SizeRiskModel#

API documentation for tradeexecutor.strategy.size_risk_model.SizeRiskModel Python class in Trading Strategy framework.

class SizeRiskModel[source]#

Bases: ABC

Estimate an impact of a single trade.

  • We are going to take a price impact hit when taking liquidity out of the market, and sometimes this hit is too large, so we cannot trade an asset with the amount of capital we have - this is call “capacity limited” in trading

  • Handle max sizes for individual trades and positions, so that we do not create trades or positions that are too big compared to the available lit liquidity

This is an abstract base class. See implementations for details.

  • capped fixed amount (no data needed)

  • historical real data (EVM archive node),

  • historical estimation (based on TVL)

  • live real data (EVM node)

Example:

from tradeexecutor.strategy.tvl_size_risk import HistoricalUSDTVLSizeRiskModel

def decide_trades(input: StrategyInput) -> list[TradeExecution]:
    position_manager = input.get_position_manager()
    pair = input.get_default_pair()
    cash = input.state.portfolio.get_cash()
    timestamp = input.timestamp

    size_risker = HistoricalUSDTVLSizeRiskModel(
        input.pricing_model,
        per_trade_cap=0.02,  # Cap trade to 2% of pool TVL
    )

    trades = []
    if not position_manager.is_any_open():
        # Ask the size risk model what kind of estimation they give for this pair
        # and then cap the trade size based on this
        size_risk = size_risker.get_acceptable_size_for_buy(timestamp, pair, cash)
        # We never enter 100% position with our cash,
        # as floating points do not work well with ERC-20 18 decimal accuracy
        # and all kind of problematic rounding errors would happen.
        position_size = min(cash * 0.99, size_risk.accepted_size)
        trades += position_manager.open_spot(
            pair,
            position_size
        )
    else:
        trades += position_manager.close_all()
    return trades

Example using AlphaModel in decide_trades():

# Calculate how much dollar value we want each individual position to be on this strategy cycle,
# based on our total available equity
portfolio = position_manager.get_current_portfolio()
portfolio_target_value = portfolio.get_total_equity() * parameters.allocation

#
# Do 1/N weighting
#
# Select max_assets_in_portfolio assets in which we are going to invest
# Calculate a weight for ecah asset in the portfolio using 1/N method based on the raw signal
alpha_model.select_top_signals(parameters.max_assets_in_portfolio)
alpha_model.assign_weights(method=weight_by_1_slash_n)

#
# Normalise weights and cap the positions
#

size_risk_model = USDTVLSizeRiskModel(
    pricing_model=input.pricing_model,
    per_position_cap=parameters.per_position_cap_of_pool,  # This is how much % by all pool TVL we can allocate for a position
)

alpha_model.normalise_weights(
    investable_equity=portfolio_target_value,
    size_risk_model=size_risk_model,
)

# Load in old weight for each trading pair signal,
# so we can calculate the adjustment trade size
alpha_model.update_old_weights(
    portfolio,
    ignore_credit=True,
)
alpha_model.calculate_target_positions(position_manager)
__init__()#

Methods

__init__()

get_acceptable_size_for_buy(timestamp, pair, ...)

get_acceptable_size_for_position(timestamp, ...)

What this the maximum position amount.

get_acceptable_size_for_sell(timestamp, ...)

get_acceptable_size_for_position(timestamp, pair, asked_value)[source]#

What this the maximum position amount.

  • Avoid exceed maximum position size

  • If the position size is exceeded start to reduce the position

Parameters:

  • timestamp (Optional[Union[datetime, Timestamp]]) –

    Historical timestamp.

    Can be set to None for onchain reading backends, and they will use the latest block number.

  • asked_value (float) – How large position we’d like to have in the US Dollar terms.

  • pair (TradingPairIdentifier) –

Returns:

Size-risk adjusted estimation how large the position could be.

Return type:

SizeRisk