Backtesting Engine Build

A written spec, a clear architecture choice, a tested core for orders, fills and accounting, corporate actions and cost models, and golden tests that prove the engine reproduces known results.

Free · Mac (Apple silicon) and Windows · v1.0.5 · Other platforms

Then open the Marketplace in the app, search for Backtesting Engine Build, and click Install. It goes into your own files, and works offline from then on.

Before you start: this plan is general information to help you organise decisions about money. It is not financial, investment, tax or legal advice, and it does not recommend any product, investment or deal. Rules differ between countries, so check decisions that matter with a qualified, licensed adviser in yours. Investing and trading can lose money, and raising or managing other people's money is regulated in most countries.

What it installs.

1

Pillar

1

Area

50

Projects

200

Starter tasks

🧮 Quantitative Finance & Trading

Backtesting Engine Build

Building a backtesting engine: event-driven or vectorised design, corporate actions, cost and fill models, reproducible runs and tests against known results.

  1. Requirements spec for the backtesting engine
  2. Build, buy or open-source engine decision
  3. Event-driven or vectorised architecture choice
  4. Core domain model for orders, fills and positions
  5. Simulation clock and event queue design
  6. Point-in-time data handler interface
  7. Pinned environment and continuous integration for the engine
  8. Hand-computed golden test cases
  9. Portfolio accounting ledger with daily reconciliation
  10. Run configuration files and run identifiers
  11. Golden and regression test report routine
  12. Engine release and versioning routine
  13. Weekly engine issue triage
  14. Monthly engine performance benchmark
  15. Shadow replay of live fills through the engine
  16. Monthly dependency and numerical library update
  17. Known limitations register for the engine
  18. Data contract tests on each new data delivery
  19. Feedback loop with researchers using the engine
  20. Studying open-source backtester designs
  21. Prices, ticks and floating point in simulation
  22. Vectorised backtesting without look-ahead
  23. Profiling and speeding up the engine hot loop
  24. Property-based tests for accounting invariants
  25. Trading calendars, sessions and time zones
  26. Deterministic runs with seeded randomness
  27. Corporate actions engine for splits, dividends and symbol changes
  28. Pluggable fill model with partial fills and volume caps
  29. Pluggable transaction cost model module
  30. Order types for limit, stop and auction orders
  31. Futures roll and continuous contract handling
  32. Multi-currency accounting and FX conversion
  33. Short borrow, financing and margin accounting
  34. Risk limits, position sizing and stop rule module
  35. Parallel parameter sweeps on a job runner
  36. Results store and standard tear sheet output
  37. Moving a notebook prototype into the engine
  38. First end-to-end run on a reference strategy
  39. Cross-check against a second backtesting tool
  40. Version 1.0 engine sign-off review
  41. Engine demo and handover session for researchers
  42. Post-incident review of a backtest bug
  43. Minimal engine for a solo personal researcher
  44. Researcher guide to extending the engine safely
  45. Inheriting a legacy backtester on a new team
  46. Student backtesting engine as a portfolio project
  47. Order book replay simulator for intraday strategies
  48. Shared strategy code path for backtest and live
  49. Options and multi-leg instrument support
  50. Latency and data delay simulation

Where it starts.

Each project arrives with its purpose, milestones and first tasks written out. Here is how the first three read.

Requirements spec for the backtesting engine

Engines built without a written scope tend to grow every feature at once and finish none, then quietly disagree with each other in edge cases. One spec naming the asset classes, bar or tick resolution, order types, cost models, run time budget and who will use the engine settles most later arguments before the first line of code.

First tasks

  • List the asset classes and data resolutions the first version must handle
  • Ask two intended users which questions the engine must answer for them
  • Write numeric run time and memory targets for a typical study
  • Get the spec reviewed by a researcher and an engineer before coding

Build, buy or open-source engine decision

Writing a backtester from scratch is a well-known way to spend a year rebuilding what an open-source library or a vendor platform already does, yet existing tools can hide fill and accounting assumptions you cannot change. Scoring the realistic options against the spec makes the choice deliberate, and the reasons are on record when someone asks later.

First tasks

  • Shortlist three to five engines or frameworks that could meet the spec
  • Run the same simple reference strategy on the top two candidates
  • Check each licence and how recently each project was maintained
  • Write the decision with the conditions that would make you revisit it

Event-driven or vectorised architecture choice

Vectorised engines run a daily cross-sectional study in seconds but struggle with path-dependent logic such as stops, partial fills and cash constraints, while event-driven engines model those faithfully at a large cost in speed. Picking the core design, or a deliberate split between a fast research mode and a detailed simulation mode, shapes every module that follows.

First tasks

  • Sort the in-scope strategies into path-dependent and path-independent
  • Prototype one moving average strategy in vectorised and event-driven style
  • Time both prototypes and confirm their trade lists match
  • Write the architecture decision record with the trade-offs accepted

Make it yours.

Install Backtesting Engine Build, then tell Aurum the date you're aiming for. It checks whether the plan fits the weeks you actually have.

Free · Mac (Apple silicon) and Windows · v1.0.5 · Other platforms

Blueprints are general information to help you organise your own work. They are not medical, financial, legal or other professional advice. See the Terms.