That's too simple. It's not just a question of choosing between 20kg of ballast or 20kg of KERS. KERS is expensive - Booth puts the number at £5-8m, which is, according to him, equivalent to enough wind tunnel time to gain them more than the 4 tenths KERS would be able to. The system also needs to be integrated into the package, it needs cooling, it needs specific testing, it will need lots of troubleshooting - time, that can be spent elsewhere, be it setup work or solving problems on the mechanical, hydraulics, engine side.
How do you achieve seconds? On an incremental basis. You don't magic a 3 second gain out of the ether. There are certain factors involved in running KERS;
Do you want to carry the extra 20kg of mandated weight as ballast or an 80 bhp power boost. One saw the largest effect of KERS during qualifying where it was combined with DRS effectively squeezing the pole time.
The argument that Marussia need seconds as opposed to the tenths KERS provides was used by Nick Wirth, a failed design engineer.
Also, Team Lotus and i think HRT too used the same reasoning to not run KERS in 2011. So whether it is the right decision or not - it's not just the idea of a "failed design engineer" as you want to put it.