NetSimulator by NetSimR
Simulation Report
Generated 14 November 2026, 22:38
Key results
Losses ceded to the reinsurance layers per simulated period.
Model settings
- Simulations
- 20,000
- Seed
- Fixed at 1
- Processing
- Single process
- Distribution
- Poisson
- Parameters
- lambda (mean claims) = 625
- Distribution
- Burr
- Parameters
- alpha (first shape) = 0.9314, gamma (second shape) = 1.2037, Scale = 23,086.3758
- Pareto slices
- Not applied
- Severity cap
- 1,000,000
- Structure
- Limited Layer
- Limit
- 750,000
- Deductible
- 250,000
- Reinstatements
- Unlimited
- Structure
- None
Gross, ceded and net
The structures are reinsurance layers, so the modelled totals are the ceded losses. Net is gross minus ceded.
| Metric | Gross | Ceded | Net |
|---|---|---|---|
| Mean | 47,335,635 | 13,460,222 | 33,875,413 |
| Share of gross mean | 100.0% | 28.4% | 71.6% |
| Standard deviation | 4,383,716 | 2,774,745 | 2,200,345 |
| Median | 47,200,390 | 13,339,890 | 33,836,420 |
| VaR 99% | 58,066,241 | 20,537,001 | 39,126,206 |
| VaR 99.5% | 59,213,612 | 21,385,738 | 39,712,635 |
| TVaR 99.5% | 60,789,084 | 22,329,438 | 40,402,726 |
Gross is after tail adjustments and the severity cap, before reinsurance. Each column's percentiles come from its own simulations, so net VaR is not gross VaR minus ceded VaR.
Gross, Ceded, Net by return period
How the structures change the loss at each return period, on a log scale.
Layer metrics
Figures a reinsurance price is usually built on, for the layers described under Model settings.
| Metric | Value | What it means |
|---|---|---|
| Chance the layers are hit | 100.0% | Share of simulations with any ceded loss |
| Average loss when hit | 13,460,222 | Mean ceded loss over the simulations that hit the layers |
| Expected loss | 13,460,222 | Mean ceded loss over all simulations |
| Loss on line | 1794.70% | Expected loss as a share of the each-and-every-loss limit |
| Aggregate capacity | Unlimited | No aggregate limit or reinstatement limit applies |
Statistics and tail risk
Summary statistics
| Metric | Value |
|---|---|
| Simulations | 20,000 |
| Mean | 13,460,221.79 |
| Median | 13,339,890.22 |
| Standard deviation | 2,774,745.15 |
| Coefficient of variation | 0.206 |
| Minimum | 3,772,920.03 |
| Maximum | 25,718,291.39 |
Percentiles, VaR and TVaR
| Percentile | Return period | VaR | TVaR |
|---|---|---|---|
| 50.0% | 1 in 2 | 13,339,890.22 | 15,677,375.96 |
| 75.0% | 1 in 4 | 15,283,210.48 | 17,071,883.42 |
| 90.0% | 1 in 10 | 17,053,629.98 | 18,578,472.72 |
| 95.0% | 1 in 20 | 18,182,906.53 | 19,581,093.43 |
| 97.5% | 1 in 40 | 19,236,868.63 | 20,524,358.13 |
| 99.0% | 1 in 100 | 20,537,001.48 | 21,622,794.33 |
| 99.5% | 1 in 200 | 21,385,738.42 | 22,329,438.28 |
| 99.6% | 1 in 250 | 21,611,872.78 | 22,541,086.33 |
| 99.8% | 1 in 500 | 22,290,796.48 | 23,189,162.16 |
| 99.9% | 1 in 1,000 | 22,953,133.43 | 23,819,291.22 |
Simulation accuracy
| Metric | Value |
|---|---|
| Standard error of the mean | 19,620 |
| 95% range for the mean | 13,421,766 to 13,498,678 |
| 95% range for VaR 99.5% | 21,223,429 to 21,611,858 |
| Simulations at or beyond VaR 99.5% | 100 |
The ranges show how much the figures could move from simulation noise alone. Narrow ranges mean the run was large enough.
The range for the mean is a normal approximation (the mean plus or minus 1.96 standard errors), which is rough for heavy-tailed totals; when no total is negative it is kept at zero or above. The count at or beyond VaR 99.5% includes totals equal to VaR, so ties (e.g. many zero totals) can make it large.
Claim frequency
Simulated number of claims per period, before severity and reinsurance.
| Metric | Value |
|---|---|
| Average claims per period | 625.06 |
| Standard deviation | 25.27 |
| Chance of no claims | 0.0% |
| 99th percentile | 685 |
| Maximum | 716 |
Claims per period
Charts
Distribution of total claims
Dashed lines mark the mean and the 99.5% VaR of all simulations.
Losses by return period
The loss expected once in each number of periods, on a log scale. Return periods are shown up to 1 in 2,000.
Cumulative distribution
Share of simulations with total claims at or below each amount.
How to read this report
- A large gap between the median and the tail percentiles points to a heavy-tailed outcome.
- VaR is the loss exceeded only with the stated probability. TVaR is the average loss in those worst cases, so it is always at least as large as VaR: TVaR 99.5% is the mean of the worst 0.5% of the simulations, rounded up to a whole number (here 100).
- The return period shows the same probability as a frequency: 99.5% corresponds to a 1 in 200 year event.
- Gross is before reinsurance. Ceded is what the layers pay, and net is what remains.
- The accuracy ranges reflect simulation noise only, not uncertainty in the chosen distributions or parameters.
- Figures reflect the tail adjustments and reinsurance structures listed under Model settings.