The fermentability range of a yeast is given by the supplier based upon the results from fermenting a standard laboratory wort which has a given fermentability. When you change the amount of fermentable sugars (such as with changing the mash temperature or introduction of multiple saccharification rests) of a wort from this 'laboratory wort', it will produce a greater or lesser the amount of sugars available for the yeast to consume. In your case, having three rests within the alpha and beta amylase activity range should produce a maximum fermentability of the wort, and the program does reflect that by setting the attenuation potential of the wort higher and predicting a lower FG. If you were not targeting a high degree of attenuation from the wort, why would you choose this mashing profile?
Having said that, the prediction of FG is subject to a number of influences from the process (mashing time/temperature, yeast health, fermentation profile) which cannot be well modeled. The model for adjustment of the wort fermentability used in BeerSmith, which can be seen under 'options' > 'advanced', is pretty standard.