Weather & seasonal downtime

Winter working is a risk your register can't see: it hits every exposed activity in a window, and it compounds — work that slips into the bad months loses more days, which slips it further. RunMonteCarlo models it Pertmaster-style: a percentage of working days lost per calendar month, sampled every iteration.

1. Tick Model seasonal downtime on 4 Ranges. A twelve-month grid appears.
2. Enter the % of working days lost per month — e.g. 15–20% across Nov–Feb for exposed civils in the UK, zero for the fit-out months. Each figure is sampled triangular between 0.5× and 1.5× per iteration, so bad winters and mild ones both happen.
   One profile for the project: activities crossing lossy months take proportionally longer.
3. Tick Date-aware for weather-driven work. Fast mode resolves each activity's lost months once from its pre-weather dates; date-aware re-evaluates as activities slip, so work sliding into January picks up January's losses — the slip-into-winter compounding that fixed calendars miss. Recommended whenever weather is a real driver.
4. Run, then decompose. Switch the results tornado to Variance: weather appears as its own source alongside duration uncertainty, threats and opportunities, answering "how much of my P80 is weather?" directly. The report's method statement prints the enabled months and which mode you used.