TSPS Trading Strategy Builder

This repo allows users to consume the historical data from TSPS and build trading strategies. This repository will provide a basis for you to be able to do analysis on the data, build a trading strategy, and backtest the strategy against historical data on all Streams inside of TSPS.

Current TSPS Data

Currently, the concentration will be on BTCUSD, but that will expand soon. Please read the instructions below to get started.

Requirements to Download Historical Data

1. Download Historical TSPS Data

You'll need the historical data in order to build a trading strategy.

To download historical performance of miners and their predictions, you first need to have Google's gcloud SDK installed.

• For Mac, you can use brew google-cloud-sdk.
• A more comprehensive documentation is also provided by Google: Install the gcloud CLI | Google Cloud CLI Documentation.

Clone repository

git clone https://github.com/taoshidev/tsps-trading-strategy-builder.git

Change directory

cd /tsps-trading-strategy-builder/historical_data

Download historical data

gsutil cp -r gs://tsps_btc .

This should automatically build a tsps_btc folder in this repository.

Expected output dir: historical_data/tsps_btc/

2. Download Historical BTC Data

Once you've downloaded historical TSPS data, you should download historical BTC data to speed up your processes.

You can do this by running the get_historical_data.py script.

python historical_data/get_historical_data.py

This will generate a data file from binance called historical_binance_data.csv which will provide the actual closes that occurred that can be paired with the predictions to analyze results.

Analyzing Data

The main logic outlined currently lives in the run.py script. The current setup takes the ongoing historical data dump from gcloud, along with the generated historical btc data, and transforms it to allow you to analyze the results by top miners on the network.

Things to consider

• Adding smoothing averages across top performing miners using weighted averages
• Concentrating trading strategies on larger moves, and consensus on larger moves
• Adding additional market indicators to add more signal
  • horizontal / diagonal resistances
  • traditional indicators such as MACD / RSI
  • futures data (such as liquidiation events)
• Using the above in a consensus model (LSTM)

Next Feature

• Build additional signals
• Create the first open source consensus model (LSTM)
• Build a backtesting engine

License

Copyright © 2024 Taoshi Inc

Source code produced by Taoshi Inc may not be reproduced, modified, or distributed
without the express permission of Taoshi Inc.