The rule, in one page

Every case on this site was sorted with the same test. It is short and it is not clever.

Why symptoms are useless for this

Because a symptom is what you can see, and what you can see is the far end of a long chain. A page that will not load looks the same whether the address was retired, the route was bad, the lookup failed, the string was mistyped or the whole network is having a bad evening.

Sorting by symptom puts all five of those in one pile and then invites you to guess. Sorting by actor separates them by mechanism, and mechanisms can be tested. That is the entire reason for the structure of this site.

The three questions on every case page

Responsible

Who was in a position to make it different. Named plainly, without hedging.

Usually blamed on

Where the blame actually goes, and why that is a natural mistake rather than a stupid one.

What follows

What changes once the attribution is right. If nothing changes, the page says so.

The third one is the point of the exercise. An attribution that changes nothing is trivia. Most of them change what you do in the next hour, and a few change it in a way that involves money.

Two biases that push in the same direction

The first is that visible parties absorb blame. The market has a name, a front page and a logo. The routing between you and it has none of those. When something fails, the mind reaches for the party it can picture, and it can picture the market.

The second is that comfortable answers get less checking. Bad luck, postal loss and network trouble all have the property of accusing nobody and requiring nothing. An explanation with those two properties will be accepted faster than one that implies you should change a habit.

Both biases push blame away from the reader and towards whatever is large, distant and unable to answer back. The corrective is to look at your own actions first, not because you are usually at fault, but because you are the party least likely to be examined.

Where the rule is hard to apply

• When two actors could both have prevented it. The one closer to the decision owns it, and the other one is context.
• When the failure is a rule rather than an event. Rules belong to whoever set them, not to whoever applied them to you.
• When the evidence is only your own memory. Memory is evidence about you and weak evidence about anybody else.
• When the answer would be inconvenient. That is the case where the rule is most worth running properly.

What the rule cannot do

It does not tell you whether something is happening right now. It does not measure anything and it produces no numbers. It will not tell you whether an address is reachable, because that is not a question this site answers at all.

It also does not settle blame in a way anybody has to accept. A dispute has its own rules and its own decider. This is a tool for working out where to aim your attention, which is a smaller and more useful thing.

• Every case, with its actor
• What you can actually observe
• What a wrong attribution costs