It’s an annual ritual: we get to late summer, something goes wrong for Barack Obama, and the press piles on. This time it’s Syria — Obama is getting widely criticized for vacillation and changes of course.

The criticism, in turn, has also produced a good discussion of how to go about criticizing a president. Here’s a rule of thumb: If you want to criticize Obama’s handling of process, then make sure you do it right. Don’t turn process criticisms into an excuse for mind-reading or magical versions of the presidency.

One question that’s attracted a lot of attention is whether a president should be criticized on process and theatrics or whether the primary focus should be on the outcomes he produces. David Ignatius today broke with much of the D.C. punditry and argued that the commentators’ obsession with appearances of vacillation — and lack of leadership — is out of touch with how the public views these matters.

In my view it’s fair to criticize a president’s handling of process. Process analysis can often illuminate what’s really going on; after all, policy outcomes are often driven by not just what the president wants, but what others in the system want, and how the president went about dealing with getting his way in the face of opposition.

The problem arises when process analysis is infected with magical thinking and outright wrong assumptions about how the presidency really works. For instance, one fallacy occurs when presidents assume a George W. Bush model in which the key question is only whether the president is able to get his way on his first impulse — as if that’s unquestionably a good thing.

Another problem arises when analysts get process completely wrong. See, for example, a Politico article called “What’s Wrong With President Obama?” We get process criticism on Syria and on the Fed chair choice with a bald assertion that if only Obama was tougher, something would have been better (it’s not clear exactly what, or how). Which rapidly turns to mind reading:

The last wave of “what’s wrong with Obama” speculation occurred almost a year ago, following his somnambulist performance at the first presidential debate in Denver against Mitt Romney. That’s no coincidence. When Obama is bored, or tired, or frustrated — whether with himself or the House Republicans or the press — he can’t hide it.

In other words, sometimes events go wrong, and the press decide it must have been because of the president’s state of mind. There’s really nothing here about why the outcome in Syria is a bad one given the realistic possibilities, or anything about why the eventual Fed pick will be worse than Larry Summers (see also presidency scholar Andrew Rudalevige on Harris and Purdum).

For another example of getting process wrong, see Matt Miller’s column yesterday. Miller is explicit about some of his policy preferences: for gun control; against the Bush tax cuts; for a jobs bill. But he manages to attack not the opponents of those policies (that would be mainly Republicans in Congress) but Obama for “blinking” and “losing his nerve” in trying to get them done.

But it’s nonsense to claim gun safety legislation failed in the Senate because Obama lost his nerve. It lost because the votes weren’t there. This is just magical presidency thinking, where it’s assumed that any president can get any policy he wants simply by wanting it enough.

Back in the actual U.S. system of government, the president is only one actor in a system of separated institutions sharing powers. A president always has to bargain, cut deals and compromise. Presidents will sometimes flat-out lose. Not “blink,” but lose. That’s what happened on Larry Summers; Obama apparently wanted him, and his party didn’t. Obama could have showed he had “nerve” by nominating Summers anyway, but that might have cost him far more than accepting “no” for an answer.

The good news: The punditry on these late-summer swoons has improved of late. Reporters and bloggers rely more on the academic literature, and academic experts are getting better at sharing what they know with journalists. But as we’ve seen these past two weeks, there’s still plenty of room for improvement.