Pitfalls in computer simulation
Though generally avoided in computer simulations, in strict logic the rules governing floating point arithmetic even apply. For illustration, the probabilistic risk analysis of factors finding the success of an oilfield exploration program engages combining samples from various statistical distributions by using the Monte-Carlo methods. These comprise normal, lognormal, triangular and the uniform distributions. Conversely, a sample from a distribution cannot maintain more important figures than were representing in the data or estimates which established those distributions. Therefore, abiding through the rules of important arithmetic, no result of a simulation can maintain more important figures than were representing in the input parameter along with the least number of important figures. If, for illustration the net/gross ratio of oil-bearing strata is identified to only one important figure, then the consequence of the simulation cannot be more exact than one major figure, though it may be presented as having three or four important figures.