The question is whether they're more concerned with national strategy or election strategy. The optimal path seems almost diametrically opposite for the two considerations.
Bush's best chance for winning the election now is probably to fail spectacularly at fighting terrorism. Switch to radically hard-line kill-our-enemies rhetoric, ramp up unfocused aggression everywhere, and watch the atrocities roll in. If he's perceived as the tough guy, maybe it doesn't matter whether he's succeeding. He just needs to make voting for Kerry look like capitulation.
The Berg murder is exhibit A. It's too early to know what effect it's had on the population as a whole, but my impression from reading wavering hawks who were on the verge of defecting from Bush support is that for them, it's completely negated Abu Ghraib. So I expect it to give Bush a small upward kick and erase some of Bush's poll losses of the past couple of weeks.
Historically these things are temporary. But if it goes on; if, say, there's some highly visible terrorist attack or prisoner-killing every week or two, that could be enough to give him an ongoing boost. On the other hand, the repetition of shock turns it into weariness, which could have the opposite effect.
