It is a partially correct argument. If you apply a method to something small and it has an x% effect then your compensation will always be limited by that upside. Applying that same method (or knowledge) to a larger entity potentially has a larger upside and so your room for negotiation expands. Whether or not the difference in percentage is the same is of course debatable but if there is an upside in both cases then the upside is likely to be larger in absolute terms when applied to the larger company. And so those large companies will be able to pay higher compensation for the same skill levels.
That's not at all the same as playing roulette and it does not mean this argument is not correct. It's correct in principle, you may want to argue about the degree to which it transfers from smaller companies to larger companies.
It failed to address the fact that it may not even be possible to increase anything. What if revenue is limited by the product? Then it's not the CEO that can impact it. What if the odds of ever changing anything fall as the company grows? Then there's no scale gain for the CEO's work.
There's a huge number of problems that may impede that scenario, and the argument just distracts people away from those.
That's not at all the same as playing roulette and it does not mean this argument is not correct. It's correct in principle, you may want to argue about the degree to which it transfers from smaller companies to larger companies.