Jay may correct me, but they took a risk with SPEs.
That's what "probabilistic model" means. In practice we categorize SPEs into classes corresponding roughly to orders of magnitude. And those within a category have roughly equivalent effects on organisms and equipment. It's not unlike earthquake magnitudes. There's a big difference in effect between a Richter 6.0 earthquake and a Richter 7.0, but little difference (in effect) between a Richter 6.0 and a 6.2. Then for each category, a fairly complicated random variable is created with parameters drawn from historical observations. Then a likelihood function is computed that is the convolution of the categorized parametric model and the mission window -- time and orbital path. It's heavy-duty statistics, but it is suitable for quantifying risk.
During a solar max peak year, you can expect up to about 6 events in all directions within the ecliptic, of the most hazardous magnitude -- the ones that would have a significant biological effect on the astronauts in the Apollo CSM, up to and including ultimately fatal exposures. If you fly a 12-day Apollo mission four times a year, you have a reasonably good likelihood of success because all your operations take place on one side of the sun.