I’ve got 4 charts below. Two were made using a completely random formula in excel and two were obtained with real market data. All four charts were made in excel in order to prevent giving away the correct ones.
Can you figure out which two of the four charts were obtained from real market data, and which two were are completely random? I will post the solution during the week…
Enough suspense… Time for the answer (If you do NOT want to see it yet LOOK AWAY! :27:)
Drum roll please…
The REAL charts - B and D
The FAKE/RANDOM charts - A and C
I was surprised myself that I could find a clear channel in Chart A and what looked like a downward trend in Chart C despite the fact that they were generated by random with no bias or weighting one way or another.
My thoughts: Perhaps a lot of the methods used in the analysis of charts (price action, channels, moving averages, etc.) have been developed from market behaviour which is essentially “random” in nature (I.e. Future bars have no relationship with past bars). How then can these techniques be used for for trading decisions? Maybe the multitude of chart analysis techniques out there have simply been forced to find relationships between cause and reaction in the market that do not exist (not unlike superstition).
It’s an interesting insight into market behaviour.
Does price action follow patterns, or is it random collections of movements to which we assign arbitrary significance.
I’m not sure I quite follow what you mean. As I understood it, each bar in the random graphs was generated with random OHLC data - making a range of different candles and these were then shown in order, effectively linking one period close with the next open.
In the real data, each period’s moves is effected by the previous movements in the market - therefore not random.
Randomly created candles or not, on any of situations you’re either buying dips, or selling rallies. Each of those charts give you a bias to one or the other. The poster children tech traders get so caught up in the “data” trees, they lose sight of the of the chart forest.
It’s not a difficult endeavor. All that makes it difficult is the space between one’s ear.
But if a chart uses random movements it omits momentum, volatility and many other inputs which act over multiple periods. Without this each is a distinct event.
The dips and rallies in the random chart happen by chance with no correlation. Forces correlated over multiple candlesticks increase the likelihood of a move in one direction (I’m not saying it causes it to happen, but it does increase the chances). Therefore I don’t understand how anyone could trade both types of charts equally.
The two random charts were generated by starting with a value and then randomly adding a number between -x to x several hundred times in a row (a completely random walk with no bias). The OHLC values were then obtained by breaking the row of data into blocks of 20 cells each, which I then used to calculate the OHLC values as normal. So each bar in the random charts summarize the values from 20 cells.
I would like to find out if there is anyway to distinguish a random collection of data which has no relationship with previous bars to actual market data. And if not, what is the validity of using technical analysis which is based on past data to predict the future?
It is elementary my dear Watson - -
Charts A and B both start at a 1.25 number
Charts C and D both start at at 1.4 number - therefore it is proof that we are comparing chart a with b to determine which is fake or real - - we are not comparing chart a with either chart c or chart d.
Hence we now just have to verify which of the charts A or B is fake - -because human nature will ‘post’ the charts in the same order - -as in
If you posted chart A first – and it was a fake – then very likely you posted chart C first which would also be the fake.
So to determine which charts are fake and which are real - it is very simple -
you have given the charts a random number between x and - x - therefore they cannot possibly go beyond the value of x or -x - which means - yes even though the bars may generate different heights etc. . . . it is quite plain that the generated charts will flow more slowly - -law of newton –
therefore
Your “random” candles did a Rich Little impression of a working chart. Define your version of technical analysis. What do you use?
What do you think your daily candles are made of? “Noise” is such an irritating way to discount the importance price movement. All those little moves make up the same type of moves on exceedingly longer time frames. Charts are fractal.
Ok. I couldn’t wait any longer and have edited the first post with the solutions!
Well done to GRIX FX who got the answer. Can I ask how you managed to figure it out?
And to PerchTird for his excellent deductive process (although wrong in the end).
@Master Tang: My definition of technical analysis is anything to do with past market behaviour (channels, trends, other indicators, which are mostly based on MAs, etc.)
