The problem with this is that as long as there are so many confounding factors, there is no reason to assume that the success of someones parents was "their own" as opposed to a result of one or more of the confounding factors such as a better network, and access to better education.
If you are going to go that route a better measure would be how much upwards mobility someones parents have demonstrated proportional to their starting point.
Of course that still suffers from a number of confounding factors (e.g. we don't know how much is down to luck)
That would be true if we were trying to find the exact reasons for success.
If you are just trying to promote success, then it stands to reason to reward success of the parents, incase some of those confounding factors are hiding actual fitness. After all, if the null hypothesis is true, you aren't losing anything by promoting some people over others, if they are all equality capable, anyway.
Does it? We don't know that without actually testing. It's very much possible that providing less assistance but to more people will have a much wider impact. If the parents truly are responsible for their own success, there is for example good reason to hypothesize that they may be able to compensate for reduced support with their abilities.
The problem is that measurable success is going to change the experiment.
If you consider that there is no constant landscape, it's always changing and always an unknown, the best option is to try a plethora of different ideas, hoping some will succeed.
It's more like fighting a war in an unknown battlefield. If you play it just one way, you are more likely to lose than if you try many different strategies at once.
The ones that succeed will be self-sustaining, for as long as they are successful, anyway, and gain you enough ground to compensate for the strategies that didn't work.
If this wasn't true, nature would only have produced one single organism rather than a billion or more.
If we don't vary the amount of support people get, then we won't be able to take advantage of situations where more support (or less) would result in greater success.
And given that everything is constantly changing, there is no way to find a perfect level of opportunity to give to people.
So, a system that provides a varying level of support spread out against the population will do better nearly all of the time than one that provides an equal amount to everyone.
I wasn't advocating for perfect exactly measured equality, I was arguing for a distribution that wasn't over the top insanely unfair to 99+% of everyone.
If you are going to go that route a better measure would be how much upwards mobility someones parents have demonstrated proportional to their starting point.
Of course that still suffers from a number of confounding factors (e.g. we don't know how much is down to luck)