Reproducible Risk Analysis: a P50 and P90 per facility, recorded so the run can be repeated.

For FID, funding, refinancing and recovery decisions, and for lenders’ technical advisers and insurers.

The range, its drivers, and how it was produced.

  • I run the risk workshop, set three-point ranges and run Monte Carlo in industry-standard software, giving P50 and P90 per key facility and the main drivers
  • Each facility’s driving path on the input file, calculated by my engine, which is the network the simulation varies
  • I record a reproducibility sheet: the seed (where the software allows one), iterations, software version, and the hashes of the sealed input file and risk register

What is recorded, so someone else can run it again.

On the sheet

Seed
The random seed, where the software allows a fixed one. Where it does not, the sheet says so
Iterations
The number of simulation runs
Software
The product and its version
Input file
SHA-256 of the sealed .xer the simulation was run on
Risk register
SHA-256 of the risk inputs: ranges, risks and their links to activities
Results
P50 and P90 per key facility, as reported by the run

What it does, and does not, show

  • The same software, version, seed, iterations and hashed inputs should give the same result. Where the software cannot fix a seed, a repeat run gives a close result, not the same one.
  • A P90 is a statement about the inputs, not a promise about the finish date.
  • A seal on the input file shows which file was used. What a seal proves, and what it does not

Turnaround and fee.

Turnaround
Agreed in writing at scoping
Fee
Quoted per engagement

Start from a reviewed baselineHow sealed files are verified

A range you can run again at the next gate.

Tell me the decision, the date and the schedule it rests on.

I reply within two working days. How I handle your files.