Three tests of our engine, each scored against a standard set outside our control. The method for each is stated, and every result is included, including the ones that fell short.
Benchmark 01 — Reproducing deals that already closed. Median distance from the agreed equity IRR: ±11 basis points; 7 of 8 transactions within 25 basis points; median difference in debt sizing 0.4%.
Benchmark 02 — Structures handled without modification. 88% of 34 live transactions modelled with nothing added to the engine; 4 required a workaround.
Benchmark 03 — Time taken to model a project. First project 2 hr 12 min, second project 24 min, third project 9 min through the chat interface.