BitSmart will be an automated system used for trading Bitcoin on Kraken platform.
It’s something I couldn’t have created without the precious help from my friend Max.
The initial idea (2018) is:
BitSmart gathers data using RPA technology and third party tools for market and social media sentiment. It further uses neural networks to process data and draw initial conclusions. Those conclusions are then refined using fuzzy logic and locally sensitive states. The whole cycle is repeated every two minutes so BitSmart can generate profit even if winning margin is low, but worth the risk.
Project is far from completion, but I’m pretty confident it will see the light of day.
UPDATE:
A few years later, after some roadblocks (mainly not being able to find any meaningful insight from data, not being able to develop LSTM that resulted in tangible profit), I’ve decided to overhaul the project.
New BitSmart is an early-alarm system that recognizes quick spikes in price for BTC/USD and sends notifications to subscribed users. It runs in the background and is subscribed to latest price using Web Sockets. It can capture quick gradient change in 30 seconds timeframe. This system I’ve built to anticipate price spike that I expect when BTC ETF are recognized by SEC. Idea is to enter leveraged (5x) trade at the beginning of a spike and “ride the wave”. BitSmart possesses the capability to execute buy/sell orders automatically, with or without leverage on predefined % of portfolio.
UPDATE 2:
After SEC ETF approval
The initial idea (2018) is:
BitSmart gathers data using RPA technology and third party tools for market and social media sentiment. It further uses neural networks to process data and draw initial conclusions. Those conclusions are then refined using fuzzy logic and locally sensitive states. The whole cycle is repeated every two minutes so BitSmart can generate profit even if winning margin is low, but worth the risk.
Project is far from completion, but I’m pretty confident it will see the light of day.
UPDATE 3:
A few years later, after some roadblocks (mainly not being able to find any meaningful insight from data, not being able to develop LSTM that resulted in tangible profit), I’ve decided to overhaul the project.
New BitSmart is an early-alarm system that recognizes quick spikes in price for BTC/USD and sends notifications to subscribed users. It runs in the background and is subscribed to latest price using Web Sockets. It can capture quick gradient change in 30 seconds timeframe. This system I’ve built to anticipate price spike that I expect when BTC ETF are recognized by SEC. Idea is to enter leveraged (5x) trade at the beginning of a spike and “ride the wave”. BitSmart possesses the capability to execute buy/sell orders automatically, with or without leverage on predefined % of portfolio.
UPDATE 4:
After SEC ETF approval developed program managed to catch spike in Bitcoin price and make some nice returns. This project was a successful one, even after more than five years of trying.
UPDATE 5, 21 OCT 2024:
I’m in process of adding almost real-time monitoring of a few selected stocks and some of their option data. Idea is to capture data for further analysis and to have early signals by calculating rolling gradient.