|
Future of investing does not lie in you trading yourself. You will be replaced by AI. Question is if you will be making money on it or not.
Problem
Imagine having an investment strategy constantly delivering high returns based on simple yet powerful ideas. Imagine having no cognitive biases affecting your investing. It is probably exhilarating and even though many can imagine and try they are not able to deliver the returns expected. Even if they try to follow moat investing as probably the only strategy proven to constantly provide high returns, they do not possess the will to believe in the idea during the hard times. This results in loss followed by trying to look for next best thing. Imagine now having an artificial intelligence based on the correct ideas and having no drawbacks of human mind as fear of losing or happiness of winning. It will rationally follow given strategy and will bring high returns. You no longer will have to imagine yourself investing successfully because you will be successful investor. Our goal is to provide artificial intelligence so this vision will become true for you.
Currently many investors invest in index funds as it is their best way to multiply their money investing. Active investment funds are shunned as too expensive because of high fees and lack of returns when compared to index and they see massive outflow of capital. Even if someone claims to invest based on difference between price and the value not many do. It is true that there are only few funds able to achieve returns higher than index long term and therefore current state of affairs it is not surprising. There is however one forgotten way to ‘beat the market’ and it is value investing.
Solution and Product and Why it Needs to be Solved
We aim to create artificial intelligence for investing based on neural network technology following value investing principles enriched by idea of durable competitive advantage. As followers of Graham and Buffett and professional programmers we see the possibility of merging the two disciplines into one as the biggest opportunity of our time. In the past we would invest and develop software separately but today, as a Buffett would say, we can stand on the shoulders of giants and build a platform bringing high returns to everyone willing to spend five minutes pondering if what we present makes sense.
Moat investing is based on three simple but great ideas:
1. Durable Competitive Edge (moat)
2. Great Management
3. Value higher than price
1. Competitive Edge
Competitive edge also known as economic moat is inherent characteristic of a business keeping competitors from taking the clients of a company. In the best case scenario competitors are not even able to enter the field to compete. This allows company to charge high fees with the client not willing to stop using the services provided. There are many competitive edges but five main are –
1. Switching Costs - client would have to invest time or pay for the change of a provider.
2. Efficiency Scale - there is only one company needed for the product to be cheap - for example it would be costly and inefficient to build more than one water or oil pipe in a given area.
3. Networking Effect - the more people use the product the more it becomes useful to them - ebay, facebook.
4. Cost Advantage - company is able to create products cheaper than its competitors and it is not easily copied method.
5. Intangible Assets - patents, brand name (Coca Cola, Google…) , licenses
Even if statistical models proved to have high returns on the data in the past they had one big flaw - inability to grasp the idea of economic moat. It is not always possible to explain the concept with numbers only and it might result in buying a business with only temporarily good returns. Current way of building investment AI is again mainly based on numbers as it is hard to get software fully understand written text. We plan to change this and deliver software able to recognize which of the provided edge, business possesses and reasons why.
We propose feeding the neural network annual reports of companies with competitive edge and let it recognize patterns in them. Maybe it will be higher usage of given word or it could be whole sentences that will be used, we do not know, but it is exciting to think about the possibilities. There are many sources of information like media, social networks or even advertisements that could be used as well. For example high frequency of media exposure in comparison to another brand could mean trouble in the short term but if the pattern is constant it could mean that people are more interested in the products of the company. Let our AI decide that for us.
2. Great Management
It is great to have a company with high returns on capital and with happy customers but if you have a management that spends investor’s money on unintelligent business ideas, or even worse on themselves. It is definitely not worth as much as quality business with highly intelligent management. In a world of constant mergers and acquisition it is important to realize that growth like this comes at a price - opportunity cost. If you buy business with lower returns on your capital you automatically lower your returns and instead of 20% of invested money per annum you receive 10%. When presented in this manner it seems obvious and you could think only dumb person could do it. But sadly it happens all the time. It is even worse if management pays high wages to themselves even when the company is constantly losing money under their reins.
