[garborg]

Second flaw is even better than that -- author mistakenly asserts 'multiplying all values times 100' is the _only_ way to 'scale all data to ensure positive values.'