The scenarios are familiar to anyone living with ME/CFS, precisely because they are not occasional:

• Not knowing in advance — the plan agreed to on Tuesday, for a Saturday whose capacity cannot be known
• Waking up unable to manage what was planned — the day arriving already spent
• Symptoms changing during the day — a morning that fit the plan, an afternoon that does not
• Wanting to attend but being unable to — the hardest version, because the wanting was real
• Other people being disappointed — and the weight of carrying that on top of everything else
• Needing to change a plan after already confirming — the second message costing more than the first

What these situations share is that none of them are planning failures. The plan was reasonable when it was made, with the information available at the time; the day then delivered new information. Responding to it is sense, not unreliability.

Between attending the original plan and cancelling it lies most of the useful territory:

One of these usually fits — and offering it (“I can't do the full day, but I could do the morning”) often lands better than a plain cancellation, because it keeps the wanting visible.

Short versions, ready for the moment the change needs announcing:

• ““My symptoms have changed and I can't manage the original plan.””
• ““Could we make it shorter?””
• ““I won't know until nearer the time.””
• ““I need to rearrange rather than push through.””
• ““I can't manage today, but I'd still like to find another way to do this.””

A brief explanation may be enough when you do not have capacity for more. One accurate sentence about symptoms covers what most people need to know; the full account is owed to nobody, and offering it is optional rather than expected.

Other people can feel disappointed by a changed plan — that is a real response, and pretending it never happens helps nobody. What is worth separating is the disappointment from the blame: someone can be sad about a cancelled dinner and still understand why it was cancelled, and the sadness is not evidence that the cancelling was wrong.

The disappointment runs both ways, too. The person who could not attend is usually the more disappointed one — they wanted the thing, and they also have the symptoms. Carrying all of that alone is the arrangement worth questioning.

Changing a plan is not automatically the same as being unreliable or uncaring. Reliability shows in how the change is handled — the message sent, the alternative offered, the honesty of the reason — and that is reliability of a real kind, even though it looks different from the version that never cancels.

When most plans change, a particular difficulty sets in: embarrassment at sending yet another change, and the temptation to stop making plans altogether — because a plan never cancelled cannot disappoint anyone. That solution costs more than the cancellations did. The isolation of not planning is heavier than the awkwardness of moving things, and people who care would rather move the plan than lose the person.

What helps instead is building the changing in from the start:

• Flexible invitations — “come if you're able” asks nothing of the answer
• Provisional plans — agreed as provisional out loud, so a change arrives as a possibility, not a shock
• Same-day confirmation, where that suits — deciding on the morning, when capacity is at its most knowable
• Plans with a built-in smaller version — the visit that can shrink to a doorstep hello without being cancelled

None of this is a counsel to avoid planning. It is the opposite: plans made moveable are plans that keep getting made.

The last piece is the one nobody completes: not knowing, in advance, how a day will go — and making plans anyway. “I won't know until nearer the time” is an honest sentence, a complete one, and a surprisingly easy one for most people to work with. Certainty about capacity is not a prerequisite for having a life with things in it; a plan that might move is still a plan, and the alternative — waiting for dependable days that do not arrive — waits forever.