AOCryptobot

Current development

Signal trading is here!! You can now develop your own strategies with some well known market analysis most famous indicators.

Changelog

[07 NOV 2021]

• Add basic database implementation for Positions
• Activate basic database implementation for Positions
• Add snippet to fulfill database from logs
• Add stop-loss feature
• Fix paper service not filling create and update times
• Delegate config parameters into config file
• Add support for whitelisted and blacklisted markets
• Add CLI implementation to send strategies via cmd args
• Fix monitor locks
• Add lock check to trading enter conditions
• Add candle limit to avoid memory overloads
• Improve database support to allow database candle inserts
• Fix TradeSignal false left open order in limbo
• Fix interval dissociation between strategies and monitor
• Fix database concurrency problems
• Add missing interval conf value
• Fix breaking error due using an unready analysis
• Add telegram output for trade logging

[07 APR 2020]

• Add signal-trading system with strategy analysis algorithms
• Add exit rules to oscillator system
• Add well-known market analysis algorithms as strategy pre and post conditions
• Refactor folders to get prepare for new trading systems
• Add fallback system to get back the unexpected residues from sell side
• Remove MarketSnapshotRecords list as an argument on MarketService functions
• Log totals on wallet update
• Add Ctrl+C detection to exit bot
• Refactor common package into 'service' root folder
• Refactor Go config to Module usage
• Fix e5 error 'The relationship of the prices for the orders its not correct'
• Log to file and simplify UI list appending
• Add .gitignore
• Add multi-threading support
• Prevent setting Market Status on empty ws response
• Prevent nil pointer reference errors on unsuccesful OrderCancel/GetOrder request
• Ensure any bid an ask is always present when market_snapshot is recorded
• Remove hardcoded pair occurrences

[09 FEB 2020]

• Add user interface: A graphic interface has been integrated. Still some bugs to make it stable, but milestone is not so far away.
• Fix order cancel and get sellingTimeout correctly working
• Resolve "Refactor binance monitor replacing REST with websocket calls"
• Update conf.env
• Add CODEOWNERS file

Signal Trader Mode

From 07 APR 2020, Market Maker mode are now combined with a Signal Trader mode.

Technical indicators

With the help of sdcoffey/techan, we can use different technical indicators such as RSI, MACD, BoilerBands, StochasticRSI, EMA...

Strategies

Tech indicators can be combined now to create strategies, with enter and exit rules over candle samples. In the own strategy logic, we can set an analysis algorithm to calculate backtest results from past samples, and a set of rules to declare it tradeable or not, depending on the analysis results. This means you can create an aggresive strategy and set the activation rules on for example "strategy had at least a 5% of profit on the backtest analysis, a custom ration between average profit per operation and std deviation", etcetc

The Signal Trader supports multiple streategy combinations, and provides another algorithm that is always running in background, analysing each strategy in a backtest, and defining if it is tradeable or not. In the same way, exists a configurable function where to develop strategy choosing.

For example, after analyse ETHEUR:

Strategy 1 (lets say RSI based):

• Backtest profit in last 5 days 2,31%, std dev 4,54%. Strategy says is tradeable

Strategy 2 (lets say EMA based):

• Backtest profit in last 5 days -1,31%, std dev 12,54%. Strategy says is not tradeable

Strategy 3 (lets say MACD based):

• Backtest profit in last 5 days 4,31%, std dev 1,54%. Strategy says is tradeable

Choice rules are taking the best profit / std dev ratio, so MACD is selected. Start trading.

Parametrizing strategies

Alanlyzer algorithms can, in addition to calculate profits with backtest, try the backtests with different parameters. Lets say you develop an strategy with the enter rule "enter if EMA increases in X (being X a parameter)" and the exit rule the same, but inverse. The backtesting algorithm can calculate the best parameters combination in terms of profit. So the bot calculates not only your past profit but the best combination of parameters to get the maximum profit. This is because if the market is, for example, more stable, the slope necessary to get the bigger profit could be different than in other situations, so the slope (or what you parametrize) changes in different market situations, the bot will calculate it and proceed to trade always with the best combination. Pretty amazing :)

Usage

1. Install dependencies with go mod download

2. Test with go test ./..

3. Run with go run main.go

Open Source

The development is open to everyone and everything, to new market strategies such as arbitrage, perpetual market making... Just use as you need.