A MathOverflow answer to a question on potential computational experiments in maths requested the computation of odd "strongly pseudoperfect" numbers. There are multiple incompatible definitions of the term: the relevant one here is a number n for which there is a subset of its factors which is closed under the bijection between a factor and its cofactor and which sums to 2n. Sieving and testing subset sum with some very minor optimisations computes strongly pseudoperfect numbers up to 50000000 in a few hours. Python code, results.