How Random Is Random? A Motivation Behind Data Science.
Being a data science enthusiast we come across hundreds of problems from random normal distributions to random poisson distribution, from random sampling to random outcomes and while studying so much of randomness throughout the course I realised that randomness has been a part of our curriculum since the very beginning of life. It goes from an arbitrary numbers in a geometric plane to a random number in a magicians play and this unpredictability in the surrounding excites me to think if randomness is truly that random or is it just that math we are not yet ready for.
Keeping this idea in mind I start to think of a few random instances taking place around me and if there is a way to predict the course those cases would take. As we roll our eyes going through this article a few thousands of people made some money on their stock while it wasn’t a lucky day for the rest of the investors and keeping these sudden ups and down of a stocks price that never fails to confuse me, I wonder if anyone could predict the price of a share, ahead of time or was it just too unsystematic to anticipate.
To my understanding the trend of a stock in the market is an indication of combined sentiment from people around the world, for that particular company. Therefore, to predict a stock price with precision would require an analysis of brain of every individual on this planet for an accurate forecast, but fortunately that’s not the game. The investors in stock market and public in general are more concerned about the shift in the prices rather than the prices itself which reduces the complexity of the problem but still does not solve the problem in hand. Now, to make the problem comparatively less complex, instead of talking about the shift, it makes more sense to talk about the probability of the shift in a stock. Simple ? Not yet ! This “less complex problem” in not that straightforward which is evident from the number of investors not being able to predict the shift in the right direction. It further requires an intense analysis of large variety of variables that play a role in the change directly or otherwise.
In layman language the price of the stock is determined by supply and demand of it’s share in the market for the most part, where an increased demand increasing the price of the stalk and vice versa, but the supply and demand are dependent on government policies, global economy and most importantly the speculation of a share by individuals. Even supposing it was easy to predict the impact of changes in government policies and economies on a stock price, the gamble becomes more interesting when speculation of an individual comes into play. A wide range of variables play a role in public opinion about a stock such as, company’s performance, past decisions by the board, news release around them and many more.
Although there are hundreds if not thousands of such factors that have significant role on a stock price still a number of reports released by organisations makes it a doable task to predict the shift of a stock in the future with some probability of success. Where reports like Consumer Confidence Index tells us about the optimism in the public with regards to the economy which has a positive effect on the stock shift, surveys like Business Outlook Surveys delineate manufacturers side of the story. A combination of similar surveys and reports from public interest organisations makes the job relatively more predictable and hence it’s just a matter of exploring more such variables that could make the prediction more accurate. Therefore, I believe predicting a shift in a stock price in not so random after-all. By using various analyses we can measure the overall sentiment of the stock and predict the direction of the trend to some extent.
Another experiment which is generally portrayed as random almost every time is the roll of a die. Someone blessed tends to throw a higher number than someone who’s out of luck, or so it is believed. In my opinion if that was the case, the casinos would be losing almost half of its fortune to the chances that may not be in their favour, but as everyone knows, the house never loses. Then how do they manage to win? I believe a bias can be introduced to an impartial die to alter the outcome with some certainty. If you would have noticed “some certainty” in the preceding line, your guess to the question above would be legit. By introducing some certainty or probability associated with the house to win, the casinos even though can lose money in the short run but eventually end up making millions of dollars annually in the long run. The question still remain, how is it done ?
Some speculation in this area enlightened me with a few ways to do it, out of which the one that caught my attention was to introduce a bias by choosing a different material for the pips ( white dots ) than that of the die itself. This simple technique would make the side with a greater number heavier than the smaller number. Therefore the die, even though looks quite normal, tends to show a smaller numbers more than the greater numbers. Although this could be realised by the customer if he/she stays on the table for too long, but plays the trick perfectly if the gamble on the table is for the short duration.
Now, thinking of it more closely, what factors decide the face of the die is nothing but how the die is thrown and more importantly how it hits the table immediately after. The way it lands on the table decides the fate of the die for the next few jumps and finally of the concluding face. Hence, factors like symmetry of the die and density of the faces in a die can change the result of the die with some probability. It’s just a matter of exploring how much does each factor effect the result and the details of the die when it hits a surface that decides the value of it on every throw. Since it is not easy to map those conditions, we tend to believe that the outcome of the throw is random otherwise the calculations could predict it all.
In the nutshell, I believe it’s just a matter of exploring variables that decides the outcome of an experiment and nothing in this world is truly random. Assuming we take into account any factor that has an impact on the experiment, the probability of predicting the result would get better and hence the so called randomness would decrease.
As Stephen Hawkings used to say, the world is governed by laws, and laws at any point are not random.
