THE SCIENCE OF ACTIVE TRADING JUST MOVED TO A NEW FRONTIER
Machine Learning Predicts Where Markets Are Moving Next
As an Active Trader, you invest in deep analysis to spot trends the more casual observer can’t see.
That discipline and dedication is both entertaining and rewarding: you like the work of uncovering the path to profitable intraday trades.
Your analysis focuses on historical trends, lagging indicators, back testing your theories, so you can then predict if that pattern will happen again. That’s the way you get home runs.
But as an Active Trader, all you have to do is look at your percentage of trading successes over losses to know how effective those labor intensive analytical strategies are.
What is a leading indicator? It’s having and acting on foreknowledge to see the future before it happens. It is NOT the latest MACD twist paraded as a ‘leading’ indicator. See Figure 1 below to see how EOTPRO Developments charts leading indicators.
Figure 1 – The green line is pointing in the direction the Dow is about to move to… sideways. The Stochastic shows the market is still moving up. The arrow is how a true leading indicator works to help the Active Trader select and time trades. This is one of 8 leading indicators.
Figure 2 – A few minutes after the Figure 1 Screen Shot, DeepStreet’s pointing arrow has turned yellow, meaning caution, do not trade. The MACD and the Stochastic are still showing there is upward movement.
Statistics show that most Active Traders make their profits for the year from a few home run intraday trades. But if your capital didn’t stretch far enough to be trading on those few knock-them-out-of-the-park days, your year could turn red fast.
So you love active trading and want more secure, reliable profits.
Then you need leading indicators, so you have a new way to look at the intraday market that might increase your probability of success.
How do you get to see the future before it happens, in the form of a leading indicator? By using a supercomputer and Machine Learning to scan news alerts no one has yet. And no, it’s not the stuff of Hollywood or what scientists used to call Artificial Intelligence.
Before you read further, why should Active Traders care about what’s happening in Machine Learning and Artificial Intelligence (AI) today?
For the simple but profoundly rewarding reason that trading with today’s AI (Machine Learning) gives you a timing edge that some of the big traders have… but you, trading for your own account or small boutique hedge fund, could never afford on your own.
Let’s just plant this idea and then we’ll break it down. What if you had advanced knowledge of where a stock or the market was moving to next, up, down, sideways and at what price it is likely to move to in the next 3 minutes to 3 hours? And you knew the level of confidence of that prediction? This is a whole suite of leading indicators.
Yes, at a glance it would be immediately obvious to you what to trade, how to trade it and what the profit potential was. Would you want to use that tool to increase your chances of better rewards?
50 traders who’ve been road testing these leading indicators inside a platform called DeepStreet EDGE have confirmed it has changed their fortunes.
They didn’t have to spend months reading books to figure it out either.
But what leading indicators are they using and how does it work?
To recognize the significance of DeepStreetEDGE, you will want to know what Artificial Intelligence is and why is it so much more reliable than historical indicators at predicting market movements.
So back in the day when we first started hearing about “AI” data scientists were just using advanced programming tricks. There really wasn’t anything artificial at work, as the programmer supplied all the intelligence to the system by programming the model.
But it sure sounded powerful. Their models did advance the power and capability of a computer to work with lots of data. But not powerful enough to predict the movement of stocks.
That is a really big data problem. You couldn’t assemble the right data in the right format, in one location for AI to effectively query the data.
When you model the question you want answered using the data available, in the old way of developing a solution, you were limited by the mind of the programmer.
That all changed a few years back when data scientists changed how they were thinking about the problem. They decided to Model the Mind, rather than just model the problem in light of the data sources.
Modelling the Mind enables AI to be capable of exploring everything about the world. As long as you provide data to train it so it learns to recognize the patterns you need as it examines all the variables, the machine ‘learns’ and can use the patterns it has recognized to make predictions.
However the AI model based on the world (the old problem mindset) becomes obsolete the moment it is finished: the world has moved on already.
The only hope to create intelligent systems is to have the system itself create and maintain its own model.
Continuously updating output, in response to sensory input. Now we are not limited by a single human mind to program a model.
Enter today’s AI: Machine Learning. It is NOT a subset of AI. It is truly artificial intelligence at work.
