Run properly, simulation answers the one question deterministic tables cannot: what is the probability coverage holds?
Deterministic sensitivity asks what happens at chosen shock levels; Monte Carlo asks how likely the bad levels are. The method draws thousands of scenarios from probability distributions assigned to the key inputs — revenue growth, occupancy or capture, operating margins, rates — and reports the resulting distribution of outcomes: the probability that year-one coverage exceeds 1.15x, the fifth-percentile DSCR, the likelihood the project ever breaches 1.00x across the horizon.
The honesty of the exercise lives entirely in the inputs. Distributions must come from evidence — the dispersion observed in comparable operations, published industry ranges, the project’s own contract structure — and correlated inputs must be correlated in the model: occupancy and rate move together in hospitality; volume and margin fight each other in fuel retail. An uncorrelated simulation over invented distributions is theater with error bars.
The committee-relevant outputs are few: the probability of clearing the program minimum in the binding year, the downside percentile coverage, and the input the outcome is most sensitive to — the tornado’s top bar — because that is where conditions and covenants should attach. A simulation showing a 92 percent probability of clearing 1.15x with the residual risk concentrated in ramp speed tells the lender exactly what reserve to require.
Wert-Berater includes Monte Carlo analysis in the standard study package alongside — never instead of — the deterministic tables, because the two answer different questions and committees deserve both.
Independent feasibility studies since 1998 — 4,000+ engagements, $40.2 billion in evaluated project value. Standard delivery in 10 to 15 business days. Fiduciary duty to the lender and agency.