stratsphera is a portal dedicated to the development of investment algorithms in different markets, with a special focus on Brazil. On the portal, users will be able to create their own strategies using the various available databases, conducting researches and competing for prizes in competitions.

Our mission is to simplify access to all the steps of Algo Trading – from creating ideas to monetizing strategies.

General Overview

stratsphera is supported by two pillars: content and simulation. All content is accessed through the SPHERA menu, where users can navigate between the Blog, Tutorials, Articles and Forum – which consolidates topics and comments from the other content areas.

The entire simulation environment, accessed through the STRATEGIES menu, works directly in the browser. It is possible to write the algorithm in Python language through the use of the Strategy module and the list of allowed external libraries. After an execution, backtest results are available for analysis.

Important Concepts

Data Base

Some datasets will be available in the simulation environment. Data is presented in 1-minute bars, but can be grouped into larger bars as per the user's choice. The data providers are listed in the table below. For a detailed list of data, please go to the Data Library. Historical data will only be available in the simulation environment, with a 2-day delay period for updates, and its use outside stratsphera is prohibited - please read our Terms of Use.

The datasets disregard direct negotiations (combined between the parties), since they are not accessible in Algo Trading. It should also be noted that the platform does not have auction data (opening and closing) yet.

Each bar is a summary of the activities of the asset in a time interval and brings the following information: opening price, closing price, highest price, lowest price, volume weighted average price (VWAP price) and volume of assets traded. As seen in the image below, the first tick data in the bar refers to the opening price, while the last tick data refers to the closing price.The average price is calculated by the average price amount weighted by the traded volumes on each tick.

Execution Simulations (Backtests)

Any purchases and sales that occur during strategy simulations will be based on a price calculated according to the steps shown below. In a simplified way, an order sent near the beginning of the bar will have a price that is approximate to the opening price, while an order sent near the end of the bar will have a price that is approximate to the closing price. Likewise, an order that intends to execute a high percentage of the remaining volume of the bar will have a price that is approximate to VWAP price of the rest of the bar. The complete steps for simulating order executions are:

• 1) If an order is sent before the bar start, the base price will be the opening price. If an order is sent during the bar window, the average between the opening price and the closing price weighted by the percentage of bar time already used will be taken as the base price.

• 2) The VWAP price is then approached from the closing price in proportion to the percentage of bar time already used.

• 3) The base price (1) is then approached to the adjusted VWAP price (2) proportionally to the percentage of the bars available volume that will be executed by the order.

• 4) Finally, user-configurable adjustments are applied:

• Bar prices are adjusted with market impact simulation;
• Execution prices are adjusted with slippage simulation;

It is important to note that the volume available at a bar linearly decreases as the bar time is used.

It must be observed that only orders that are still active at the end of the bar will be executed. That is, if an order is sent and canceled between the start and end of a single bar (i.e., the order existed only inside the bars time window), it will not have it`s execution simulated. The only way to get around this is by increasing the strategy frequency.

The use of other datasets that are not available on the platform is currently not allowed.

For better strategy development, we recommend that an updated data window be saved for the out-of-sample test. More details on how to handle the backtest are available in the Tutorials.

Paper Trading

Paper Trading is the act of simulating a strategy using real-time data, so it is different from the simulation done in Backtest since there is no way to manipulate the data. Paper Trading is very important in testing an algorithm as it will eventually reveal strategy creation problems, such as overfitting historical data. Paper Trading is not yet available in stratsphera.

Live Trading

Live Trading is the act of executing the strategy live, using real time data and sending the strategy orders to the exchange. Therefore, it is the final stage of the strategy development. To clear a strategy for Live Trading, it is vital that the Backtest and Paper Trading phases have been appropriately executed. Live Trading is not yet available in stratsphera.