catalyst/docs/source/live-trading-blueprint.md

Live Trading Blueprint

The purpose of this document is to allow project contributors navigate ongoing live trading implementation.

Components

At a high level the following components have been modified to coerce zipline into live trading.

Exchange

Exchange is a new package which introduces the concept of cryptocurrency exchanges to zipline. The package contains all new component implementations adapted to charasteristics of exchanges.

catalyst/exchange

Here are some key characteristics which makes exchanges different from equity and futures currently implemented in zipline.

• They trade around the clock.
• Currency symbols are inconsistent across exchanges.
• They trade currency pairs, i.e. the base currency is not always be USD. This is a significant departure from the equity market. Additional business logic will be required both to assess performance and manage trades.
• The cryptocurrency market being relatively immature, there are still significant price arbitrage opportunities between exchanges. In contrast with the equity markets, trader usually trade directly against an exchange (as oppose to using a broker). Consequently, to extract maximum alpha, the platform should not only support multiple exchanges, but also multiple exchanges per algorithm.
• The fee model is usually more complex than that of an equity broker. It can vary drastically between exchanges.
• There are no splits, mergers, etc to worry about.
• Their order book is publicly available, the platform should access to it as it can be used to drastically reduce slippage.

New Components

These components of the exchange package were added to the zipline sources.

Exchange

catalyst/exchange/exchange.py

Abstract class which acts an interface for the implementation of various exchanges. It also contains logic common to all exchanges.

Bitfinex

catalyst/exchange/bitfinex.py

The Bitfinex exchange implementation. It extends the Exchange class.

DataPortalExchange

catalyst/exchange/data_portal_exchange.py

Extends the zipline DataPortal to route spot data to the exchange. This is critical because it allows the algoritm to request data in real-time.

For example,

    data.current(asset, 'price')

retrieves the current price of the asset, not the price at the time of yielding the bar this is critical to minimize slippage.

ExchangeClock

catalyst/exchange/exchange_clock.py

An implementation to the zipline Clock which runs 24/7. It yeilds a bar every minute.

AssetFinderExchange

catalyst/exchange/asset_finder_exchange.py

An alternate implementation of AssetFinder which locates each asset against the exchanges instead of bundle databases.

For example,

symbol('eth_usd')

retrieves an Asset object against the exchange as opposed to querying a database of equities.

I have created a dictionary of currencies for the Bitfinex exchange. The primary goal is to standardize the symbol notation across exchanges. Here is a snippet of the file.

• Each key represents the exchange specific symbol.
• The symbol attribute represents the standard symbol which should be common across exchanges for the given currency.
• The start_date attribute should correspond to its first trading day on the exchange.

{
  "btcusd": {
    "symbol": "btc_usd",
    "start_date": "2010-01-01"
  },
  "ltcusd": {
    "symbol": "ltc_usd",
    "start_date": "2010-01-01"
  },
  "ltcbtc": {
    "symbol": "ltc_btc",
    "start_date": "2010-01-01"
  },
  "ethusd": {
    "symbol": "eth_usd",
    "start_date": "2010-01-01"
  },
  "ethbtc": {
    "symbol": "eth_btc",
    "start_date": "2010-01-01"
  }
}

ExchangeTradingAlgorithm

catalyst/exchange/algorithm_exchange.py

Extends the TradingAlgorithm class which orchestrates the api operations. This class brings together most of the components described above.

Modified Components

The following components have been modified to include conditional business logic to enable live trading.

run_algorithm

catalyst/utils/run_algo.py

The run_algorithm interface is an entry point to execute an algorithm in zipline. This component was already modified for the catalyst concurrency bundles. I added conditional logic to which should not break any of the existing backtesting implementations.

At a high-level the run_algorithm method now contains two additional parameters:

• live: If True, zipline will attempt to trade live. If False or not specified, it will run a backtest as normal.
• exchange_conn: A dictionary containing the attributes required to instantiate an exchange. Here is an example for Bitfinex:

exchange_conn = dict(
    name='bitfinex',
    key='',
    secret=b'',
    base_currency='usd'
)

The following sample algorithm uses the run_algorithm interface:

catalyst/examples/buy_and_hold_live.py