You are sitting in the parking lot with the engine off, and you have been sitting there for four minutes. You know this feeling. It is not laziness, exactly, and it is not fear. It is a low, flat dread that has nothing to do with the hour in front of you and everything to do with the hour you remember. Here is the uncomfortable part: that memory is wrong. Not slightly wrong. Structurally wrong, in a way psychologists have been able to reproduce in a lab for more than thirty years. The workout you are dreading is not the workout you did. It is an edit. And the editor has terrible taste.

Two selves walk into a gym

Daniel Kahneman spent the back half of his career on a distinction that sounds like a parlor trick until you notice it running your life. There is the experiencing self, who lives in the present moment and feels things as they happen. And there is the remembering self, who writes the story afterward, keeps the score, and makes the decisions. The experiencing self does the squats. The remembering self decides whether you ever squat again.

The trouble is that the remembering self does not keep an honest ledger. It does not add up the minutes and average them. It samples. Two rules govern the sample. The first is the peak-end rule: an experience is remembered mostly by its most intense moment and by how it ended. The second is duration neglect: how long the experience lasted barely registers at all.

Kahneman and his colleagues showed this with a study that still feels faintly absurd. Participants held a hand in painfully cold water for sixty seconds. In a second trial, they held a hand in the same cold water for sixty seconds, then for thirty more seconds in water that had been warmed by a single degree, still unpleasant, just less so. Then they were asked which trial they would repeat. Most chose the longer one. More total pain, remembered as better, because the ending was gentler. A later study with colonoscopy patients found the same thing: how much pain patients remembered was predicted by the worst moment and the final moments, not by how long the procedure took.

What this does to a workout

Now run a typical strength session through that machine. Fifty-five minutes. The warm-up was fine. The first few working sets felt strong, even good. There was a moment in the fourth set of squats where the bar slowed and your vision tunneled. And then, because the program said so, you finished with a final set taken to failure, grinding, failing on the last rep, racking the bar with your ears ringing.

The remembering self takes that hour and keeps two frames. The peak: the bottom of that fourth squat. The ending: the failed rep. The forty minutes of fine do not make the cut. That is duration neglect doing exactly what it does. Tomorrow, when you consult your feelings about going back, what you are consulting is a two-frame highlight reel of the worst parts.

The exercise psychologist Panteleimon Ekkekakis has spent decades mapping how exercise feels moment to moment, and his dual-mode theory explains why the peak is so reliably ugly. Below a certain intensity, how exercise feels is variable and shaped by your thoughts, your music, your mood. Above it, when the body is near its limit, pleasure drops for almost everyone, no matter how tough they are. A heavy set near failure lives above that line by design. That is not a flaw in the set. It is what the set is for. But it means the most intense moment of any serious session is almost guaranteed to be unpleasant, and the peak-end rule guarantees that moment gets a starring role in your memory.

And most programs are written to end there. The last set is the hardest. The top set comes last. Leave nothing in the tank. Finish strong. We have built an entire training culture around making the final frame the worst one.

The ramp-down experiment

In 2016, Ekkekakis, along with Zachary Zenko and the behavioral economist Dan Ariely, tested whether you could simply rearrange a workout to trick the remembering self. Participants did a fifteen-minute cycling bout. One version ramped the intensity up, finishing hard. The other version ramped it down, finishing easy. Same total work in both. The people who finished easy felt better right afterward, remembered the session as more pleasant, and, most importantly, forecast that a future session would be more pleasant too. Nothing about the physiology changed. Only the order did.

That forecast matters because it is the actual lever on behavior. A 2015 systematic review by Ryan Rhodes and Alison Kates found that how exercise feels during the session is one of the better predictors of whether people keep doing it. And a 2011 study led by Matthew Ruby found that people consistently underestimate how much they will enjoy a workout, partly because they fixate on the unpleasant opening minutes when imagining it, a habopsychologists call forecasting myopia.

Put those together and the parking-lot dread finally makes sense. It is a forecast, built from a memory, built from a peak and an ending, and every link in that chain is biased toward the worst moment. You are not dreading the gym. You are dreading a caricature of it.

Why "finish strong" is backwards

This is not an argument for training easier. The muscle needs the hard set. It needs the fourth squat where the bar slows. But the muscle does not remember. It adapts to the stimulus and forgets the story. The remembering self is the only one keeping notes, and it is the one who decides whether you come back.

So the fix is not to remove the hard work. It is to move it. Put the thing that produces the peak earlier, when you are fresh and strong anyway. Then let the session taper. End with a set you can complete with reps to spare, or a lighter movement you actually enjoy, or five minutes of easy walking. You have done the same work. You have simply refused to let the worst moment also be the last one.

There is a quieter benefit here too. A failed last rep does not just feel bad. It becomes evidence in a story about who you are: someone who cannot finish, someone who is losing it, someone who should probably stop. A completed last set tells a different story with the same hour of effort. You get to choose which story the remembering self files, and it costs nothing but sequencing.

Your next moves

  • Rewrite the end of today's session before you walk in. Take whatever your program puts last and move it to the middle. Finish with a back-off set at roughly eighty percent of your top weight, stopping with two or three reps still in the tank. Same volume, different final frame.
  • Never let a failed rep be the last thing you do. If a set goes badly, strip some weight and do one clean, controlled set immediately after. You are not training the muscle at that point. You are overwriting the ending.
  • Add a five-minute ramp-down. Easy walking, light rowing, a stretch you actually like, the same song every time. The body's last sample of the session should be calm, not gasping.
  • Rate the session the moment you rack the last bar. One number from one to ten, and one sentence about how the ending felt. Do it before you leave the floor, while the experiencing self is still in the room. This is the only honest witness you will ever have.
  • Before you skip, read the ratings, not your gut. Open your last three sessions and look at what you actually wrote. If the numbers say seven, seven, eight and your dread says two, trust the witness over the editor.

Where the log comes in

A training log is, at bottom, testimony from the experiencing self. It records what happened, set by set, before the remembering self gets its hands on the footage. Rep is built to make that testimony fast enough to actually give: log the set, see the PR, close the app, done, so the final act of every session is writing down something you completed rather than replaying something you failed. And when the parking-lot dread shows up, the log is the counterevidence, the forty good minutes your memory threw away. If you want a place to keep that record that you pay for once and own, you can find Rep at rep.lumenlabs.works. Either way, change the ending. The muscle will not notice. You will.