The Fix Without a Story
The troubleshooting post that ends with "reinstalled it, no idea why it was broken" is more useful than the one that ends with a clean paragraph explaining exactly what went wrong. Not more satisfying. More useful.
The honest version only claims what the author actually watched happen: thing was broken, I did X, thing works now. That's the whole epistemic content, and it's airtight, because it's just a report of an observation. Nobody can misuse it, because there's nothing in it to misuse beyond "try X." If X doesn't help you, you've lost five minutes.
The confident version claims more than that. It names a mechanism: a stale lock file, a corrupted cache entry, a race between two services starting in the wrong order. Somewhere between "I did X and it worked" and "here's why," the author crossed from reporting to inventing. Usually they didn't verify the mechanism. They verified the fix, then worked backward to a story that would explain it, because a fix without a story reads as incomplete and a story makes the post feel like it earned its length.
This is the part that actually costs something. Readers don't take away "X worked for someone once." They take away the mechanism, because the mechanism is the part that's portable. "It was a stale cache" generalizes. "I reinstalled it and it worked" doesn't, by design, it refuses to generalize. So the next person with a superficially similar symptom goes and clears a cache that was never the problem, or restarts a service in a different order because some blog post implied ordering mattered, and when that doesn't fix their issue they assume they've misdiagnosed rather than that the original diagnosis was fiction with a plausible shape.
The tell is usually in the specificity. A real root-cause finding comes with evidence of the causal chain being traced, not just asserted: a log line, a stack trace, a before-and-after of some internal state. If the writeup jumps straight from "it broke" to "here's the mechanism" with no artifact in between showing that mechanism was observed rather than inferred after the fact, that's a backfilled story wearing the outfit of an investigation. It reads confident because declarative sentences always read confident. "The cause was a race condition" and "I'm not sure, but it was probably a race condition" convey the same amount of actual knowledge about half the time, and only one of them admits it.
None of this is an argument for not finding real causes. A real root cause, traced and verified, is worth more than either version, obviously. The argument is narrower: between a fix with no explanation and a fix with an explanation nobody checked, the one with no explanation is closer to the truth, because it's not claiming to have one. Confidence is being used here as a stand-in for rigor, and the two aren't the same thing and don't reliably travel together. A post can be completely honest about its own ignorance and still be the more trustworthy artifact in the room.
The genre deserves its own respect rather than getting treated as the unfinished version of the real post. "No idea why, but here's exactly what I ran" is a complete, bounded claim. It's the format that matches what debugging actually produces most of the time: a fix you can reproduce and a cause you can't, because you stopped looking the moment the symptom went away, which is also exactly what everyone else does. The polished writeup just declines to say so.