Stock Market Distributions again

THIS MAY BE FLAWED

I’ve returned to this problem again. Working out the expected distribution of votes at random in the Eurovision song contest (which will be biased because countries get to vote against themselves!) I noticed a similarity with the stacked charts I had produced for stock market data (simply ordering the raw data). In the Eurovision the game is simply dropping in one of the 10 scores to the pigeon hole of your chosen acts and then adding the scores up at the end. A variation I used for this stock market simulation was repeatedly adding a line of bricks of random length randomly along to a wall. The result is very like a stock market chart but has a more normal distribution (essentially it creates a random walk – this remains to be proven). However by fortune I made a mistake and forgot to reset the array I was using to hold the daily change data before producing the frequency histogram. The result was that over time the array became very large and the distribution narrowed and became more like the stock distribution.

The area of the green normal is the same as the blue FTSE data but the size of the peak and the squeezing is very pronounced.

 

 

 

 

 

 

 

 

 

 

Why this particular game should lead to this distribution I cannot explain at the moment. A similar game of repeating a unit with a percentage of Gaussian random changes produces the expected result that only certain events are produced more often depending upon when they enter the unit and how long they survive in the repeating unit. Need to look at again.