Recently Keefer and Verdini (Management Science, 39 (9), 1086-1091 (19
93)) and Lau, Lau and Zhang (IIE Transactions, 28, 183-192 (1996)) sho
wed that well-known PERT-type formulas for computing standard deviatio
n from fractiles are very unreliable, and they presented alternative f
ormulas that are more accurate. This note extends their results and pr
esent a formula that is simpler, more meaningful and yet more accurate
.