Regular users of ORTEX will be familiar with our Alpha Trading Signals; they have highlighted valuable opportunities for a few years now. 2020 was a turbulent year, and we thought it might be a good time, a year after the markets reacted to the coronavirus, to review the performance of our signals.
“Our 5 star EPS trading signals returned 2230% over 4 years”
ORTEX’ approach to trading signals is rather cautious, markets can be volatile, and history doesn’t always repeat itself. Hence, we have a two-pronged approach to our signals. We start by statistically determining rare events, and on top of this we use a machine learning algorithm to determine the likelihood of success. This can all be seen from the data we display on our Alpha Trading Signals.
We will start with a little explainer of the different trading signals we currently provide. If you are already familiar with them, please jump to the performance section, where we have a look back at the performance a portfolio has had if it had followed our signals.
For each signal we tell you how many times the signal has triggered in the past, the success rate and the average return, as well as what holding period has previously given the best return. We also help you visualize this with star ratings, success bars and charts.
Types of Trading Signals
ORTEX has five families of trading signals: Short Squeeze, EPS, RSI, MACD and Company Events. The different types of signal look at different aspects of the metric that is the main base for that signal.
For example, the Short Squeeze signal comes in Type1, 2 and 3. They look at slightly different aspects of how to detect a Short Squeeze, considering everything from Short Interest, Utilization, Cost to Borrow and the Stock Price as well as the rate of change in these. You might see all three of them triggering at the same time, as we did on GameStop, NYSE:GME, just before the epic price moves at the end of January.
The small icons on the image, 3 stacked on top of each other in three places are the three types of Short Squeeze signals firing on 3 different days before things really took off.
After the first Squeeze, we detected two more Short Squeeze events in GameStop as the price continued to climb, however on these occasions only one of the Short Squeeze signals triggered.
Below is the signal performance chart for a recent signal. This shows aggregated data for all signals of the same type for a stock. At the far left of the chart is the date of the signal, and at the furthest-right is 90 trading-days after the signal, or 30 days depending on the trading signal family. The colored lines are the median and the 25th and 75th percentiles. The gray lines are the most recent signals of this type, and we can clearly see that these have had a strong trend, reaching heights of 100% return on some occasions, as a median we have reached 40% return after 47 days.
To evaluate the performance of trading signals, you can look at a few different statistics: how often is it profit making? How much does the average signal return? Or how much would following the signals return? ORTEX provides you with many different signals and a significant number of statistics about them to enable you to pick which signals to trade. The options are endless. Below we have made some simple decisions regarding what to trade on, and we have backtested this over the last 4 years. For each example, we will outline exactly how we picked these, for you to replicate if you so wish.
Please bear in mind, the last 4 years have seen some extraordinary market conditions, something you might not expect a machine learning trading signal generator would be able to handle.
The Casual Trader
The parameters for this signal selection were set for someone that does not trade all the time, and does not have the opportunity to monitor the markets every second of every day. So fewer trades, with more profit is a good idea. We are just going to trade on the absolute best signals, the 5 stars, or to be precise, everything with a star rating of over 98. A typical casual trader typically only holds a long position, and only trades US stocks.
Over the last 4 years, that actually ends up being 1107 signals. We thought that was probably too many, so we decided to focus on our strongest signal family, EPS, which reduces it to 338 trades, or 1-2 trades a week over four years.
We have to decide how much money to put on each trade as well, so to simplify things here we have just decided to invest 25%, a quarter of the cash we have, on each trade. So if we start with $100, the first trade we do we will invest $25, the next one, providing we haven’t closed the first position yet, will be $18.75 (= $75 * 0.25) and so on.
This is how it would have looked.
That’s a return of 2230% over 4 years!
Of course, that is compounded over 4 years, hence it goes up more quickly at the end. But still!
Year by year the return was:
2017 – 67% on 95 trades
2018 – 21% on 83 trades
2019 – 226% on 46 trades
2020 – 118% on 74 trades
(We only consider positions opened and closed in the same year, hence the sum of the trades here is 298 and not 338, 40 trades are triggered in one year and would have been closed in the next year)
The Active Trader
Okay, so you love the markets and you want to trade more frequently. No problem! Let’s consider all our 5 star (90+ significance) signals. That’s 6078 signals over the last 4 years, or 29 a week. With so many signals, we lower the percentage of cash we invest in each trade. Let’s say 10% of available cash in each trade.
A much smoother line, as we are running more positions side by side. With a more diversified portfolio we carry less risk, and still a great return, achieving 1090% over 4 years.
Year by year the return was:
2017 – 76% on 1268 trades
2018 – 32% on 1530 trades
2019 – 75% on 1196 traded
2020 – 142% on 1596 trades
For simplicity, these back tests and trading simulations have been done without any stop-losses or take-profit strategies, and solely relied on the recommended holding period. These profits, and the risks, can be greatly improved by using stop-losses and take-profit strategies. Please also bear in mind, that any signal can be wrong and, because they are based on statistics and machine learning, events that are outside of the scope of data they consider, they can trigger for the wrong reasons.
To see all our trading signals – start a free trial of the ORTEX platform