Predicting the markets.

If one could predict the stock market, forex market or crypto market one would become very rich in a matter of hours. This fact has led many to study the possibilities to do so.

image source

Fundamental analysis
Of course one should always first look at the fundamentals. What would be a rational value of the product or service, how is the demand, what are the risks, competition etc.
Are these fundamentals reflected correctly in the current price?
Currently we have the example of IOTA in the cryptosphere. It is something new and thus it may have great potential. But on its day of introduction it ended up in the top 10 crypto currencies, among BTC, ETH, LTC to name a few. These currencies have long track records and are built on proven technology. Honestly, IOTA may be good, but it certainly does not belong there yet. It still has a lot to prove. And based on just this fundamental observation I expect the price to go lower for the remainder of this year. If all goes well it may come back in the top 10 after 1 or 2 years.

Technical analysis
After we have looked at the fundamentals we look at the price performance against peers. Many techniques have been developped, such as candle-stick patterns, Elliot Wave etc. But one should first ask the question “Can the market be predicted?”.
To answer that question we look at how prices are made, and to make a long story short we can establish that it is largely a feedback mechanism; People do something, that changes the prices, people see the changes and respond to that.
A feedback system can be predicted as long as we know all the parameters. Some parameters we can not easily predict, such as wars, outbreak of diseases, social unrest and those sort of things, but apart from that markets should be relatively predictable based on historical price movements and volume.

Puting it to the test
Using simple pattern recognition algorithms in an AI type setting, I wrote a program to analyse well over a 100 years of DJIA data and predict future values. The program was very simple -a sort of brute force pattern recognition- and used a lot of CPU, but after a little tweaking it worked.
So why am I not insanely rich right now?
The problem I ran into was that every time after a few correct predictions there came a point where prediction and reality diverted. One day I was lucky enough to monitor the gold market just at such a moment and I saw a ridiculously large gold transaction. I forgot the exact amount, but I was working in the forex markets at that time so I knew what numbers are ‘normal’ and this gold transaction absolutely did not qualify as normal. What made it even more suspicious is that it was repeated exactly 12 hours later.
So I started investigating these transactions and found more, sometimes they repeated up to 3 times.
My conclusion was that some large institution, most probably very close to the clearing bank (which is a FED-bank, btw), if not the clearing bank itself was doing these transactions to manipulate the markets.
Therefore the stock markets are not predictable, unless you know what the manipulators have in mind.

The sub-prime crash?
An interesting side note here would be that my program accurately predicted the 2007/2008 crash. Which makes me wonder if the reasons for this crash are indeed what we are told they are, or perhaps something much more fundamental. But that goes beyond the scope of this post.

@furion
I recently read a post by furion in which he used python to analyse the crypto markets. This brought back the memories of my experiment and I wondered if there would be sufficient data available to use my program here. So I looked for sources of historical data and found a few, but I’m afraid that this market is just too young for my program. Of course I could add many altcoins but the fundamentals of those coins are incompatible, also in the first couple of years there was only a couple of nerdy BTC users, which is an entirely different situation now.
So we’ll have to wait and then, when we have sufficient historical data, the markets will probably be manipulated again...