How a Video Game Algorithm can Beat the Markets

Corona Del Mar, CA

Howdy Friend!

When I was waiting in line at Disneyland, I noticed the kids were playing a guessing game app to pass the time.

It's called Akinator.

Here's how it works:

The game begins with the player thinking of a character, either real or fictional.

The Akinator app then begins asking the player a series of 20 yes-or-no questions about the character.

The questions may include things like the character's gender, age, occupation, physical appearance, and personality traits.

Based on the player's answers to these questions, the app uses an algorithm to eliminate certain possibilities and narrow down the list of potential characters that the player is thinking of.

After a certain number of questions, the app will make a guess as to the character that the player is thinking of.

If the app's guess is incorrect, the player can indicate this and continue answering questions until the app is able to correctly guess the character.

Asking 20 yes or no questions allows the app to go from 2^20 possibilities (1,048,576) to 1 (the correct answer).

Narrowing it down, but not too much

A trading strategy allows you to ask a few questions, and have the computer spit out a simple "buy" or "sell" signal.

Each question is a trading rule.

The simpler the better.

Why? Well, if you have 20 trading rules, you'll rarely see a trade.

I remember watching the late Kevin Samuels, a dating advice YouTuber, tell single women that their checklist consisting of men over 6ft tall, making over $200k, under 40, and fit would mean they would never find the guy they were looking for.

They had better dial it back or get comfortable with cats.

If you're looking for the perfect trade, it rarely appears. And that of course means it won't pass the statistical significance tests.

One of the simple rules from my new book, Outfoxing Wall Street, is comparing today's price of the S&P 500 to 80 trading days ago (about 4 months).

Is today's price greater than 80 days ago? Yes, then it's a bull market? No, then it's a bear market and it's time to use a different approach (mean reversion).

Simping for Gartley Patterns

It's funny because many many years ago I decided to program a "Gartley Pattern" into the computer.

According to a trading academy website I found on Google: "The Gartley pattern is in close relation to the Fibonacci numbers. Gartley patterns are considered high-profitability patterns with a success rate of over 70%"

I have high confidence that "70%" number is about as real as Bernie Madoff's trading gains.

Just this one pattern has 8 rules. So of course when you feed all those rules into a computer, it barely ever finds 'em. Even when you relax the rules and give them some room for error, they hardly ever show up.

And backtests of those that did show up didn't find any alpha.

We spent 18 months on a project that identified ALL possible price patterns. It absolutely doesn't work.

So I find this meme appropriate on a topic I spent mountains of time and hundreds of thousands of dollars on:

More from me soon!

Trade smart,

Dan "Prince of Proof" Murphy

Disclaimer: The results listed herein are based on hypothetical trades. Plainly speaking, these trades were not actually executed. Hypothetical or simulated performance results have certain inherent limitations. Unlike an actual performance record, simulated results do not represent actual trading. Also, since the trades have not actually been executed, the results may have under (or over) compensated for the impact, if any, of certain market factors such as lack of liquidity. You may have done better or worse than the results portrayed.