The blade quantitative system provides comprehensive and complete market data all over the world. Data format including (but not limited to) ticks, minute, daily and other periodic data of shares ,futures and cryptocurrency.
The blade quantification system provides data processing functions, including data cleaning, data aggregation, data feature calculation and data splicing. Market data of various granularity can be directly processed into training and test data sets by using the data processing function of blade, without additional programming by users.
Blade may be the most advanced automatic strategy seeking software on the market:
1. The kernal of blade strategy production uses the machine learning algorithm accelerated and optimized based on GPU graphics card, which has the fastest speed in the industry. It can try thousands of strategies according to the set strategy type with high efficiency and find profitable strategies. For example, on a Geforce 1080ti GPU graphics card of NVIDIA, it only takes 3 milliseconds (0.003 seconds) to complete the back test of one strategy and 14 million strategy in 12 hours for the futures data with 50k rows and 1000 columns.
2. The process of making blade strategy only needs to set parameters without programming. The process of making strategy is completely left to blade. It has built-in 58 types of strategies, including several categories: shock, trend, grid, mixing and combination. It supports both timeframe bar or non timeframe bar, meeting the needs of quantitative trading strategies for futures, stocks, precious metal spot, cryptocurrency and other varieties.
3. The blade has 28 kinds of performance objectives, including net profit, sharp ratio, profit factor, winning rate, etc., to meet the different performance requirements of different customers.
4. Blade has strong expansion ability. It naturally supports cluster deployment based on distributed architecture and can run in large GPU clusters and call hundreds of GPUs at the same time. It does not limit the amount of information of the input algorithm (unlimited number of rows and columns), which greatly increases the probability of finding a good strategy.
The strategy evaluation function can continuously evaluate the strategy, test whether the strategy is still effective, and obtain important information such as transaction signal and position weight, which can be used for manual position adjustment.
Blade supports the translation of generated strategies into Powerlanguage / Easylanguage language programs, which can be directly copied to Multicharts(www.multicharts.com) and TradeStation(www.tradestation.com) platforms for online operation to generate trade signal.From version 2.6 blade support translate strategy to MT4/MT5 code. It also supports C++,Python,.Net interface. We also provide examples to help users translate into China CTP interface based on futures and stocks according to customized development rules. The convenient translation function comes from the blade's strategy, which is a regular body plus calculation formula rather than a black box, so it can be translated and run on any online real offer platform.
The translated strategy code is directly copied and pasted into MultiCharts/TradeStation platform, compiled and run without any coding !
It also fully supports the MT4 or MT5 platform. It can be directly translated into their programs and compiled and run without writing any code!
General investors evolve into quantitative investors through the blade.
Professional quantitative investors get a powerful tool to quickly experiment with a variety of new ideas.
Real Future 1
Real Future 2
Stocks Portfolio Simulation 1
Stocks Portfolio Simulation2
Simulation Online Trace
Real future account profit performance chart trade with shfe.rb
Real Future Account trade with dce.m