I remember Shabir Ally also mentioning this in one of his lectures on the numerical relationships in the Qur'an. He said something of the sort that the "even/odd" relationship got established in a natural way, that the total number of verses wasn't already accepted and the 6,236 number was the one Muslims generally accepted by the 20th century. I always assumed that he was talking about only the difference of opinion on the matter of whether Bismillah... (at the beginning of almost all Surahs) should be counted in the total number of verses or not. The above citation shows that there was more to it.
I find it odd that there is resistance to accept this idea. Shabir Ally considered this as a strength of the "even/odd" relationship; an additional evidence for the uniqueness of this phenomenon. The most valid/popular criticism of the "even/odd" relationship has been that this was designed by whoever the human author of the Qur'an was (i.e. Prophet Muhammad). Unlike other relationships like the "golden ratio" one, a human could easily have designed and structured the Qur'an to conform to this "even/odd" relationship. All that human would've needed to know was addition.
However, the idea that the Qur'an didn't have a universally accepted number of verses, provides evidence that no human knew about this "even/odd" relationship prior to its discovery in the 20th century.
If Prophet Muhammad had himself painstakingly worked to create the "even/odd" relationship, even if we were to accept that for some reason he chose not to disclose it and therefore risk wasting all his hard-work; it becomes quite difficult to accept that he would further choose to not even, at the very least, disclose the total number of verses either.
At this point the critics' claim devolves to as follows, Prophet Muhammad himself designed the structure of the Qur'an in order to conform to these mathematical "miracles" and then, apparently, he did pretty much everything to make sure that no one found about it.
