Fault coverage of a long random test sequence estimated from a short simulation

Forecasting the fault coverage of a long random test sequence from the results obtained from the simulation of a short test sequence looks like an inaccessible dream. As a matter of fact, a short test sequence detects mainly easy to detect faults, while the test length required for a fault coverage close to 1 depends mainly (almost only) on the most difficult faults. A new 2-parameter model, linking the fault coverage to the test length, was proposed recently. Although these parameters are very difficult to obtain, they can be roughly estimated from a relatively short simulation. According to a previous comment, the results can not be accurate. However, they are more accurate that ones obtained from other models proposed in the literature.