Prediction of random charts w/poll | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Is it just me or is cable acting random? 49 replies

Is random really random? 231 replies

Charts, Charts, Charts! 1 reply

Random charts 2 replies

EURUSD prediction poll for fun 5 replies

Attachments: Prediction of random charts w/poll
Exit Attachments

Prediction of random charts w/poll

  • Poll

  • can chartists be successful on random chandlestick charts?
  •  
  •  
  •  
  • #1
  • First Post: Aug 25, 2013 9:23am Aug 25, 2013 9:23am
i've tested random charts with chandlesticks. i've also tested a lot of normal charts. what works in normal works for random.
so my conclusion would be that they behave exactly the same way and that they are the same and that the markets are just as predictable as the atmospheric noise.

a prediction will sound like this . if price reaches 100 then it will go to 150 without reaching 50 with a 70-75% winrate.
is it a good prediction?
what winrate and how many predictions do you have to do in order to prove for instance that random.org translated into chandlesticks can be predicted with technical analysis?
  • #2
  • Aug 25, 2013 10:23am Aug 25, 2013 10:23am
i perceive price to be random for the vast majority of the time.

i also perceive price to be non random for short periods of time when liquidity is lacking and a dominant force will take price to the next level of interest.this is clearly structured and can be seen in the way price moves.this repeats its self all day everyday,on all t/frames.

there are also periods when the whole markets have total agreement this is like a harmonic symphony playing a master piece,these moments of total agreement can not be manipulated.although they happen less often once a week would be a reasonable average.
  • #3
  • Aug 25, 2013 10:37am Aug 25, 2013 10:37am
i asked for a specific test and not your opinon on markets. did you read the topic?
  • #4
  • Aug 25, 2013 10:48am Aug 25, 2013 10:48am
Quoting deltatrade
i asked for a specific test and not your opinon on markets. did you read the topic?
this is my conclusion to the topic .the initial statement is somwhat vague
  • #5
  • Aug 25, 2013 12:25pm Aug 25, 2013 12:25pm
hot thread eh? FF ruling as per usual
tradewith60
  • #6
  • Aug 25, 2013 1:00pm Aug 25, 2013 1:00pm
no responses? couple years ago i would have gotten some responses from knowledgeable people.
  • #7
  • Aug 25, 2013 2:25pm Aug 25, 2013 2:25pm
They don't behave in exactly the same way. You can distinguish a random chart from a non-random chart by using any of more than 20+ statistical tests that allow you to evaluate the random walk null hypothesis. You can use the "vrtest" package in R to evaluate any of these methods quite easily. This test :

Escanciano, J.C., Lobato, I.N. 2009a. An automatic portmanteau test for serial correlation. Journal
of Econometrics 151, 140-149.

Is my favorite. There are many pairs and time frames where you can reject the random walk hypothesis with a more than 99% confidence (but I won't do your homework for you). Many pairs have data with serial correlations -- which means they don't follow a random walk.
  • #8
  • Aug 25, 2013 3:08pm Aug 25, 2013 3:08pm
Quoting deltatrade
i've tested random charts with chandlesticks. i've also tested a lot of normal charts. what works in normal works for random. so my conclusion would be that they behave exactly the same way and that they are the same and that the markets are just as predictable as the atmospheric noise. a prediction will sound like this . if price reaches 100 then it will go to 150 without reaching 50 with a 70-75% winrate. is it a good prediction? what winrate and how many predictions do you have to do in order to prove for instance that random.org translated...
how many predictions?
That's science 101, should have been studied in high school. Things like these can't be proven right, they only can be proven wrong.
So, if your hypothesis is that markets and charts are random at all times, then it is impossible to prove that this is always the case. But if I can show you one chart which doesn't act randomly, then your hypothesis is proven to be wrong. That is how it works in science.
Best example is gravity. We can't prove that gravity exists everywhere on the planet. We just assume it to be true, and that is our hypothesis. However, if you can show me a place where the earth has no gravity, then you have proven that the hypothesis is wrong.

