Your latter point is a tautology. You win arguments by making really good ones and gaining an advantage over those you don't. By the measure of "able to win arguments" the former viewpoint is better than the latter. If you're using other measures to evaluate viewpoints, you shouldn't do so in the framework of arguments.

Regarding this particular argument and how difficult it is to value ones that sound valid: at the time, this argument was quite cogent and I would have agreed with it. The author made a well-researched, well-formulated, convincing argument. He (along with many others, I'm sure) was later proven wrong, but at the time of the argument, he had a convincing case. Nothing wrong with that, in my opinion.

Often old articles with predictions that turned out to be horribly wrong come up on here. I think that whether the person turned out to be right or wrong is not very relevant. When you make an argument you are stating your opinion at that time, and so the quality of the argument should be measured in that context. That isn't to say that you can just make wild predictions without risk of losing credibility; again, your prediction can and should be measured at the time you make it, and I think that baseless ones should be critiqued appropriately.