Poor AI Performance Explained

Our goal for this feature is to be transparent to users and to instill their trust in our models based on its historical performance. In layman's terms, if the model performance is historically good, chances are it will continuously perform well, and if it is has been historically poor performing, it's probably not a good bet for now.


This article discusses some limitations of our algorithmic models and reasons why you might be experiencing some bad signals. We wrote this in the spirit of transparency. This is meant to educate our users on our predictive technology and develop trust.

This is not financial advice, nor are we registered financial advisors. Everything mentioned here is meant for educational purposes only. You are ultimately responsible for the individual investment decisions you make, and we cannot be held liable for any losses incurred from your investments.

How our algorithms work

Our algorithm works by compiling historical data dating back anywhere from 8-10 years. We trained a machine learning model to analyze the price movement, we use some common indicators such as RSI, Stochastic K, MACD, etc. to aid with our predictions. In layman's terms, we will look at factors such as trend strength, volatility, and a whole lot more.

We stream our data through our partners at Polygon.io and NASDAQ where we apply our trained model and apply it to the data as it is being streamed. We then display the signals for our users to see! This process has anywhere from a 1-5 minute delay.

Why you might see poor model performance

The primary reason behind poor model performance is market irregularities such as chunk buy or sell orders. For example, if somebody buys $20M of TSLA stock, or if somebody dumps $20M of TSLA stock.

It is important to keep in mind that these are human driven events and are unpredictable by even the world's most "accurate" models.

Primarily because stock market algorithms are only meant to react to human behavior not predict it.

How our most successful users leverage our signals

After speaking with a lot of you, we have come up a unique way that our users to leverage our signals properly. Now it is important to note that our users have found many different ways to profit off of Blackhedge, this is just the popular strategy.

The recent buy signals list is extremely valuable, the way we rank it is by looking at where we have seen a recent buy signal and order it by trading volume descending.

IMPORTANT: this is not investment advice, just an explanation of how our most successful users have used our app.