Since 2012, Deep Learning, a Machine Learning technique, has made huge strides in sophistication in Image Understanding, Signal Processing, Voice Understanding, and Text Understanding.
Machine Learning doesn’t care about intelligent behaviour: it is focused on the accuracy of its patterns so that it’s output: predictions, are relevant and reliable.
Artificial intelligence cares about maximising the chance of success… not accuracy.
When it comes to predicting the movement of stock prices (leading indicators) you need accuracy you can trust. You don’t want to analyze the world of stocks in its historical form!
That just gives you lagging indicators on steroids and you are back to not knowing anymore about the future than history can predict.
You have to give the Machine the right base to learn from. In simple terms, it means you feed it news, show the machine what happened as a result of the news and increase its confidence that it can predict price movement better with the next piece of news. Then you multiply all that activity by thousands of news items every day.
You also need to aggregate those thousands of changes in price movement of every stock so you can make the prediction as to how all that news will move an entire index.
You are now talking about collecting all the inputs that affect price and movement so that Machine Learning can spot thousands of patterns and turn the patterns into predictions.
Up until now, this was an immense computation problem as a well as a data source problem. It is very difficult to get all the right input feeds that contribute to stock price movement… and house that data in one big location so it can be analyzed.
You can imagine the size of that hard drive.
When you get disparate data sources, in the old problem frame approach, a data scientist had to organize and label all that data before any analysis could be done.
In Machine Learning, using natural language processing, the computer can move from one data source to another without the data needing to be labeled. It learns and broadens its model off of every bit of information… with the correct algorithms.
While the data and natural language processing side of the problem in analyzing stocks has now been conquered, the other side of the problem is where Machine Learning and leading indicator trading, is in its infancy.
How do you turn a Machine Learning pattern recognition prediction system into a stream of information that makes sense to the human in a way that they can immediately understand what it means, and then know what to do with the result?
The User Experience, as this is called, turned into the bigger nut to crack. If you don’t have a graphical, easy to navigate, interpreter for the average human who has no Data Science credentials, you don’t have a way of extracting the value from all that Machine Intelligence.
A few Wall Street companies are now investing in startups to help them grapple with how to pose questions using Machine learning to see if they can predict macro trends for oil, for instance, using many variables.
Despite these investments, the bigger surprise for Wall Street has been how hard it is to find a way to interpret the Machine Learning results in a way that easily reveals what to trade and when to trade.
It’s great that Machine Learning can be trained to give an answer. But if you can’t use that answer confidently to time entry and exit for the right trade, Wall Street traders can’t capitalize on that answer.
It’s really an exercise in futility.
But asking the question of big data has a big ‘cool’ factor, like when Watson, IBM’s supercomputer played chess and finally learned how to win after many games. Nice. But not profitable.
So where are we in terms of being able to use leading indicators to find the right trade before the stock has moved to a new price?
Back to DeepStreetEDGE. EOTPRO Developments has cracked this complex user experience interpretation problem, and is now able to deliver on the promise of Machine Learning… so the answer of what to trade is distilled from vast data into discrete actionable tradable strategies.
Getting to the bottom of this ‘user experience’ problem was no easy feat.
There were two failed attempts under the belt that gave EOTPRO Developments the insight into how to wrestle big data from 45 different sources, discover the patterns, back test predictions to prove accuracy and funnel all that down into a dashboard that presents the best trading opportunities.
Then the founders and their team led by active trader Bill Dennis, went eight steps further.
As Active Traders, the founders didn’t just want to ask big data questions to uncover the sentiment of thousands of news stories. They wanted true predictive leading indicators to use for their own trading. And they wanted to put that power in the hands of people like themselves to give everyone a leg up on the system.
They wanted to give Active Traders an edge, so they did not have to rely on lagging indicators to select and time trades because it’s not a fail safe system. It’s just all that has been available up to now and what hundreds of trading research tools relies on.
You can’t buy what has yet to be built and tested.
Historical trends, (lagging indicators), never got them the consistent, reliable results they wanted.
So Bill and his own team of data scientists and product developers started with the news. The markets move on news. Such predictive sentiment analysis on news and social media feeds is everywhere today.
