What is the FatKelly?

The FatKelly answers the most important question for the Fractional Social Trader of today: How much should I buy? It is simple, and provably makes betting/investing better. Don’t take just our word for it: Warren BuffettCharlie MungerMohnish Pabrai, and Bill Gross, and other impressive financial types love it too.

How does it work?

The Kelly criterion is a time-tested formula for sizing bets or investments. The idea is simple, and has been called Fortune’s Formula: Given a sum of money to invest, allocate it in proportion to your edge/odds (The edge is the amount of gain Vs loss. The odds, the chances of a gain Vs loss). 

The FatKelly, is our extension of the traditional Kelly criterion. With two modifications:
  1. We use Options data implied win/loss probabilities and expected returns, and
  2. Make fractional adjustment to preserve dry powder for opportunities ahead 

How do I use it in my trading?

It’s a three step process:
  1. Consider how much money you have to invest (the purse/cash you’re willing to risk without FDIC insurance)
  2. Source and screen available investments/bets on the basis of edge and odds (Surprise: Social-media and Options data do help)
  3. Pick the ones you like and use some type of modified Kelly to size the bet given your purse
Check out the 💎 Oracle if you like. It is a tool we made for ourselves and friends to make these steps easier.

Where can I learn more? (References)

There is some math, but it is surprisingly simple. Examples and some simple formulas can be found in the references below and the others they cite. 
  1. The Kelly Criterion: You Don’t Know the Half of It, CFA Institute
  2. Option Prices Imply A Probability Distribution, Global Capital
  3. Understanding the Kelly Capital Growth Investment Strategy, CAIA

The 💎 Oracle Design Principles: SAVER

Simple - to understand and explain
Actionable - to use and test
Value - for users. Saves money/time/effort/pain
Enthusiastic - energize with exciting possibilities
Reliable - always works and signals are robust