Conclusion

 

• Performance Wheel compares and provides credible inputs to buy-side trading desks to create broker-algorithm wheels based on clean and validated broker-algorithm data.

• The machine-learning-based model builds on a large broker routes dataset and captures essential characteristics of the equity electronic algorithms to accurately predict broker-algorithm performance for various slices of order flow.

• Comparing client’s own routes data with the wider Bloomberg dataset provides an effective decision-making framework to create and manage algorithm wheels.