Thread: Can you tell with naked eyes?

I want assistance from seasoned traders. I have generated random price movement and plot OHLC charts. If it's possible to differentiate them, I wonder they are price charts with eyes?

Originally Posted by ;

The reality is slightly more complied: [...] If you turn enough coins, then you'll receive sequences of coin flips that seen in isolation from the rest[...]

Exactly. In isolation is the main point here. This is exactly what I said. These patterns might not be viewed to learn whether it truly is a pattern or whether it's completely normal and expected this sequence appeared.

Originally Posted by ;

To create freshness at the experimentation, they may get a random start point on the price chart to present. It's highly possible that some chart-portion are repeated, and this rep manifest the memory factor. This problem has not been addressed in their paper.

But I just realised from trad3r, he is one of the winner there. The MIT fake charts can't compare to the quality of fakeness on this thread and seem really fake. This is another weakness of the analysis.

Nope, wrong again . New information was used every time. I am not certain what is so difficult about understanding ? The analysis specifically said they used the tick information of a different day .

I don't think you actually read the study

Originally Posted by ;

Nope, wrong again . New data was used each moment. I am not sure what is so difficult about understanding that? The analysis specifically said they used another day's tick data .

I do not believe you really read the analysis

I've read the analysis. You're wrong. You'll discover that I've read it, if you read my comment; it is really obvious. But I have to admit that I haven't read each sentence . Allow me to know where you discover the description say what you mentioned. I.e. they delibrately choose various days for testing and training. If they delibrately do it, it should be fine. If they randomise to generate different item, repetition is possible.

Anyhow this is simply the first portion of the research, in the next part, they perform daily and each time they use the same sequence of daily.

Other issues also happen that make their research cannot use to here as said by two responsents here, trad3r and luqmanz.

Originally Posted by ;

To assess the robustness of our experimental design, we diverse different parameters of
the experiment across data sets, as indied from the Results section below. In addition,
we presented subjects with data charts in two different manners. For half of the data collections
corresponding to transaction-by-transaction (or #8220;Query #8221;-RRB- data, each subject was shown a fresh
group of charts, based on a succession of returns disjoint in the strings shown to additional
subjects.

Originally Posted by ;

For the other half of the data, corresponding to daily data, the charts revealed to each topic were based on precisely the same sequence of returns. On the other hand, the data was changed by a random amount for security reasons, i.e., to prevent the possibility that two subjects could coordinate their suspects, such as by simultaneously playing the same charts on </b>two nearby machines. </b>

P.S. each subject was shown a brand new set of charts as said in the paper not a topic was shown a brand new set of charts each time for the first half
.

Originally Posted by ;

As I have said before....an inexperienced trader will see randomness whereas an experienced trader will see construction.

Are you experienced enough to do this? If so, please allow me to know how to perform it.

Here is one I made earlier...

I took 2^22 amounts from a normal distribution with zero mean and standard deviation 1. Then each 1024 points were taken by me and made a OHLC candle out of it.

That I havn't done exhaustive experimentaion but the price changes on actual market data is generally distributed on everything I've tried so far. I wonder if distinctions are made if you check out ordering or correlation with data points.

I havn't browse mandelbrot's novel or anything like this so I want to read to it I was just wondering if anybody can give me a direction to go in to mathematically inspect data sets taht are great.
https://www.forexsoutheast.asia/atta...4312923336.xls

Originally Posted by ;

Here's one I made before...

I took 2^22 amounts from a normal distribution with zero mean and standard deviation 1. Then I made a OHLC candle and took every 1024 points.

That I havn't done exhaustive experimentaion but the price changes on real market data is normally distributed on everything I have tried thus far. I wonder if distinctions are made if you test out ordering or correlation with previous data points.

I havn't read mandelbrot's book or anything like that so I need to read to it I was wondering if anyone can...

no I believe you're making a mistake, market price fluctuations doesn't follow normal distribution. Prove us a plot if you found any price changes that is normally distributed.

Btw I cant read your excel sheet.

Hmmm, I had a problem uploading... I'll try again at some point but I'm a little distracted.

I tried it using a few 15M charts for EURUSD and GBPUSD, getting the information into a spreadsheet was a faf but I do recall it looking distinctly gaussian.

I'll withdraw the announcement until I've collected the information to back it up, obviously 3 or 4 samples is not sufficient to confirm a concept. Has anyone got some substance / literature (on the internet) on this type of thing?

Originally Posted by ;

HmmmI had a problem uploading... I'll try again at some stage but I am a little distracted.

I tried it with several 15M charts for EURUSD and GBPUSD, getting the data into a spreadsheet was a faf but I do remember it appearing distinctly gaussian.

I'll draw the statement until I have collected the data to back it up, obviously 3 or 4 samples isn't enough to verify a theory. Has anyone got any material/ literature (on the internet) on this type of thing?

Looking distinctly is not a scientific technique to strengthen your hypothesis. Yes it looks like gaussian, but it is not, not even near.

That is true, my motive was to provoke a dialogue about the subject.

Please could you shed a bit more light onto it ?

What methods do you use to fit the supply?