# What Compound decides for you

> Compound decides your next training day, exercises, sets, reps, rest, and weight automatically, leaving about a dozen settings for you to tune.

**Source:** https://trycompound.cc/blog/what-compound-decides/  
**Author:** Scott Shumaker, maker of Compound  
**Published:** 2026-08-16  
**Topic:** The app (the app, simplicity, programming, privacy)

## In short

Compound decides the next training day in a fixed pull, push, legs, auxiliary rotation, the exercises that fill it, the sets and rep range, the working weight carried from your last session, the rest between sets, and what you warm up and stretch with. About a dozen settings, from rest length to per-slot exercise choice to a per-exercise weight increment, are still yours to change. Everything else, including which day comes next, is fixed.

- Compound decides the day, the exercises, the sets and reps, and the working weight. About a dozen settings are still yours to change.
- A 2010 meta-analysis of 50 studies found the average effect of a bigger assortment on choice satisfaction was close to zero.
- The most famous decision-fatigue finding, a 2011 study of parole judges, has a documented alternative explanation, and the broader theory behind it failed a 2016 multi-lab replication.
- A review of 17 medication-simplification trials found 14 of them improved adherence just by cutting the decisions a regimen asked for each day.
- Training history lives in one file on your phone. There is no server, no login, and no cross-device sync beyond a paired Apple Watch.

---

Compound decides which day you train next, which exercises fill it, how many sets,
what rep range, and what weight goes on the bar. It decides how long you rest, how
you warm up, and what you stretch when you're done. It makes these calls every
session, automatically, because there is exactly one program and it already made
them. About a dozen settings are still yours. Everything else has already been
decided.

## What does Compound decide for you?

Eight things, every session, without a prompt: the day, the exercises, the sets, the
reps, the weight, the rest, the warm-up, and the cooldown.

| Decision | What Compound does | What you can change |
|---|---|---|
| Which day comes next | A fixed pull, push, legs, auxiliary rotation. Push always separates the two heavy days | Nothing. The order is permanent |
| Which exercises | One movement per slot. A handful of slots, lifts and mobility both, carry a second option | Your pick, per slot, from the built-in alternatives |
| Sets and rep range | Set by the plan: 3 sets of 3 to 5 for the deadlift, 3 sets of 12 to 15 for lateral raises, and so on for every exercise | Nothing |
| Working weight | Carried forward from your last session, raised only once you've hit every set at the target | A manual stepper override, and how fast automatic raises come (1 to 5 sessions) |
| Rest between sets | A tier per exercise, 180, 120, 90, or 30 seconds | Timers on or off, and LONG, NORMAL, or SHORT scaling |
| Warm-up | Cardio, the day's dynamic movements, and ramp-up sets at 40/60/80 percent before the first lift | On or off |
| Cooldown | About 10 minutes of mobility matched to the day you just trained | On or off |
| Weight increment | 5 lb on barbell compounds, 2.5 lb on isolation work | A per-exercise override: 2.5, 5, or 10 lb, or the matching kilogram jump |

You didn't choose PULL today, choose 4 sets on the barbell row, or choose to add 5 lb
because you hit every rep last time. Compound made those calls before you opened the
app. Double progression is the rule behind how the weight moves, and the
pull-push-legs-auxiliary order is the logic behind which day comes next; both get the
full treatment in their own posts.

## What does Compound deliberately not have?

Compound ships without an exercise library to browse, a program builder, a social
feed, a leaderboard, a streak badge, an AI coach, or a workout of the day. None of
these is a feature we haven't gotten to yet. Each is a decision we made not to hand
back.

A library exists so you can pick an exercise out of it, which is a decision.
Compound picked already: deadlift for the posterior chain, standing overhead press
over seated, hip thrust alongside the squat and the RDL rather than instead of them.
Four lift slots in the whole plan offer a second exercise, with four more in the
mobility blocks, and every option carries its own progression attached, so the choice
is equipment, not programming.

