How well to humans know random?

To OP:

Can you plug in a price series to your computer program and determine whether it is random?

Like plug in the closing prices for crude over the past three months and let me know what it says.

Thanks

5yr
 
Quote from nutmeg:

Could the results be different if you asked people to post random numbers if varying amounts of money was at stake?
More...

Probably but it would depend on the specifics of the situation and exactly what the perceived trigger for the payoff was.
 
Quote from florida_pilot:

Perhaps you meant imaginary numbers? There's a rather large difference semantically between the two.
More...

Yes I was going to say imaginary numbers and got distract. LOL.
 
Quote from 5yrtrader:

To OP:

Can you plug in a price series to your computer program and determine whether it is random?

Like plug in the closing prices for crude over the past three months and let me know what it says.

Thanks

5yr
More...

You can't really put things through a test that says either "yes its definetly random" or "nope not random". All you can do is look for a model that can determine futures states ie If it is sunny and yesterday was up then today is up if not then today is down. If you find that this model works for a given time period it might because of some causal relationship or it could be coincidence. It is impossible to know for certain but you can be more confident if you can find that your model works on longer timeframes. The longer the more confident you can be. If you find this then whatever you are looking at is not random. If you cannot find deterministic dependence on past states then you can build a distribution of what might happen (figure out the chance of events happening). For example an unbiased coin flip does not care what happened on the last flip so it does not matter for the next flip and is random. The distribution is 50/50. BUT if you have a bent coin that is biased it still does not matter what happened on the last flip so it is still random even if the distribution is 0.1/99.9.

Since I dont know any model to predict the movement of oil all I can say (just by looking at the chart) is that is it a very bent coin right now.

You can say that there is dependence on past states because the price now depends on the past price alot and does not change to huge values and go to tiny fractions. The thing that matters though is the relative price fluctuations which is what you are trying to understand in this case.
 
0110011010
0101011001
1001010011
0011001010
1001101011
1010010011
1011100111
0011011010
1110110011
0010111110

I think this is binary for "psnbdt kdjrn t gfd rekjgf kfjed DGD!!!!"

but Im not sure. :D

Let me know, I think its 45 0s and 55 1s but it seems that they are random even if unequal. If I did 1000 the results would be better. :)
 
Quote from 88888accountant:


0110011010
0101011001
1001010011
0011001010
1001101011
1010010011
1011100111
0011011010
1110110011
0010111110

I think this is binary for "psnbdt kdjrn t gfd rekjgf kfjed DGD!!!!"

but Im not sure. :D

Let me know
More...

19 AB 94 65 6B 2B 96 DB 37 C<--added 0

=

«”ek+–Û7À
 

Attachments

Question for the OP:

This might be a dumb question, but aren't the odds of any given sequence the same as any other given sequence? So that a sequence of all zeros is just as random as a more "normal looking" mixture of zeros and ones?
 
Quote from trendo:

Question for the OP:

This might be a dumb question, but aren't the odds of any given sequence the same as any other given sequence? So that a sequence of all zeros is just as random as a more "normal looking" mixture of zeros and ones?
More...

Yes. Given 100 binary digits there are 2^100 different possible combinations. Therefore the chance of getting all 0's is 1 in 2^100, not very likely but it can happen. This is why you can't ever be 100% sure that your causal model really explains things or is just coincidence but if presented with odds of ~ 1 in 2^100 of being wrong I would probably trade the model.
 
Within the pool of Elite members, one could argue it is random as to who posts here.

But could one put a program together based on past behavior that would give probably as to who would post here.

So , what is random and what is probablity?
 
You must log in or register to reply here.
Back
Top