Introduction

This page explains the philosophy behind the PathLit engine. The key notions here are unopiniated, transparent, explainable. If you are interested in the api aspect, please head to api - overview section of the documentation.

PathLit approach

We studied the wealth management value-chain and identified portfolio optimisation - asset allocation as a field with a lot of variables with a penchant for increasing computational complexity. Our hypothesis is that even if the wealth management industry is highly opinionated when it comes to “making-money”, there are steps for portfolio optimisation which are fairly standard and re-used, regardless of the chosen strategy.

We have identified that each standard step (for e.g. computing log-returns) could be a block which could work in orchestration with other blocks that are thematically organised into layers to create optimisation paths to cover a full range of strategies.

All the blocks in the current version are found in academic and practitioner papers and journals. They are transparent and provided for all to view and critique.

Every block and every path would be made accessible to the client for internal verification as well. Given the nature of the approach, the core of the product would not lie in the method of the optimisation or the paths that it takes, but in how:

1. Fast (7 layers with 7 blocks per layer generates 823543 paths for computation)
2. Reliable (Reproducible results, deterministic in nature, audit trails and log available)
3. Secure (Best practises in the industry for APIs)
4. Easily it can integrate with current systems (Clean API infrastructure and documentation)

The heavy-lifting is primarily mathematical and technical in nature.