A program builder hands you the hardest decision in strength training, which
movements, in what order, at what volume, and calls the hand-off customization.
Compound made that decision once. You run it.

A social feed and a leaderboard turn a set into a comparison with strangers, which is
a different decision than the one training actually requires: did you do the set. A
streak badge rewards showing up regardless of what happened once you got there, and
punishes the session you skipped for a good reason exactly as hard as the one you
skipped for no reason at all.

An AI coach reopens the exercise, volume, and rest decisions a program already has to
make, answered a little faster than you'd manage alone. It's still a decision, made
again every time you ask. Compound would rather make it once.

A workout of the day changes daily, decided by someone else, for everyone training
that day at once. It optimizes for variety. Compound optimizes for the same lift
getting one rep heavier than last time, which needs the opposite: the same exercise,
again, until it's earned a change.

No account exists because an account exists to hold your data somewhere other than
your phone, in case you need it on a second device or the company needs it for
something else. Compound doesn't need it for anything, so there's nowhere for it to
live but your phone.

## Does having more choices actually make training harder?

Probably not by itself, and the research is more mixed than the popular version of
this argument suggests. A 2010 meta-analysis in the Journal of Consumer Research
pooled 63 experimental comparisons from 50 studies, more than 5,000 participants
combined, testing whether a bigger assortment of options reduces satisfaction or the
motivation to choose at all (Scheibehenne, Greifeneder & Todd, 2010). The average
effect across every study was close to zero. Choice overload shows up reliably in
some conditions and vanishes in others, and the analysis could not identify a single
precondition that reliably predicted which.

So the claim that a longer exercise list makes you a worse lifter doesn't hold up as
stated. The real moderator the same literature keeps flagging isn't the raw count of
options, it's how hard the options are to tell apart. Ten near-identical
cable row variations is a harder decision than ten obviously different exercises,
because the work is figuring out which difference matters to you, not counting the
list. A typical fitness app's exercise library looks a lot more like the first kind.

## Does deciding less actually help you keep training?

Yes, though not for the reason usually given first. The popular explanation is
decision fatigue: a shared pool of willpower that drains a little every time you
choose something, until there's none left to choose well. The single most famous
piece of evidence for it doesn't hold up. A 2011 study reported that Israeli parole
judges granted parole around 65 percent of the time early in a session and near 0
percent right before a food break, and attributed the pattern to that draining pool
(Danziger, Levav & Avnaim-Pesso, 2011). A later analysis showed the pattern is fully
explained by something duller: easy, favorable cases get decided first, and harder
ones pile up before a break, a scheduling artifact rather than mental exhaustion
(Glöckner, 2016). The broader theory behind it, ego depletion, the idea that
self-control draws on one shared, exhaustible resource, failed to replicate in a 2016
multi-lab study built specifically to test it under registered, adversarial
conditions (Hagger et al., 2016).

What actually holds up is plainer. Regimens that ask for fewer ongoing decisions get
followed for longer, for reasons that have nothing to do with a willpower tank
running dry. A review of 17 medication-simplification trials, each one cutting the
number of pills or the times a day a patient had to decide to take one, found 14 of
them improved adherence (Elnaem et al., 2020). In exercise, a systematic review of
adherence among older adults found completion rates of 65 to 86 percent across
studies, consistently higher in supervised programs, where someone else had already
made the session's decisions, than in self-directed ones (Picorelli et al., 2014).

That's the defensible version of the argument. Not that a willpower tank runs empty
mid-workout. That a program asking you to decide less, every session, tends to get
finished more often.

## What can you still change?

Almost everything that isn't the program itself. Rest timers on or off, and their
length: LONG, NORMAL (the default), or SHORT. How many successful sessions it takes
before your weight goes up, from 1 to 5. Sounds, haptics, and rest alerts, each its
own switch. Warm-up on or off. Mobility on or off. Exercise figures on or off.
Appearance: system, light, or dark. Your weight unit, pounds or kilograms. Which
exercise fills a two-option slot: barbell row or cable row, barbell curl or dumbbell
curl, Copenhagen plank or adductor machine, standing calf raise or seated, and a few
mobility alternatives besides. And, per exercise, the weight jump itself, 2.5 lb, 5
lb, or 10 lb, if the plan's default doesn't match the plates you own.

