A direct comparison between two combatants abilities in a fight that just takes in to account attacking and defending, nothing else, could have some value. It would ignore maneuvers, blocks, declarations, grapples, supplemental actions, and a whole slew of other things that would make the comparison pretty meaningless, but it would be an interesting at a glance type of thing IF:
1.) We knew the computations behind the arbitrary values, and we knew the reasons those computations were chosen.
2.) The analysis actually managed to get the basics right
This fails on both accounts. On 1) we are given arbitrary numbers, and an arbitrary formula and told that it works. On 2), we can clearly see that even without explanation its wrong, because it values armor and defense as equal, when this isn't even remotely the case. I suspect that some kind of binary non-dynamic analysis was done, that only compared things against each other one at a time, rather than the way these skills all interface together. In a one dimensional analysis, it would be possible to think that armor is as good as defense because you aren't accounting for the way weapon value stacks with attack value, and the way this effects the results of defense rolls or such.
A correct analysis probably wouldn't be able to give us a single number for each stat, because dynamic interaction would make different things valuable in different circumstances. Facing a high weapon value low attack value opponent? High defense is a hell of a lot more valuable than a high armor. Is a combination of high attack high initiative worth more than a combination of high defense high armor? That answer can't be given by simply taking each skills arbitrarily assigned value and calculating from there. You have to know what combinations work well together, rather than what works in isolation.