Only with an extremely oversimplified version of events. You’re essentially saying that selection is 100% effective. Anything that is big enough to be selected is always selected. This ignores one very big problem: noise. There are many, many things affecting reproduction that have nothing at all to do with genetics. And this fact alone is a major limiting factor on the scope of what selection is really able to accomplish. (It’s certainly not the only countervailing factor).
If beneficials and deleterious mutations were of equal strength and frequency, then we might get away with saying that noise doesn’t matter, since it affects both sides equally. But as we know, that’s not the real world. Since beneficials are very rare, the overriding of noise is much more problematic. Our rare beneficials are significantly likely to be lost.
Our DFE shows a threshold of selection. It should also show a horizontal threshold of noise intersecting the Y axis. I believe realistically speaking, virtually none of the selectable beneficial mutations would go above that threshold.