There is a strange experience that sometimes accompanies slow change. You can spend years watching something develop, talk about where it appears to be heading and even make decisions based on that expectation, yet when the thing you anticipated finally becomes concrete, it can still feel surprising.

“I knew this was coming” and “I can't believe this happened” can coexist more easily than they seem to.

Part of the reason may be that prediction and adaptation are different activities. I can believe something is likely without living as though it has already happened, and most of the time that's a useful distinction. If every plausible future immediately reorganized my expectations, I would spend my life adapting to events that never arrived.

Uncertainty gives us a reason to wait.

Suppose a business appears to be declining. Revenue has been falling, customers are leaving and the owner has been covering losses personally. Someone looking at the numbers may say that the business probably can't continue this way indefinitely. That doesn't tell us whether it closes next month, survives another five years, finds a new customer, changes its model or unexpectedly recovers.

The direction can look clearer than the destination.

While the business remains open, people continue interacting with the business that exists. Employees go to work. Customers place orders. Bills get paid. Someone may quietly prepare for the possibility that it fails without behaving every morning as though today is definitely the day.

Then the doors actually close.

The event can still land differently from the prediction because something has changed. Yesterday there was a struggling business that might fail. Today there is a closed business.

Those aren't the same condition.

This becomes harder to perceive when change happens gradually because each individual step can be small enough to absorb. A price rises a little. A service becomes slightly worse. A process takes longer. An institution loses one capability and compensates somewhere else. People adjust, and after the adjustment the new condition becomes the starting point from which the next change is experienced.

Nothing requires us to assume that people are unaware of what's happening. They may see it quite clearly. They are also living during the transition rather than observing it afterward from a completed timeline.

That difference matters.

Looking backward, a sequence often acquires a shape that wasn't available while we were inside it. We know which warning mattered, which temporary improvement didn't last, which event became a turning point and which apparent turning point turned out to be noise. Earlier possibilities disappear from the story because we now know they weren't the possibilities that occurred.

Living forward is messier.

A trend can continue.

It can reverse.

Several trends can interact.

Something nobody was watching can become more important than the thing everyone was watching.

Even when the general prediction proves correct, the path can be wrong.

That's one reason I would be careful about interpreting surprise as evidence that someone failed to understand what they had previously seen. The new event may contain information the earlier prediction didn't. “I knew this was coming” can mean that the event fits the broad pattern someone expected. “I didn't think it would actually happen” may mean they had never assigned certainty to that expectation, or that its concrete form carried consequences they hadn't anticipated.

Sometimes, though, there really is a lag between what we've concluded and what we've allowed to change our behavior.

I've experienced that too.

An arrangement can look increasingly unsustainable while I continue treating it as basically normal because it still works today. I can recognize a dependency as dangerous and postpone replacing it because other problems are more immediate. I can know that something probably won't last while still being surprised by how much of my routine depended on it when it finally disappears.

Knowing the argument isn't the same as having reorganized around its implications.

The difficult question is when we should.

Adapting too late has obvious costs. If I wait until the basement is flooding to discover whether the sump pump works, I've chosen an expensive moment to learn something I could have learned earlier.

Adapting too early has costs too. I can spend enormous amounts of money protecting myself from failures that never occur. I can abandon useful arrangements because I became too confident in a prediction. I can interpret every new event as confirmation that the break I've been expecting has finally begun.

Preparedness can become another way of overfitting the future.

So the useful distinction isn't between people who see and people who don't. It's between different levels of commitment to an uncertain expectation.

I can notice something.

I can consider it plausible.

I can decide it's likely enough to prepare for.

I can make changes that are useful even if I'm wrong.

Or I can reorganize everything around the assumption that my prediction is correct.

Those decisions shouldn't require the same amount of evidence.

This also changes how I think about long timelines. When something finally goes wrong, there is a temptation to search backward for the moment when the process “really” began. Sometimes that produces useful history. Causes accumulate, incentives change, earlier decisions constrain later ones and arrangements that appeared stable may have contained problems for a long time.

There is rarely only one defensible starting point.

Choose a different question and the timeline can change. The immediate failure may have a proximate cause from last week, an organizational history extending twenty years and a broader background much older than either. Those explanations can coexist without one secretly being the true timeline.

Short timelines can hide accumulated causes. Long timelines can create their own illusion by making everything that happened afterward look like part of an inevitable trajectory toward the outcome we already know occurred.

Neither direction gets hindsight for free.

That's especially important because predictions affect the people making them. Once I believe I understand the trajectory, new information becomes easy to classify as another step along the same path. Sometimes it is. Sometimes I've become so good at explaining events through the model that I've stopped noticing evidence that should change it.

Reality still gets another turn.

Maybe the anticipated break finally happens and confirms much of what I thought.

Maybe it happens for reasons I didn't anticipate.

Maybe the thing I spent years calling decline turns out to have been transformation into something I hadn't learned how to recognize yet.

Maybe the decline was real, but my proposed cause was wrong.

Being surprised can therefore contain information rather than merely reveal that my emotions hadn't caught up with my analysis.

I still think there is something worth learning from the experience of seeing an event coming and being shaken when it arrives. Prediction can make us feel more prepared than we are because describing a possibility is much easier than living through its consequences. Thinking seriously about those consequences beforehand can expose dependencies we hadn't noticed and give us options while there is still time to create them.

The goal isn't to emotionally inhabit every bad future in advance.

It's to ask what preparation makes sense while remembering that the future hasn't agreed to our model yet.

Sometimes that means changing something now. Sometimes it means keeping a backup, preserving another option or watching a particular signal more carefully. Sometimes the sensible response is simply to know what we would do if the situation changes and continue living in the world that still exists.

Then, if the anticipated thing finally happens, we may still be surprised.

That doesn't necessarily mean we failed to see it coming.

It means seeing the path and arriving at the destination were never the same experience.