanks for your previous info.
  • #83
  • Apr 10, 2010 11:16am Apr 10, 2010 11:16am
They are so many interesting points in this thread, here is my 1st opinion.

The excel formula NORMINV(RAND(),0,1) does not reflect the true market condition because it's 99% normally distributed but a real price change is not normally distributed. Therefore i dont think this is a proper way to simulate the market price.

Here is the example distribution chart & Q-Q Plot for a return generated by excel and real GBPUSD Hourly return. It is obviously that real market return has an extreme fat tails.
Attached Image(s) (click to enlarge)
Click to Enlarge Name: Excel RNG.jpg Size: 25 KB
Click to Enlarge Name: GBPUSD.jpg Size: 19 KB
Click to Enlarge Name: Excel RNG QQPlot.jpg Size: 19 KB
Click to Enlarge Name: GBPUSD QQPlot.jpg Size: 21 KB
  • #84
  • Apr 10, 2010 11:56am Apr 10, 2010 11:56am
Quoting Bytebodger
In almost every instance, something identified as "random" can more accurately be identified as "we don't (or can't) properly understand all of the factors that determine the outcome".

Probably the most common example of a "random" event is a coin flip....
I totally agree in his example. In theory if you can measure all other variables in 100% accuracy then the outcome is predictable. However, 100% of measurement accuracy is not possible therefore the exact outcome cant be calculated. Even 99.999999999999999999% of accuracy will create a totally unexpected results. This phenomena is well known as butterfly effect.

Since we have no way to do an accurate measurement then statisticians try to describe the phenomena statistically and surprisingly some cases like coin flip/Quantum mechanics can be model well as a random outcome. Maybe we can say that coin flip is a non-deterministic chaotic process, where the result is guided by some simple hidden rules but no way you can measure them.
  • #85
  • Apr 10, 2010 12:02pm Apr 10, 2010 12:02pm
Quoting sangmane
anyone know how to measure the randomness of a set of data? i think it's better to state the randomness with a number (like 90% random) than just saying "not completely random".
Hurst exponent.
  • #86
  • Apr 10, 2010 12:07pm Apr 10, 2010 12:07pm
Quoting dr_zeus

Hanover, your comments on roulette are apropos. I once made a roulette simulator in excel and tried to find an edge using ranges and trends - there isn't one. It could be that the random charts posted here have a real difference from real price charts since some traders feel like they have found an edge. I certainly find inside bars to be a good trade. It could also be that they are benefiting from fundamental drift, carry interest, and increased ranges per Mikkom's chart.
It doesnt matter whether the outcome is random or not random, the edge of a static system like roulette is skew towards the house during the design of the game, therefore IMHO there is no way to reverse the edge.
  • #87
  • Apr 10, 2010 12:10pm Apr 10, 2010 12:10pm
Quoting ForexQuant
It doesnt matter whether the outcome is random or not random, the edge of a static system like roulette is skew towards the house during the design of the game, therefore IMHO there is no way to reverse the edge.
There is a way to reverse the edge: be the house

BTW, roulette is not "random" in the strict sense (since it's based on Newtonian mechanics): http://www.timesonline.co.uk/tol/new...icle399673.ece
  • #88
  • Edited 4:54pm Apr 10, 2010 12:12pm | Edited 4:54pm
Quoting kk007
But Microsoft told me that they no longer use the pseudo-random number generator in their 2007 version.

"It (2007 version) passes the same battery of standard tests. The battery of tests is named Diehard (see note 1). The algorithm that is implemented in Excel 2003 was developed by B.A. Wichman and I.D. Hill (see note 2 and note 3). This random number generator is also used in the RAT-STATS software package that is provided by the Office of the Inspector General, U.S. Department of Health and Human Services. It has been shown by Rotz et al...
IMHO there is no way to generate a 100% true randomness because the random number generator always guided by some hidden complex rules, if there are rules then it's not 100% true random.

But i believe that 99.9% of randomness is achivable. I think this is good enough for practical use.
  • #89
  • Apr 10, 2010 12:21pm Apr 10, 2010 12:21pm
Quoting ForexQuant
IMHO there is no way to generate a 100% true randomness because the random number always guided by some hidden complex rules, if there are rules then it's not 100% true random.

But i believe that 99.9% of randomness is achivable. I think this is good enough for practical use.
I think you can increase the radomness percent to 99.9999% or even more.
http://en.wikipedia.org/wiki/Hardwar...mber_generator

To simulate a random walk you don't need the best random number generator in the universe.
  • #90
  • Apr 10, 2010 12:43pm Apr 10, 2010 12:43pm
Quoting Adal
There is a way to reverse the edge: be the house

BTW, roulette is not "random" in the strict sense (since it's based on Newtonian mechanics): http://www.timesonline.co.uk/tol/new...icle399673.ece
Obviously it's not because it can be explain by newtonian mechanics, just you cant measure the initial condition accurately.

I once focus on study & research of market randomness but i eventually stop halfway because i really have no idea how this can improve my profitability. It does not matter whether the system is random or non random, the most important element of profitable trading is EDGE! It is not our job as a trader to prove that the market is random or non random, our job is find the calculated edge!

