A Superior Trading Intelligence Feeding from the Exploration of the Trading Space
A superalgo is an artificial intelligence that dynamically combines any number of AlgorithmsRefers to trading algorithms; pieces of software automating trading Strategies. within a pool, using a vast bank of trading experiences gathered from hundreds of thousands of live-trading runs and CompetitionRefers to trading competitions; events in which Algorithms and human traders compete to determine the top performers under a given set of rules.Learn more about Competitions... matches carefully recorded and catalogued within the supermind's collective memory as the food for its learning processes.
The production of Algorithms is driven by an open worldwide community of developers, traders and analysts collaborating by engaging in iterative competition cycles. Competitions measure the performance of individual strategies and different forms of Algorithms, identifying and selecting promising traits for Algo-MakersA subset of the Community counting professional and amateur traders, developers and analysts, organized in Teams, commited to create, develop, maintain and promote the evolution of Algorithms. to breed and help evolve.
The Competition system is designed to explore the widest possible trading universe in order to produce a comprehensive bank of memories covering as many experiences as possible through an array of variables such as markets, market conditions, strategies, or technical implementations.
Competition dynamics drive the evolution of different kinds of Algorithms, all of them existing as autonomous agents with a Financial LifeA type of artificial life in which agents are alive as long as they have enough money to pay for their expenses. of their own. These Financial BeingsA perspective on Algorithms and other sort of bots (i.e. Indicators, Sensors, etc.) as autonomous agents with a financial life of their own, meaning that they live or die depending on their ability to produce value in the form of products (data) and services (trading), which they sell to obtain financial resources and guarantee their subsistence.Learn more about Financial Beings... feature varying degrees of intricacy, resulting from the combination of simple Algorithms into complex organisms:
An algobot is the basic structural and functional unit of Algorithms running on the platform. They are coded to implement a single trading strategy.
They consume data from Sensors and Indicators and use their built-in logic to make trading decisions such as opening, moving or closing positions.
If an Algobot is not successful in its trading job and its performance drops below a pre-defined threshold, it dies.
Algonets are organisms which dynamically deploy swarms of algobot clones, assigning each one a unique set of parameters within predefined ranges, allocating a percentage of working capital to each clone. They monitor algobots performance over time and decide which algobots to keep alive and which to let die depending on their performance.
Each algonet is made out of algobot clones implementing the same trading strategy. The job of an algonet is to find the best possible structure of algobots to do the trading.
Advanced algos function as higher order trading organisms, as their main activity is deploying algonet clones appropriate for the job, taking into account to the strategies each algonet implements.
They monitor the performance of algonets and eventually decide which ones to keep and which ones to let die.
Thus, advanced algos use multiple strategies in the quest to adapt to changing market conditions and other variables.
The nesting of simple Algorithms into more complex organisms is –in essence—an approach to learning algorithms which use live experiments to explore ranges of possible behaviors and dynamically select the most efficient ones while discarding the less desirable. These learning mechanisms not only supply a live feedback loop to the organism doing the actual trading, but also stores valuable information to be recalled by the system at any point in time.
It is this wealth of memories and real-life trading experiences that superalgos will harness. How exactly they are going to do that will be the outcome of the relentless evolution of the ecosystem and the work of the specialized intelligence segment of the supermind.
In that light, we invite researchers and computer scientists within the field of artificial intelligence to join the team and help design strategies to make the most out of the base layer of Algorithms and trading experiences the CommunityThe collaboration of an open and heterogeneous group participating in the Superalgos Project. A subset of the Community counts professional and amateur traders, developers and analysts, organized in Teams, referred to as "Algo-Makers". is developing.
We also invite experts with a deep understanding of financial markets, the capital investments space within the financial industry and algorithmic trading in general to join too, so as to help guide the IT efforts from the user and business case perspectives.
We believe your input can be of value in our quest to reorganizing the existing intelligence in the trading space so that more people have more opportunities to participate.