FAQ

What is Preddix?

Preddix is an information service for investors. It uses sophisticated machine learning algorithms to help investors determine when to buy or sell various Exchange Traded Funds (ETFs).

What is machine learning?

Machine learning involves the use of algorithms that can learn from data and make predictions. Rather than following strictly static instructions, machine learning algorithms dynamically generate statistical models from various inputs and then use them to make data-driven predictions. The algorithms used by Preddix are state-of-the-art machine learning algorithms. They are trained using vast amounts of historical and current market data to learn how a particular fund will to react to market conditions.

How does Preddix works?

At the end of each business day, Preddix runs algorithms to predict if the price of an ETF will go up or down over the next five business days. Preddix then issues a ‘buy’, ‘sell’, or ‘hold’ signal for each of the ETFs it tracks. As a Preddix subscriber, you can then use that information to inform your investments decisions.

What kind of machine learning algorithms do you use?

We use an ensemble of advanced machine learning techniques. Our approach involves the use of multiple learning algorithms to obtain better predictive performance than could be obtained from any individual learning algorithms.

What kind of data do you do you use to feed your algorithms?

We use a variety of data sources and analysis techniques to feed our algorithms, including historical market data, technical indicators, economic indicators, market sentiment indicators, and other quantitative information.

Are your algorithms published or described anywhere?

No. To maintain our competitive advantage, we do not publish our algorithms.

What is the ‘strength’ indicator?

Members have access to a long list of daily trade signals (buy, hold, or sell). With each signal comes a ‘strength’ indicator. It is an estimate of how likely a signal is to be correct based on past algorithmic performance. It can be loosely interpreted as a level of confidence in the trade signal. It does not represent the magnitude of the swing in one direction or the other. Low-volatility ETFs tend to have higher signal strengths than volatile ETFs. Low strength is the main factor in issuing a ‘hold’ signal.

Why are you tracking ETFs only, and not shares of individual companies?

ETFs have become hugely popular investment vehicles and are well-suited to the average investor’s needs. they give investors the diversification of an index fund and, in some cases, the ability to sell short or buy on margin. Though the algorithms used by Preddix are currently optimized to make predictions on equity ETFs, we plan to extend them to other types of ETFs and shares of individual companies in the future.

Why aren’t you tracking some ETFs?

There are various reasons why we may not track a particular ETF:
• The ETF is fairly new and there is not enough historical data to properly train and validate our prediction algorithms. We usually prefer to have at least five years of historical data to do training, testing, and validation of our algorithms.
• The ETF composition is not well suited to our prediction algorithms. For example, the ETF may track bonds instead of equities.
• The ETF is a small fund that we haven’t started to track yet. Contact us if there is an ETF you would like us to track.

Why aren’t you tracking inverse of leveraged ETFs?

In many cases, such ETFs are only a few years old and thus do not meet our tracking criteria yet. However, there is a good chance that we track the index ETF that the leveraged/inverse ETF is related to.

Are you paid by some ETF companies to track their funds?

No. We are not paid by any company. We are completely independant.

When are predictions made?

Predictions are made at the end of each business day. They are normally posted on the website before 3AM Eastern time.

What’s the time frame of each prediction?

Predictions are made with an horizon of five business days.

Do you offer automated trading?

No. Preddix is purely an information service. We are not brokers or investment advisers and we do not place trades on behalf of our subscribers.

I want to actively trade based on Preddix results. How many trades per year should I expect to make for a given ETF?

It depends. Based on historical data, somewhere between 10 and 20 trades may be required per ETF. Trading based on signals published by Preddix should be done with caution and you should consult with an investment adviser before doing so.

I’m not an active trader. How can Preddix be useful to me?

Even if your preferred strategy is to ‘buy and hold’ ETFs, you still have to buy shares from time to time, and you will need to place at least one trade per fund you will ultimately sell. If you make these trades on the wrong days, you may lose a few percents on your investment, which can represent a very large sum over many years. The signals published on Preddix can help you reduce the risk of such loss. Again, trading based on signals published here should be done with caution and you should consult with an investment adviser before doing so.