So how many predictions do you need? none at all, you just don't want any predictions that prove your randomnes hypothesis wrong.
  • #9
  • Aug 25, 2013 3:13pm Aug 25, 2013 3:13pm
Quoting Custos
{quote} Best example is gravity. We can't prove that gravity exists everywhere on the planet. We just assume it to be true, and that is our hypothesis.
the fact that the earth doesn't have gravity in one place doesn't make the observations that we have less true. we should change the hypothesis to include the new data.
  • #10
  • Aug 25, 2013 3:20pm Aug 25, 2013 3:20pm
there are several statements.
1 atmospheric noise can be predicted with ta.
2 financial charts have the same predictability as atmospheric noise.
now you need proof that this is the case.
  • #11
  • Aug 25, 2013 3:29pm Aug 25, 2013 3:29pm
Quoting deltatrade
what winrate and how many predictions do you have to do in order to prove for instance that random.org translated into chandlesticks can be predicted with technical analysis?
100% win rate
∞ number of trades ...

and it still wouldn't be enough
  • #12
  • Aug 25, 2013 3:36pm Aug 25, 2013 3:36pm
Quoting chazzle
{quote} 100% win rate ∞ number of trades ... and it still wouldn't be enough
high standards you have there. nothing in real life is 100. ask the fukushima builders. only math can be 100%.
  • #13
  • Aug 25, 2013 3:38pm Aug 25, 2013 3:38pm
Prediction of random (charts)? now that's not trading. There can't be any consistent data in a random chart.

Prediction of trader's actions. Their orders are what move the financial markets.
  • #14
  • Aug 25, 2013 3:44pm Aug 25, 2013 3:44pm
Quoting EmeraldEyes
Prediction of random charts? now that's not trading. Prediction of trader's actions. Their orders are what move the financial markets.
do you communicate with 1 mil traders in the world ? do you know where traders are going to enter and at what second? do you know all the strategies including the winning ones?

are you a chartist?
  • #15
  • Aug 25, 2013 4:45pm Aug 25, 2013 4:45pm
Quoting erik1
They don't behave in exactly the same way.
they give the same signals. they have the same winrate. the targets are met.
  • #16
  • Aug 25, 2013 4:57pm Aug 25, 2013 4:57pm
Quote
they give the same signals. they have the same winrate. the targets are met.

This does not mean that they behave in the same way. They are statistically different. Run statistical tests for serial correlation and you'll see. Statistical fact.
  • #17
  • Edited 10:37pm Aug 25, 2013 6:44pm | Edited 10:37pm
.
Riding the order flow
  • #18
  • Aug 25, 2013 8:44pm Aug 25, 2013 8:44pm
Quoting deltatrade
i've tested random charts with chandlesticks. i've also tested a lot of normal charts.
what is a random chart? how does it differ from a normal chart?
  • #19
  • Aug 25, 2013 9:50pm Aug 25, 2013 9:50pm
Quoting deltatrade
{quote} do you communicate with 1 mil traders in the world ? do you know where traders are going to enter and at what second? do you know all the strategies including the winning ones? are you a chartist?
No, don't need to communicate w/ individual traders when dealers report theirs and larger player's orders & positions for me.
Not at which second, but at what price level(s) and usually who they are trading for. Ex; Eur/Chf has a very large trader's buy order @ 1.2005. If the market moved randomly, then why did it not go down below 1.20 last year.
Nope, don't care to waste time knowing 'strategies'
Idk what a chartist is.
  • #20
  • Aug 25, 2013 11:36pm Aug 25, 2013 11:36pm
I voted no, but let me say this - you actually can profit from a random market, provided there is either a floor on the market, a ceiling on the market, or both, and assuming the market will not ever reach a value of zero. I've proven it to myself through thousands of simulated runs on a program I wrote but the catch is the profit is so small that you already need a fortune to make anything decent in the real world. It did account for spread; interest was assumed to be neutral or positive so not universally applicable but it's all largely moot anyhow.
0 traders viewing now