Where the model came from
The denial-anger-bargaining-depression-acceptance sequence comes from Elisabeth Kübler-Ross's work with dying patients in the 1960s. It was a description of what she observed, not a route anyone was supposed to follow, and she later said so explicitly. It was never validated as a model of romantic loss at all.
That matters practically, because the model gets used as a ruler. People ask whether they are 'in the right stage for week six', and the question has no answer. There is no schedule to be behind.
Why you loop
Most people revisit stages they thought were done. You reach something like acceptance, then a song or an anniversary puts you back in anger for two days. This is the ordinary shape of grief, and it is the part the neat diagram leaves out.
Looping is also not a sign the relationship was uniquely significant. It usually means your brain is still updating a large number of small assumptions — who you text at lunch, what you do on Sunday — and each one gets revised separately.
What stages are actually useful for
They are useful as vocabulary, not as a map. Noticing 'this is bargaining' can stop you from acting on it, because bargaining mostly produces messages you regret. Noticing 'this is anger' can keep you from treating a furious thought as a considered conclusion.
What they cannot tell you is how much longer. Any source promising a timeline per stage is inventing it; the variation between people is larger than the pattern.
What progress actually looks like
Duration and frequency, not intensity. A bad Tuesday in month four can feel as sharp as week one and still represent progress, if the previous bad day was ten days ago rather than yesterday.
The other marker is reach: whether a hard hour stays an hour or takes the whole day with it. When ordinary tasks survive a wave of feeling, the wave has stopped running the day — and that is a better measure than any stage name.