In anticipation of our AMA with Chainlink tomorrow we wanted to circulate some preliminary information to get the attention of the developer community.

In a nutshell, what we do at QDT is use machine learning to forecast financial instruments, like stocks, commodities and currencies, for large enterprise customers. That said, our team is full of crypto enthusiasts, so it only makes sense that we would build models for crypto on the side and use them ourselves.

Using Quantum ML, we have built a series of APIs for the Chainlink developer community that will allow you to interact with and use these same Bitcoin signals.

Our Bitcoin machine learning models use market price data from all cryptocurrencies, with blockchain and sentiment data (totalling 3000+ predictors), and then use multiple algorithms and machine learning parameters to find the best signals.

Those signals are then converted into a strategy, so users can answer the ultimate question,

“if I had traded these signals, would I have beaten the market?”

Figure 1: Bitcoin (BTC) Composite Signal Performance using Quantum ML

Above is the historical performance of three strategies, each using a slightly different logic to trade the same signals. The details of each strategy and their performance is summarized in the table below.

Strategy Name	Colour	2yr Return (%)	Alpha (%)
Bitcoin (Buy and Hold)	Orange	633.03	na
Binary Long Strategy	Red	988.42	+56.15%
Simple Fractional Strategy	Cyan	1050.55	+65.95%
Fractional Long Strategy	Blue	1100.99	+73.92%

We can get into the nuances of the signal aggregation methodology another time, but what we’re hoping will get your attention is the performance in the far right column.

Thus, while compliance dictates that we must remind you that “historical performance isn’t an indication of future performance,” we at QDT have been trading the signal ourselves for over a year and have been very happy with the results.

We hope this gets your attention and that you’ll join us for the AMA tomorrow, November 9th at 10am EST to learn more about how we do what we do, and how YOU can use and experiment with these Bitcoin signals yourself.

See you there,

Sean Bockhold
Produce Owner
Quantum Data Technologies

Join the LIVE AMA here:

About Chainlink

Chainlink is the industry standard oracle network for powering hybrid smart contracts. Chainlink Decentralized Oracle Networks provide developers with the largest collection of high-quality data sources and secure off-chain computations to expand the capabilities of smart contracts on any blockchain. Managed by a global, decentralized community, Chainlink currently secures billions of dollars in value for smart contracts across decentralized finance (DeFi), insurance, gaming, and other major industries.

Chainlink is trusted by hundreds of organizations, from global enterprises to projects at the forefront of the blockchain economy, to deliver definitive truth via secure, reliable oracle networks. To learn more about Chainlink, visit chain.link and subscribe to the Chainlink newsletter. To understand the full vision of the Chainlink Network, read the Chainlink 2.0 whitepaper. Want to discuss an integration? Talk to an expert.

Solutions | Docs | Twitter | Discord | Reddit | YouTube | Telegram | GitHub 

About QDT

Quantum Data Technologies is an engineering company with a passion for using advanced technology to solve complex business problems. We are fascinated with the pursuit of innovation and to constantly push the boundaries of what’s possible with AI, Big Data, and Machine Learning. Our core product, Quantum ML, is a data platform that uses machine learning to predict the movement of financial instruments such as stocks, bonds, indices, fiat currencies, cryptocurrencies, commodities, ETFs, and more. Quantum ML utilizes data from 150,000+ unique sources such as global exchanges, financials, macroeconomics, and fundamentals, to automatically generate and backtest ML-driven strategies that inform financial and trading decisions.

To learn more about QDT visit our website or contact us at info@qdt.ai.

Website | LinkedIn | Twitter | Blog

Financial Disclaimer

Quantum Data Technologies Ltd. nor its employees or contractors are investment advisors or brokers, and are not registered as such.  All opinions expressed by Quantum Data Technologies Ltd., its employees and contractors are solely their opinions and neither reflect the opinions nor are made on behalf of Chainlink or any of its respective affiliates. The trading and investing ideas and strategies discussed are not recommendations to buy or sell any security and are not intended to provide any investment advice and/or recommendations of any kind, but are made available solely for educational and informational purposes. Investments or strategies mentioned may not be suitable for your particular investment objectives, financial situation or needs. Any opinions, including that of Quantum Data Technologies Ltd., its employees or contractors are based upon information they consider reliable, but no representation or warranty is made by any person as to its completeness or accuracy.
We are obligated to inform you that past performance is not an indicator of future results.
This is not an offer to sell or a solicitation of an offer to buy any securities. Before acting on information provided, you should consider whether it is suitable for your particular circumstances and strongly consider seeking advice from your own financial or investment advisor. Any investment in any security is subject to a number of risks, and that discussions of any security will not identify or describe all or, in some cases, any relevant risk factors. You should be aware of the real risk of loss in following any strategy or investment discussed in this broadcast, including the acquisition of any security.

Our approach when designing and implementing our technology is inspired by, and in many ways, builds upon foundational concepts outlined in the following books:

• “Superforecasting, The Art and Science of Prediction,” by Philip Tetlock and Dan Garder
• “The Beginning to Infinity, Explanations that Transform the World,” by David Deutsch
• “The Wisdom of Crowds,” by James Surowiecki 

Concepts and Ideas

These books offer a wealth of information and wisdom when it comes to forecasting, but there are three main concepts that inform our solution design that we want to highlight as an introduction to our conceptual methodology:

1. In Superforecasting, Tetlock summarizes his findings by defining two broad groups of experts: hedgehogs and foxes. Those who displayed poorer superforecasting skills (hedgehogs) tended to organize their thinking around Big Ideas. They sought to squeeze complex problems into preferred cause-effect templates. The other group consisted of more pragmatic experts; these foxes, “gathered as much information from as many sources as they could. They talked about possibilities and probabilities, not certainties, and they readily admitted when they were wrong and changed their minds” Foxes are better at forecasting than hedgehogs.