With neural networks it will be much easier to see if the management always does the correct thing than when human does the job. AI is not biased and cannot be tricked by false claims or by false kindness. Using proxy statements it will be easy to compare returns and wages of the management. Also if data are presented, it will be easier to sift through all acquisitions to find out if the returns grew up due to the action or if inactivity would be better.
Next step where management can go wrong are share repurchases. If management buys their own stock even when it is highly overvalued and taunts how cheap it is you can sense incorrect motivation. Maybe they own stock options soon to be exercised or the company has high debt with a caveat of given market capitalization and they need to keep the price of the stock high. Or they are in love with their company and see potential as much higher than it truly is. In some cases they just cannot differentiate overvalued and undervalued opportunities. In any case it is better to find a company buying their own stock at price cheaper than its internal value.
It is hard to pinpoint exact value of a company and it is highly subjective so this will present a challenge for AI as well. We believe that using metrics as price to cash flow or price to book value will help our software recognize cheap and expensive. As the project will be unveiling maybe there is some new idea to be discovered how to value the company. Currently our best way is to discount future cash flows to today value and that is what we will try to teach our AI to do.
3. Value higher than price
The discussion before gave us some ground on which to build. We said that value is the present value of discounted future cash flows. Of course it is hard to predict future cash flows unless you possess clairvoyant abilities. What makes this exercise easier is durable competitive advantage. Due to constant buying and inability to easily leave the company’s product there is higher probability of constant returns. Even with this advantage it is hard to exactly say what the real value is.
With the given assumption of our inability to pinpoint the value we need to find some other way to make sure we do not pay more than the value. Solution is to buy only companies that are so cheap that it hits you. Ben Graham presented idea of margin of safety in his book Intelligent Investor and that is exactly what keeps value and moat investors safe today. Let’s assume the value you accounted for is $100. It might seem logical that at this price it is a buy opportunity. In a reality it is a sell opportunity unless a company comes up with a plan how to grow. You want the price to be much lower so even during turbulent times you do not lose money. It also gives you opportunity to gain the difference between the price and value as in the long run market will appraise the company at its value.
Software we are developing will have more than one way of accounting for a value so instead of one strictly given value it should be range of them. Unless the price is significantly lower it will not be a buying opportunity. Of course there might be differences in industries as to which metric best presents the value. For example for banks to sell under their book value means that the market gives you opportunity to buy 1 illiquid dollar for 50 liquid pence and even if future cash flows do not seem to be high if the bank was liquidated you would most likely get more than 50p invested. Feeding the cash flows, book values and earnings will be the most vital for this part to be successful.
Token Implementation
Token will be divided into two theoretical parts.
First part is an asset with underlying investment in hedge fund. Fund will be using value investing strategy to buy tokens and shares and will short sell obvious frauds. Token will therefore always will be worth at least the value of the underlying assets. We plan to achieve at least 20% yearly returns with this type of investing.
Second part will be right to use Value Investing AI when the product is created. After creation of AI all funds will be managed by it with little or no interference from us. Our goal is to achieve 100% yearly return using this strategy.
We must warn you that numbers are goals and there is probability of not achieving it as investing is always risky.
Team
We are willing to work on every issue until it’s solved. If we lack some skills, we are working on gaining them as fast as possible. Our goal is to use every smart mind we can find to temporarily provide us with their knowledge. We do not believe in big teams that are just wasting money on PR activities and marketing. We want to keep the costs to minimum and therefore we do not want to pay for unnecessary staff and space for them. We might be proven wrong one day but currently this is the strategy we see as best fit for execution of the plan given above.
Token Deployment and Plan
Token is divided into two parts. One part presents ownership of underlying assets. We will invest in tokens, shares and debt. Second part presents right to have the underlying assets managed by AI we will create. Token will be standard ERC20 format.
Money Usage
-data - compustat, bloomberg, quandl and alike
-newspapers subscriptions for evaluating competitive edge
-temporary employees (university professors and people currently searching for opportunities)
-our own wages
-hardware, software, server needed
|
Show Posts
