Somewhere in your notes app there is a number with a date next to it. A 315 squat by June. A bodyweight bench before the wedding. You wrote it in a January mood, and you meant it, and you were not being lazy or naive when you wrote it. You were doing honest math with the wrong inputs. The date came and went, and now the number sits there with the particular quietness of a promise you made to yourself and didn't keep. Here is the uncomfortable part: you would have missed that date even if you had done almost everything right. The mistake wasn't in the lifting. It was in the forecast, and the forecast error has a name.

The thesis that took twice as long

In 1994, psychologists Roger Buehler, Dale Griffin and Michael Ross asked university students finishing their honours theses to predict when they would be done. They also asked for a worst-case date, the one the students would give if everything went about as badly as it plausibly could. Then the researchers waited. Most students missed their own estimate. A large share missed even the worst-case date. Yet when the same kinds of students were asked to predict other people's projects, or were made to think carefully about how their own past projects had gone, the optimism shrank. The blind spot only appeared when they looked forward at themselves.

Daniel Kahneman and Amos Tversky had already named this pattern the planning fallacy: the tendency to underestimate how long a task will take, even when we know perfectly well that similar tasks in the past took longer. Notice the second half of that sentence. This isn't ignorance. The students remembered previous late projects. They just didn't believe those delays applied this time.

The inside view and the outside view

Kahneman and Dan Lovallo later described two ways to forecast anything. The inside view builds a story: I'll train three times a week, add five pounds each session, and in twenty weeks I'll be there. It feels rigorous because it has steps. The outside view ignores the story and asks a duller question: when people like me have tried things like this, how did it actually go?

The inside view fails in a specific way. A story is made of the things you plan to do. It cannot contain the things you don't plan: the cold that takes a week, the work trip, the tweaked lower back, the fortnight where you keep showing up but the bar simply will not move. None of those is likely on any given day. Across five months, at least one of them is nearly certain. The story has no slot for something, I don't know what, so it quietly assumes nothing.

And when a delay does happen, memory files it as an exception. That was a weird month. This one won't be like that. So the next forecast starts clean, built once again from the plan rather than the record, and the cycle repeats. Buehler's group found that the fix was not telling people to be realistic. Vague humility doesn't move estimates. What worked was making people recall how past projects had actually unfolded and explicitly connect that history to the task in front of them. The past only helps when you force it to testify.

Why the barbell is unusually good at fooling you

Strength training sets a trap that most projects don't, because early progress is genuinely fast. A new lifter can add weight to the bar almost every session for months, and that is real, not a fluke. But the inside view does what it always does: it takes the slope of the last eight weeks and draws a straight line into the future. Adaptation doesn't run in straight lines. Gains come fastest when you are furthest from your ceiling and slow as you approach it, so the same effort that bought twenty pounds last quarter buys ten this quarter and five the next. A plan extrapolated from the easiest stretch you will ever have is wrong on day one and becomes more wrong every week.

There is also a cruel asymmetry. A missed week cannot be reclaimed by doubling the next one. The body will not absorb the extra dose, and trying to force it usually costs you a second week. Progress is capped on the upside and uncapped on the downside. Any forecast that treats a good week and a bad week as cancelling out is already optimistic.

None of this means the goal was wrong. It means the date was invented, and that invented date is now doing something corrosive: it is turning a perfectly normal rate of progress into evidence of failure. You are not behind. You are roughly where the outside view would have put you, had you consulted it.

Forecast from the record, not the plan

The remedy is to build your reference class out of your own history. Pull up your log and find out what you actually did over the last three or six months: how many sessions you planned versus how many you completed, and how far a lift really moved from start to finish, including the weeks that didn't count. That number is your rate. It already has the sick days and the travel and the flat fortnights baked in, which is precisely what makes it trustworthy. Divide the distance to your goal by that rate and you have a date that might actually hold.

Then stop making the date the goal. Make the rate the goal. Five pounds a month on the squat, on average, over the next six months, is something you can check, adjust, and keep believing in. It survives a bad week because bad weeks are already inside the average. And it quietly changes what a stall means: not a verdict on you, just a data point that the rate needs updating.

Your next moves

  • Audit your attendance before you plan anything. Count the sessions you intended to do over the last twelve weeks and the sessions you actually did. Use that real ratio, not the intended one, in every forecast from now on.
  • Compute your true rate for one lift. Take the top working weight on your main lift from three months ago and today. Subtract, then divide by three. That monthly number, not the five-pounds-a-session fantasy, is what you can plan on.
  • Rewrite one goal as a rate with a window. Replace 315 by June with something like add about X per month on average through the end of the year, using the rate you just calculated. Put it exactly where the old goal lived.
  • Add a disruption buffer on purpose. Look back over the last year and count the weeks you lost to illness, travel or injury. Assume the same number for the coming year and add them to any timeline before you say it out loud.
  • Schedule a monthly outside-view check. Once a month, compare your actual rate to the planned one. If it runs lower for two months in a row, change the plan or the date, not your opinion of yourself.

Every one of those steps depends on having a record honest enough to argue with, which is the real job of a lifting log beyond remembering what you did last Tuesday. Rep was built to make that record effortless: tap in a set in seconds, and your history becomes the outside view on its own, with every lift's trajectory laid out so the next date you set comes from what you have actually done rather than what you hoped. Pay once, keep it forever, at https://rep.lumenlabs.works.