So the question to profitability is can you extract a statistical significant edge from the market? If yes then random or non random is not your business.

Just my 2 cents. As usual, I could be deathly wrong
  • #91
  • Apr 10, 2010 1:18pm Apr 10, 2010 1:18pm
I admit I'm stepping out of my box with this one, but to my knowledge many quant traders use mean reversion stratagies. My math skills make their home on old arthritic lane. My math deficient understanding tells me that it's stat based OB/OS type thing where price returns to an average price somewhere. I'm actually just guessing here. So what are mean reversion strategies all about?

Lets throw stat arb in here to. I have more experience with this, but I've never seen a good example of stat arb as far as parameters are concerned..
  • #92
  • Apr 10, 2010 1:37pm Apr 10, 2010 1:37pm
I don't need arbitrage in general explained to me BTW. But what kind of statistical measures could find price discrepancies between instruments?
  • #93
  • Apr 11, 2010 12:14am Apr 11, 2010 12:14am
Thanks ForexQuant,

There was no attempt (as far as I know) to simulate real market movements. (I don't think it is possible). The attempt was to simulate a completely random movement. It sounds to me that Excel Rand passed the normally disturbution test as a RNG.

cheers,

Quoting ForexQuant
The excel formula NORMINV(RAND(),0,1) does not reflect the true market condition because it's 99% normally distributed but a real price change is not normally distributed. Therefore i dont think this is a proper way to simulate the market price.
  • #94
  • Apr 11, 2010 9:32am Apr 11, 2010 9:32am
Quoting ForexQuant
They are so many interesting points in this thread, here is my 1st opinion.

The excel formula NORMINV(RAND(),0,1) does not reflect the true market condition because it's 99% normally distributed but a real price change is not normally distributed. Therefore i dont think this is a proper way to simulate the market price.

Here is the example distribution chart & Q-Q Plot for a return generated by excel and real GBPUSD Hourly return. It is obviously that real market return has an extreme fat tails.
A thousand kudos to you, sir. This is the most useful post in this thread I think. I'd like to hear your interpretation of results.

It appears to me that the GBPUSD shows results very similar to random for most of the sample period, but a few of the hours show extreme moves that are not predicted by a purely random movement.

So for most of the time, we should expect price to move randomly and we would have poor ability to predict what it would do. However, sometimes price makes very large moves - break outs.

I think we can make some trading decisions around this. I'll need to think for a bit.
  • #95
  • Apr 11, 2010 11:01am Apr 11, 2010 11:01am
Quoting dr_zeus
Unproven assertion.

I'm busy now, but I bet I can show the opposite using the same random excel file.

Just gonna chime in to say. Use your brain. You can't profit from random movement. It's not an assertion that needs any proof.
  • #96
  • Apr 12, 2010 8:50pm Apr 12, 2010 8:50pm
Quoting dr_zeus
A thousand kudos to you, sir....
Although the GBPUSD chart indicated that the actual return is not normally distributed but that does not mean the price is random or not random. The extreme moves only tell us that Black Swan does exist more frequent than expected.
  • #97
  • Apr 13, 2010 8:54am Apr 13, 2010 8:54am
Quoting ForexQuant
Although the GBPUSD chart indicated that the actual return is not normally distributed but that does not mean the price is random or not random. The extreme moves only tell us that Black Swan does exist more frequent than expected.
I have used Hurst exponent to test Excel Rand(). Each sample has 10000 Rand() values and I tested 10 times. The results are:
0.51, 0.50, 0.52, 0.51, 0.52, 0.47, 0.50, 0.48, 0.45, 0.50. Sounds to suggested that it time-independently random?
  • #98
  • Apr 13, 2010 9:16am Apr 13, 2010 9:16am
"Just gonna chime in to say. Use your brain. You can't profit from random movement. It's not an assertion that needs any proof."

Yes you can..................... run a bucketshop

SORRY
  • #99
  • Apr 13, 2010 9:21am Apr 13, 2010 9:21am
Quote
On each of them, it seems to me there are clear support/resistance and Trendline. Why is that????
The same reason we see clouds in the shape of an animal. We log in everything we see, and it's likely that random data will appear as one of the many patterns we have logged into our memory. TA patterns are simple (as opposed to a bunny or Jesus' face), so we're likely to see a lot of them in random data. It doesn't matter that the data is random, the points correspond to patterns we have logged in our memory. I usually see an animal when I look at the chart. For example, the 4hr euro looks a bit like a camel at the moment!

Quoting mikkom
There is your million dollar question, how does real data differ from random data.
One has bias.
  • #100
  • Last Post: Apr 13, 2010 12:30pm Apr 13, 2010 12:30pm
Quoting kk007
I have used Hurst exponent to test Excel Rand(). Each sample has 10000 Rand() values and I tested 10 times. The results are:
0.51, 0.50, 0.52, 0.51, 0.52, 0.47, 0.50, 0.48, 0.45, 0.50. Sounds to suggested that it time-independently random?
I guess it depends on the excel version. No matter how many times i tried, my old excel 2002 version will never had hurst 0.5 or below it.
0 traders viewing now