Earlier quoted context omitted.
This means that a small difference in skill leads to a large difference in sales That does not follow at all from your initial observation. You could just as easily state that a small difference in exposure leads to a large difference in sales, thereby taking skill out of the equation again, and thus supporting the GP's initial argument.
But exposure in the form of publishers costs money. Therefore publishers try to find authors whose work will sell, meaning that skill also makes exposure easier. And once your books sell your skills are validated, so you will have even more opportunities. But if the books doesn't sell enough then you are thrown away, and the publisher lost money. If you used your own money then you lost money betting on yourself, mon…
It is precisely advantages like this that drive the "trickle up" effect. To use your original example, let's say one author legitimately wrote a book that was 10% 'better' than the next best effort. As a result, many more people bought her book (let's leave aside for now the notion of winner-take-all markets, which is another factor in the skewed distribution), and now she (and her publisher) are enriched. On her next book, they can afford a lot more promotion, and with her name recognition, her next book is a huge success, even though it is 10% worse than her nearest competitor. Now they've both written two books, and they're both 1-1 in terms of quality. But the one with the first good book has two best-sellers, and the second is a failure who will probably drop out.
No one is saying that the first author is morally wrong, they're simply saying that if we, as a society, want to reward people according to their skill and effort, the current system fails in numerous cases because it is heavily biased in favor of existing wealth over talent.