Historical simulation with point-in-time data, transaction-cost assumptions, and look-ahead safeguards.
Back Testing is the historical simulation workbench. Every bar, every trade, every position snapshot is clickable to the data as it would have looked on that day — no look-ahead, no survivorship bias. Transaction-cost and slippage assumptions are configurable per run, and results route into the same Performance, Risk, and Attribution views you use for live portfolios.
4 reports in this category. Every cell is clickable to the source row.
Can I prove a strategy on real historical data with no look-ahead?
Event-driven simulator that replays prices, corporate actions, and portfolio state as they existed on each historical day. Supports walk-forward and rolling-window runs; strategy logic is authored in plain Python against a point-in-time API.
What would this strategy have actually cost to run?
Configurable per-security cost model — commission, spread, market impact — applied to every simulated fill. Supports linear, square-root, and Almgren-Chriss impact models for sizing-sensitive strategies.
Did the strategy deliver after costs?
Reuses the live Performance reports (TWR, MWR, Brinson attribution, GIPS-style statistics) against the simulated NAV series. Directly comparable to live production results, so a paper strategy's numbers mean the same thing as the live book's.
Where did this strategy take risk, and what was the worst-case exposure?
Covariance, VaR, stress tests, and exposure concentration computed over the simulation window. Identical views to the live Risk category so you can compare simulated and realized risk profiles side-by-side.