The moments where switching is the right move

When it is right

1. One route is not answering and you have not tried the others. This is what the set exists for.
2. One route is slow enough to be blocking you right now, and you have nothing unsubmitted in front of you.
3. You are starting a visit and are deliberately rotating, which is rotating on purpose rather than switching.

When it is harmless

• Out of curiosity, at the start of a visit, with nothing in progress.
• Because you want to confirm you can still reach the market by a route you have not used lately, which is a reasonable thing to check occasionally.

When it is a reflex rather than a move

1. Anything to do with signing in, credentials or the account, which is one object on all three routes.
2. Anything that looks like content missing, which is far more often a stale page or a signed out view. See real and apparent.
3. Immediately after two other routes behaved the same way, which points at something common rather than at the routes.

The test to apply before switching

Ask what you expect to be different afterwards. If you can answer that in a sentence, switch. If the answer is that you are not sure but it might help, you are pulling a lever that is not connected, which is the subject of what switching cannot fix.

It is a low bar and it filters out most wasted switches.

The check to run before you go

Is anything typed and unsubmitted on this page. If yes, deal with it first, because that is the only thing a switch can actually cost you.

That check takes a second and it is the entire safety procedure for this section.

Why timing beats frequency

How often you switch matters far less than when. Ten switches at the starts of ten visits cost nothing at all. One switch in the middle of writing something long costs the thing you were writing.

Advice that says switch more or switch less is missing the variable. The variable is what is on the screen when you do it.

The bias worth correcting

Readers switch too readily when something looks wrong and too rarely when something is simply slow. The first is the reflex case and produces nothing. The second is the case the set exists for and is often skipped because moving feels like a bigger act than waiting.

Correcting both biases is most of the practical value of this section.

The summary

Switch when a route is failing you, at a moment when nothing is unsubmitted. Do not switch to solve a problem that is not shaped like a route. Everything else in this section is detail attached to those two sentences.

The question that replaces all the rules

What do I expect to be different afterwards. If the answer is specific, the move is justified. If it is a vague hope that something might change, the move is a reflex and it will cost you a minute and possibly a form.

That question also works in the other direction. A reader who can answer it and is not switching is probably waiting when they should be pasting, which is the more common of the two errors and the less examined.

What this study is not. It is not a rule about how often to change address, which is not a variable that matters on its own.