Results indicate that climate conditions similar to the 2005 season could occur again, but they are rare. Note that given the stochastic nature of the tropical cyclone formation process, it is possible that variations in the rate of tropical cyclone formation could occur even when forced by the same climate conditions. Therefore a third climate model with finer horizontal resolution is employed to directly simulate tropical cyclone formation rates, and these lows are then detected in the model output. The tropical cyclone climatology of the model is evaluation when forced with observed SSTs (an AMIP-II style simulation) over the Atlantic for the period 1951-2011. To estimate the stochastic component of tropical cyclone formation, the variation in directly-detected formation in a 100-member ensemble forced with 2005 SSTs is determined. Results show that the stochastic component of variability in tropical cyclone formation in the Atlantic is about 40-50% of the mean, in agreement with previous results (Done et al. 2014). Even when allowance is made for the possible amplification of tropical cyclone numbers by this stochastic component above what is generated on average by the climate conditions, the formation rate observed in 2005 remains rare. Thus the year 2005 might serve as a benchmark of the maximum formation rate in the Atlantic for risk management purposes.