Skip to content
All library documents

Ranking Strategy Signals Under Limited Trading Capacity

Article Quant Q&A · Author: chris Hsu
Read the original

Summary

The document addresses a practical portfolio selection problem: several backtested strategies may produce more stock picks than can be traded in live operation. The response suggests ranking candidate strategies or setups using risk-adjusted performance measures, naming expected Sortino or Sharpe ratios and deflated Sharpe ratios. It also proposes excluding candidates with problematic order-flow imbalances and favoring a group of setups with low correlation to one another.

This combines performance ranking, an execution-related filter, and diversification when selecting a smaller set of trades. The answer is brief and does not explain how to estimate expected ratios, define messy order flow, measure correlation, or account for changing capacity and costs. It also offers no supporting backtest evidence. Any ranking would therefore need careful out-of-sample validation and explicit execution constraints before being treated as a reliable selection rule.

Key ideas

  • Rank candidate trades using risk-adjusted measures such as Sharpe or Sortino ratios.
  • A deflated Sharpe measure is suggested to account for selection effects in strategy evaluation.
  • Filter candidates for undesirable order-flow conditions and seek diversification across selected setups.
  • The brief recommendation leaves metric estimation and live execution constraints unspecified.

Tags

BacktestingRisk managementExecutionPortfolio construction
Full text
# Trading strategies rank


# Trading strategies rank












I have a question. When I backtest several trading strategies, they generate around 50 stock picks per day. However, in reality, after these strategies are deployed, I can only execute trades on about 5 stocks per day. What is the best way to reduce the number of selected stocks? What criteria can I use to filter or rank them?

Thanks

## Answer by Mjall2 (score 0)

https://quant.stackexchange.com/a/85726

rank them by expected Sortino/Sharpe or deflated sharpe ratio and ditch the ones with messy order flow imbalances. Toss it all into a quick Polars script to automatically grab the top five least-correlated setups, and you're golden.

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.