This is still a Strawman, as well as a Red Herring. Nobody is saying that it's "like real life." That's you twisting the argument and distracting from the point. We're saying that it's MORE realistic or in the case of Frogreaver, that it's "closer to real life." We are not saying that it is mirroring the real world. How about you respond without twisting arguments and attempting to move the point away from what people are talking about?
Not to speak for pemerton, but I think that's what he disagrees with. The idea that a GM making a decision yields a more realistic result than other methods.
While you could likely provide examples of a GM making a decision that seemed to result in a more "realistic" (by which I think we mean something like "mathematically likely") outcome, others can just as easily provide examples where a player makes a decision that results in a more "realistic" outcome.
It's a preference that some folks in this thread have, and it's something that makes more sense to them, and that's fine....but it is in no way objectively more realistic than most other methods used in an RPG.
To illustrate, let's look at a basic example. The party is making their way along a path. They come to a fork in the path. Which branch shows more signs of traffic?
In a game like D&D, where the DM knows the surrounding area and its inhabitants and their goals and so on, he may declare the left branch as the more traveled because he has the map and knows that way leads to more populated areas, so common sense indicates that would be the answer. This is "realistic" in the sense that some form of logic is applied to the answer.
But what about a game where the surrounding area and inhabitants are not known by the GM ahead of time, but are instead determined through play? Why would the GM picking the left branch be more "realistic" than the right branch? Perhaps the game calls for a roll from the players, and then based on the results of the roll, the GM narrates things accordingly. The player rolls well, so the GM decides that the character is capable of accurately determining that the left branch sees more traffic.
The "realism" of the result is no different in either example. The "realistic" method used in the first example simply doesn't work for the second. So ultimately, what is being discussed is a preference in game mechanics and how they're applied.