2. In the Beginning of Infinity, by David Deutsch, he argues the need to constantly adapt to changing inputs to make precise predictions. This is reinforced by Tetlock in Superforecasting, where he introduces the concept of “perpetual beta”, which acknowledges and reinforces the benefit of constant reevaluation, updating and self-improvement. 

3. In Wisdom of Crowds, Surowiecki shows that with the aggregation of information in groups, it results in decisions that are often better than could have been made by any single member of the group.

Solution Design and Technology

These three ideas go a long way to introducing the conceptual methodology behind our platform. They are therefore reflected in our solution design and technology, firstly in the form of our rapid-prototyping solution architecture; secondly, through our approach to data, algorithm and parameter exploration via the Auto-series feature; and thirdly through our UI, in the form of the Group Dashboard, which ensembles and stacks the models into an aggregate output visualization.

These three aspects of our technology in many ways underpin the success of our platform insofar as they facilitate a dynamic process of continuous improvement via exploration, assessment and adaptation. A process that isn’t focused on one “Big Idea” (hedgehogs), but explores thousands of permutations and combinations of data, algorithm selection and parameter possibilities (foxes), and then aggregates the individual model outputs into a challenger/champion framework (wisdom of crowds). We then revisit each hypothesis periodically to adapt to changing conditions and inputs, where winners are reselected based on the new results, and the process repeats ad infinitum (perpetual beta). The more compute resources the user can allocate to this process, the better the results.

Where do we go from here?

This post is intended as an introduction to the conceptual methodology and these components by no means define the entirety of what we have designed into Quantum ML. In the following series of posts we will walk you through different aspects of our solution so as to better outline, define and deep dive into what’s behind our technology and solution. Our primary goal is to bring clarity and confidence to our existing (and potential) customers, so we encourage you to reach out to us with any questions, suggestions and even criticisms. Our philosophy with regards to constant adaptation and improvement isn’t just limited to our technology, but our business practices too.

---

About QDT

Quantum Data Technologies is a data science services and solutions company based out of Vancouver, BC, and New York. We specialize in Machine Learning and AI, and build robust, cross-industry data science solutions for major global enterprises. Our core product, Quantum ML, is a cloud-based data platform that uses Automated Machine Learning to help users predict markets, understand market drivers, and manage risk. With Quantum ML, users can easily deploy highly accurate ML models to predict any financial instrument – no data science or coding expertise required. We also offer consultancy and data services in addition to, or in augmentation of, our ML solutions, including data sales, data services, and custom software development.

To learn more about QDT visit our website or contact us at info@qdt.ai.
Website | LinkedIn | Twitter | Blog

QDT got its start building one Machine Learning model for Mars Wrigley to predict cocoa futures. To deploy the first model it took us 6 months of development at a cost of approximately $1million USD. Our technology has come a long way since then, and the results speak for themselves. The following section lays out our solution evolution in terms of:

1. Challenges 
2. Features/Benefits
3. Business outcomes
4. What’s next?

1. Challenges

A few of the major challenges we faced with our original methodology were:

• Data prep is slow and labour intensive
• Model updates are manual and therefore time-consuming
• There is so much data, and different data is predictive in different contexts
• An individual model is vulnerable to changes in market dynamics

To address these challenges and more, our team developed the beta version of Quantum ML, inspired by some open source code. To make the code effective for time-series predictions we did a significant amount of customization, and, amongst other things, implemented a more sophisticated machine learning methodology, as well as additional analytical tools and visualizations.

2. Features/Benefits

Some examples of developed features and their benefits are:

• Multi-model deployment – allows us to build and run multiple models at once.
• Auto-updates – enables daily updates of models without any human input.
• Group Dashboard (multi-model view) – provides a comprehensive view of multiple model predictions and performance metrics across multiple timeframes.
• Driver Analysis – Quantifies relative impact of specific data inputs across time.
• News Page – Daily news aggregation from 173 news sources with Theme and Keyword filter function to identify impactful news stories.
• Added Algorithms – Expanded library of machine learning algorithms.
• Algorithm improvements – Hyperparameter customization.
• Hyper-optimization protocol – Algorithm optimization protocol.
• Centralized Data Lake – for ease of data access
• Auto-series – Auto-model generation and input optimization.
• Expanded Use-Cases and Timelines

Building off of these new features and the success achieved with Mars Wrigley predicting Cocoa, we expanded the number of target variables to other agricultural commodities. 

• Sugar
• Wheat
• Rice
• Corn 
• Milk

In contrast to the original cocoa model deployment timeline, building better models for all of these targets took only three months. In addition to doing things faster, we also achieved better results.

3. Business Outcomes

As of today we have worked with the Mars’ Buying Team to build over 4000 models, and have more than tripled Mars’ original cost savings. Current yearly procurement cost reduction ranges from 2-8%, depending on the commodity.

4. What’s next?

In late 2020 we expanded the core function of Quantum ML to also provide Revenue and Sales forecasting. We call this Superforecasting. Simply put, what makes Superforecasting so effective is that it incorporates external (global) data inputs like regional GDP, Unemployment, and a long list of additional alternative data types, like Covid case and death rates, that improve accuracy and granularity of analysis.

As of June 2020 we are halfway through our third project with Mars: SuperROI. SuperROI uses the same underlying ML engine as Quantum ML and Superforecasting, but with a different use-case: Marketing Optimization. With SuperROI we have built off of Superforecasting to provide an additional layer of ML and analytics that tells Mars’ how effective their marketing spend is across advertising channels, and how to optimize that spend to maximize sales.