I'm thinking we should start keeping in mind Game Informer covers again. The October cover, which will be revealed at PAX in September, is definitely not TES V, so we don't have to start thinking about it yet. But Bethesda is going to announce the game before the cover is revealed so once October starts we should prepare for a potential announcement, which will, frankly, likely lead to disappointment. However, the way I see it is, the more GI covers turn out to not be TES V, the higher the chance that the next one will be the one. That would have sounded crazy a few days ago when we had no idea when TES V would be announced, but after the comments from Todd it isn't so crazy.
Statisticially, a outcome has no effect on a variable. For instance, if you roll a dice, and it comes up six, there isn't a more likely chance that it won't be six the next time. Just because a cover isn't ES5 dosen't make it more likely that the next one will be.
Time however does in face effect the outcome. If there is a 100 percent chance of an annoucement, as time passes, the chance that any given day will be a announcement increases.
So covers to game magazines have no impact, it's possible, what with so many unannounced games from develoers right now, that there won't be a cover at all, making probability in that instance extremely unreliable.