New SPERT-Beta release for July 2016!

It’s been a long time since I played around with the next edition of Statistical PERT:  SPERT-Beta, which will use the beta distribution functions inside Microsoft Excel.

I’ve retooled the SPERT-Beta template so it’s simpler with fewer choices.  More importantly, I’ve changed the philosophy I used to create the initial SPERT-Beta template.  This new pre-release template of SPERT-Beta uses ratio scales that more accurately create a skewed probability curve, placing the most likely outcome — the mode — at or very near the top of the curve.  In my earlier version of SPERT-Beta, the mode was often nowhere near the top of the curve for many of the most likely confidence choices.  Now, choosing any of the most likely confidence choices will still place the mode in its proper place:  at the top of the mountain.

With this new version, choosing “Medium Confidence” in the most likely outcome will yield probabilistic results that closely mirror what you would get by running a Monte Carlo simulation using Palisade’s @Risk7 (specifically, the RiskPERT function).

I ran a comparison between SPERT-Beta and @Risk7’s RiskPERT, and the results were very often very close, even for skewed uncertainties.

As always, feel free to download and use, but be aware that this is a development version and may still contain serious flaws anywhere in the spreadsheet.

Leave a Reply

Your email address will not be published. Required fields are marked *