Yet sentiment analysis does not inspire confidence to know what to trade and what the benefit of the trade will be in advance for a stock never mind an index, intraday.
The EOTPRO founders wanted to know:
• When to trade any stock or option in the Dow or top 100 stocks on Nasdaq.
• When to trade the Dow or Nasdaq futures or ETFs
• What price change to expect from the prediction before stepping into a trade.
• What level of confidence they could expect from the trade, based on the prediction.
• How long it would take to get to that price.
• When to get out.
• When not to trade at all.
Figure 3 – DeepStreet EDGE prediction dashboard
With seven goals achieved, their eighth goal took a year of programming and testing to figure out. Having succeeded where no else has to date, the EOTPRO team took on the last challenge.
There had to be a way to estimate the pace of momentum so that traders can track when the market is about to reverse or retrace AND how slow or furious the pace of that change will be.
When you are trying to do something that no one else has, and you don’t want to be just a little bit better, you want to deliver incredible value, far more easily than the current practice.
Kind of like the difference between adding a battery to a gas powered car as an innovation in energy use vs developing a reliable, attractive electric car that people really want to drive as a revolution in energy use AND the driving experience.
EOTPRO could have focused on using Machine Learning to make historical pattern recognition (lagging indicators) more accurate.
Instead they said “we don’t drive looking in the rear view mirror, we look forward. We want to trade looking forward so we see where we need to go next… and what to avoid!”
Conquering the eighth prediction as to when momentum would stall and the market would reverse, almost didn’t happen.
All the work had been done on DeepStreet EDGE. It was delivering all seven leading indicators flawlessly. The beta subscribers were using it to trade and making better returns.
But the momentum indicator seemed random. It didn’t seem to point out anything in particular but a changing set of numbers from plus 4 to negative 4. As Active Traders love to do, everyone kept watching the patterns, noting what happened next and sharing their insights (The EOTPRO team has a vibrant Active Trader community that meets daily to select their trades using DeepStreet EDGE).
Then one day, the numbers came into focus.
It turns out that the EOTPRO’s algorithms fueled by machine learning had ‘learned’ to tell exactly when momentum and news were going to fade and stop fueling the direction of the Dow.
It just took the humans awhile to notice how the number sequences worked. The EOTPRO machine learning algorithms had already figured it out.
Figure 4 – DeepStreet EDGE dashboard and Dow Prediction and Momentum Indicator Chart
As EOTPRO’s commitment to the user experience was paramount, the numbers were soon turned into easy colored coded symbols that everyone could use to trade confidently.
In the end, CEO Bill Dennis notes that the hardest part of the project was building the visual analytics so that it made simple, easy, at-a-glance, sense to an experienced active trader. Very few active traders are data scientists. So without the combined expertise and collective intelligence of his team and subscribers, this project never would have succeeded.
With the Dow prediction completed, it was just natural the EOTPRO team of data scientists believed they would achieve the momentum prediction for the top 100 stocks in the Nasdaq index, despite it being an equally complex problem.
With the Eighth step completed, they keep improving the Active Trading experience so everyone loves to use leading indicators and will soon have the S&P index prediction conquered as well.
For seasoned traders, using the platform for futures and options trading has been a revelation as the old laggard indicators aren’t part of their daily trading routine unless they need to look up history.
Bill and DeepStreet EDGE’s user community set their sights on the future, following DeepStreet EDGE as it points to the next high, the next low or says it’s the wrong time to be trading.
EOTPRO’s tenacity and persistence is paying off. In fact, when Bill’s wife Shelly had 19 (and counting) profitable futures trades in a row over 3 weeks, based on DeepStreet’s prediction of where to enter and when to exit, they knew they had succeeded well beyond their goals.
That’s how far Artificial Intelligence has come. There is an intelligent machine that can predict what to trade, when to trade and how far a stock will move. DeepStreet EDGE knows where the markets will move next. And that’s all an Active Trader really needs to see, before everyone else does, at a glance.
DeepStreetEDGE is now ready for sophisticated Active Traders to road test for themselves. As a way to introduce it to the industry, EOTPRO is offering active traders a chance to try it out for free for 30 days, and see how to trade with leading indicators by joining the daily trading club.
To get your 30 day complementary pass, just click the link to register.