When two addresses seem to disagree

The situation

You have two windows open, or you looked at one route and then another, and they are not showing the same thing. Something is different and you did not expect anything to be.

The first useful move is to notice how strong an observation you actually have. Two pages seen minutes apart, on a screen, remembered rather than compared, is a weaker observation than it feels like, and most reported divergence dissolves at this step.

The full list of causes

1. The two pages were fetched at different moments, and the thing that differs is something that changes.
2. One of them was not fetched at all, because your browser already had it.
3. You are signed in on one and not the other, which changes what a page shows in ways that look like missing content.
4. The two pages are not the same page. On a route you use rarely, it is easy to be one level away from where you think you are.
5. One of the two is not the market.

Why that order

It is the order of frequency, and frequency is the only sensible order in which to check things that all take the same amount of effort to rule out.

The first four cost seconds each. The fifth costs real attention and produces a conclusion that changes what you do next. Spending the attention before the seconds is how readers end up alarmed by their own cache.

How to make the observation stronger

1. Load both again, deliberately, within a minute of each other.
2. Check the signed in state on each before comparing anything.
3. Compare the same page rather than two similar looking pages.
4. Compare text rather than impressions, because impressions of a layout are unreliable and text is not.

What a difference is never evidence of

It is never evidence that the market keeps separate copies per address, because that is not what an address is. The not a backup study deals with that belief at the root.

It is also never evidence about which route is better, because a route showing you a stale page is not a worse route. It is the same route that your browser did not ask.

The one thing that survives all four checks

If two freshly fetched, signed in, matching pages genuinely differ in substance, the interesting possibility is that one of the two hosts is not the one you think. That is a different subject with a different response, and it is set out in full in the difference that matters.

Reaching that possibility properly, having ruled out the others, is worth much more than jumping to it.

Why plurality creates this whole category

A reader with one address sees one page and has nothing to hold it up against. If that page is stale, or they are signed out, or they are somewhere they did not intend to be, they find out slowly and by other means.

With two routes open, a reader has a comparison. The comparison is genuinely useful and it also produces a stream of false positives, because all the ordinary causes above are things a single view would simply not have revealed.

The honest summary

Divergence is mostly a symptom of having a second window rather than a symptom of anything being wrong. The ability to compare is a real advantage and the price of it is a lot of noise, and knowing the causes is what turns the noise back into an advantage.

The disposition this section is trying to build

Not scepticism, which is exhausting, and not credulity, which is worse. Something closer to patience: a willingness to spend one minute on cheap checks before spending any attention on expensive conclusions.

That disposition is worth more than any individual fact in this section, because it generalises. A reader who has it will handle a symptom this site has never described, and a reader who does not will need a page for every possible case.

What this study is not. It is not a claim that differences never matter, and the one case where a difference matters a great deal has a study of its own.