Many investors are turning to the Kelly Criterion for deciding investment size. Seldom, most of them never have walked through the inherent mathematical derivation or read Ed Thorp on the best way to use the Kelly Criterion in the stock exchange.
Kelly Criterion Stock Trading Example
Firm A is now studying 3 new products that are distinct. In a upcoming custom, we realize that A might declare the start of the newest products of one. We also can estimate the effect of consequences that are different on the stock price:
- 30% increase in the stock price of A if Product 1 is found. There are 20% probability with this to happen.
- If Product 2 is found 10% increase in A’s stock price. There are 15% probability because of this to happen.
- If Product 1 is found, 12% increase in A’s stock price. There are 25% probability with this to happen.
- If no product is found, 15% decline in A’s stock price. There are 40% probability with this to happen.
Now you have $100 dollars in your bankroll, just how much could you spend on the stock of A in order for your bankroll to have the maximum increase in the future?
The Kelly Criterion cannot assist you to solve this issue as it presumes only two potential results: negative or favorable. In addition, it presumes when the result is negative, you may lose 100% of that which you invested.
In the stock exchange, you frequently have multiple consequence scenarios, and you also practically never lose 100% of your investment within a trade. Thus, the Kelly Criterion isn’t directly related to the stock exchange.
Defining Kelly Criterion Variables
Let us define some variables:
F = % of your bankroll which you spend money on A
Of Introduction Product 1 = 30% W1 = ROI
Of Introduction Product 2 = 10% W2 = ROI
Of Introduction Merchandise 3 = 12%, W3 = ROI
W4 = ROI of No Products Found = -15%
P1 = Likelihood of Product 1 Introduction = 20%
P2 = Likelihood of Product 2 Introduction = 15%
P3 = Likelihood of Merchandise 3 Introduction = 25%
B = First Bankroll
Utilizing the aforementioned info, we are able to invent: We can locate the utmost M by choosing the most Ln(M): P2*Ln(1 W2*F) P3*Ln(1 W3*F) P3*Ln(1 W3*F)