Liquidity Control for CEXes

Objective

Determine whether an exchange has sufficient liquidity for a specific asset.

Input Data

• Exchange (e.g., Binance, Coinbase, Kraken, etc.)
• Asset for which liquidity is being assessed (e.g., BTC/USD, ETH/USDT, etc.)
• Thresholds for each liquidity metric, which can be set individually for each feed and also have default values.

General Algorithm

1. Obtain liquidity data for each metric for the specified asset and exchange using the API of CEXClient.
2. Check whether the liquidity data for the given asset on the exchange meets the set thresholds for each metric.
3. If the exchange meets all thresholds for all metrics, consider it liquid for the given asset; otherwise, exclude this exchange from the price calculation for the given asset.

This logic will be implemented in CEXLiquidityCheckMixin and added to CEX Feeds.

Liquidity Metrics

Using multiple liquidity control metrics allows for a more comprehensive and accurate assessment of an exchange's liquidity for a specific asset, as each metric accounts for different aspects of trading.

Thresholds can be set individually for each Feed, but also have default values (which may be revised during implementation and operation).

Trading Volume

This is the amount of the asset that has been bought and sold on the exchange over a specific period. The higher the trading volume is, the greater the liquidity of the exchange.

Algorithm

For simplicity in the initial implementation, the trading volume is calculated over a period corresponding to the price weighting window, using trade data that is also required for price calculation, specifically CEXClient.get_trades().

Threshold

min_trade_volume (default: US $1,000)

Order Book Depth

This metric represents the total amount of the asset available for sale or purchase at specified prices on the exchange. The greater the order book depth is, the higher the liquidity of the exchange is considered.

order_book_depth_percentage must be defined. It is a percentage of the current market price of the asset that determines the price range for analyzing the order book depth. For example, a 5% value indicates that orders with prices within the range of 95% to 105% of the current market price of the asset are analyzed. This parameter is set globally for all exchanges (default: 5%).

Algorithm

1. Obtain the open order data using the exchange's API, specifically CEXClient.get_order_book().
2. Calculate the total amount of the asset in the range.

Threshold

min_order_book_depth (default: US $10,000)

Spread

Calculate the spread as a percentage, which represents the relative difference between the best ask (sell) price and the best bid (buy) price on the exchange, expressed as a percentage of the average price between ask and bid. The smaller the spread is, the greater the liquidity of the exchange is considered.

Algorithm

1. Take the open order data, already used for calculating the order book depth, specifically CEXClient.get_order_book().
2. Determine the best ask price and the best bid price for the asset.
3. Calculate the average value between ask and bid: average_price = (ask + bid) / 2.
4. Calculate the spread as the difference between the best ask price and the best bid price: spread = ask - bid.
5. Calculate the spread as a percentage: (spread / average_price) * 100.

Threshold

max_spread_percentage (default: 1%)

Potential Improvements

• It may be necessary to use retries when checking the spread or the order book depth, as their values may temporarily change due to large market transactions.
• Accumulating trade data over time, rather than solely at the moment of price calculation, will allow for evaluating trade volume over a desired time period.
• Relative metrics, such as "market share" could be added, where we calculate trade volume in proportion to all available exchanges. This could be done in a derived Feed instead of the MedianFeed.
• Default threshold values and parameters might need adjustments based on real-world usage data.