freqtrade_origin/docs/advanced-orderflow.md

Advanced Orderflow

This page explains some advanced tasks and configuration options that can be performed to use orderflow data by downloading public trade data.

Quickstart

enable using public trades in config.json

"exchange": {
   ...
   "use_public_trades": true,
}

set orderflow processing configuration in config.json:

"orderflow": {
    "scale": 0.5, 
    "stacked_imbalance_range": 3, # needs at least this amount of imblance next to each other
    "imbalance_volume": 1, # filters out below
    "imbalance_ratio": 300 # filters out ratio lower than
  },

Downloading data for backtesting

• use --dl-trades to fetch trades for timerange

For example

freqtrade download-data -p BTC/USDT:USDT --timerange 20230101-  --trading-mode futures  --timeframes 5m --dl-trades

Accessing orderflow data

Several new columns are available when activated.

    dataframe['trades'] # every single trade
    dataframe['orderflow'] # footprint chart: see below
    dataframe['bid'] # bid sum
    dataframe['ask'] # ask sum
    dataframe['delta'] # ask - bid
    dataframe['min_delta'] # minimum delta reached within candle
    dataframe['max_delta'] # maximum delta reached within candle
    dataframe['total_trades'] # amount of trades
    dataframe['stacked_imbalances_bid'] # price level stacked imbalance bid occurred
    dataframe['stacked_imbalances_ask'] # price level stacked imbalance ask occurred

These can be accessed like this:

def populate_indicators(self, dataframe: DataFrame, metadata: dict) -> DataFrame:
    # calculating cumulative delta
    dataframe['cum_delta'] = cumulative_delta(dataframe['delta'])

def cumulative_delta(delta: Series):
    cumdelta = delta.cumsum()
    return cumdelta

dataframe['orderflow']

This includes a dataframe that represents a Footprint chart of the Bid vs Ask type. Footprint charts are a type of candlestick chart that provides additional information, such as trade volume and order flow, in addition to price. The scale of the price is set by orderflow.scale (see above) and thus binned per price level.

Following columns are available:

orderflow_df['bid_amount'] # how much bids were traded
orderflow_df['ask_amount'] # how much asks were traded
orderflow_df['bid'] # how many bids trades
orderflow_df['ask'] # how many asks trades
orderflow_df['delta'] # ask amount - bid amount
orderflow_df['total_volume'] # ask amount + bid amount
orderflow_df['total_trades'] # ask + bid trades