<aside> ⚙ The trading bot that doesn’t trade live with an exchange and is made with the sole purpose of creating a dataset to train a ML model

</aside>

The basic flow of the trading bot should look something like this:

It needs to start up and scan the market for possible trade setups the trade indicator should be made with a logarithmic strategy once a trade is started it should loop until it is ready to sell and then log how the trade went and log the information needed to train the ML model to filter out bad trades and boost profitability.

The strategy

The most important part of the bot is it’s strategy. The bot will be written to trade BTC

— Pre trade

???

— Trade phase

A trade will be closed once a certain % loss or gain is incurred so no specific sell trigger is needed

The Flow Of The Bot

The bot will exist out of 3 phases

Phase 1 - Pre trade phase

The bot will scan the market using an API and request the needed data then compare against it’s strategy and enter a trade once the trade setup aligns with the market.

Phase 2 - Trade phase

The bot should save all the needed data in memory and check how the trade goes if it turns out and save the outcome of the trade.

Phase 3 - Reset phase

At the final phase the bot should log all the data of the trade in a file. and then reset again to phase 1.

Programming the bot

Flow function

bot() this will be the main function that packages the functionality into a working product

Create a loop that loops over trade_signal(<market_data>) until it returns true then start a trade and go into a new loop that checks on the trade check_trade(<coin_price>) once the trade is considered a loss or win it should log the trade log_trade(<trigger_trade_data>, <outcome>) with the market data that triggered the trade signal and the outcome of the trade. Then go back to the first loop and do it all over again.