None of that is a decoy. It's the equipment layer: how loud, how long, how fast,
which of two nearly identical movements, how heavy a jump. It doesn't touch the
program layer: which day, which exercises, how many sets, what rep range. Tune any
of it, or turn it off. The defaults are the program.

## What do you lose with no account and no cloud?

Cross-device sync beyond a paired Apple Watch, and a web dashboard. Training history
in Compound is a file in Application Support, on your phone alone. There's no server
and no login, so there's nothing to breach and nothing to hand to an advertiser, and
it works in a basement gym with no signal because it was never talking to anything in
the first place.

That's a real trade, not a hedge. If your phone is the only place your training
history lives, that actually is the only place it lives: no laptop dashboard, no
second phone with the same numbers. Anyone who wants their lifting log on three
devices and a browser should weigh that honestly against what going without an
account buys back.

## What does "the watch decides nothing" mean?

It means the phone is the only source of truth, and the watch only ever shows what
the phone just told it. Every time something changes during a session, a set logged,
a rest timer starting, a weight adjusted, the phone sends the current state across.
The watch draws that and sends taps back. It never runs its own copy of the logic,
and it never decides anything the phone hasn't already decided. One brain, one
program, two screens.

## Who is this the wrong app for?

Several kinds of lifter, and it's worth naming them plainly rather than pretending a
fixed program suits everyone. An athlete training for a sport with its own specific
demands needs a program built around that sport, not four general strength days.
Someone rehabbing an injury under a physio's program needs to follow that program
exactly, not one that has never seen their chart. Someone who genuinely enjoys
programming their own training, picking exercises, setting volume, adjusting week to
week, is better served by a tool built for that than by one that took the decision
away. And someone who needs a specific exercise the plan doesn't carry, a specialty
lift, a machine their gym doesn't stock an alternative for, will hit a wall Compound
was never built to move.

A fixed program is a constraint. Compound bets that most lifters would rather have
the deciding done for them than keep making it every session, and that bet is wrong
for the sport-specific athlete, the lifter in rehab, the hobbyist programmer, and
anyone missing an exercise the plan doesn't carry.

## Common questions

### Can I choose my own exercises in Compound?

Only within the plan's built-in slots. A handful of exercises carry a second option, such as barbell row versus cable row or barbell curl versus dumbbell curl, and you pick between them. There is no way to swap in an exercise outside the plan or build a routine from scratch.

### What settings does Compound actually let me change?

Rest timers on or off, with LONG, NORMAL, or SHORT length. Progression speed from 1 to 5 successful sessions per increase. Sounds, haptics, and rest alerts, each its own switch. Warm-up on or off. Mobility on or off. Exercise figures on or off. Appearance. Your weight unit. Your pick within each two-option slot. And a per-exercise weight increment override.

### Does Compound sync my workouts to other devices?

Only to a paired Apple Watch, which mirrors the phone live during a session. There is no account and no cloud, so there is no web dashboard and no automatic sync to a second phone or a tablet.

### Is there real research showing fewer choices helps people keep training?

The best-supported evidence is not about willpower running out. It is that regimens with fewer ongoing decisions get followed for longer, a pattern seen in exercise adherence research and, more strongly, in studies of simplifying medication regimens.

### Who shouldn't use a fixed program like Compound?

Anyone training for a sport with its own specific demands, anyone following a physio's rehab program, anyone who enjoys designing their own training, and anyone who needs a specific exercise the plan does not include.


## Sources

- [Scheibehenne, Greifeneder & Todd (2010): Can There Ever Be Too Many Options? A Meta-Analytic Review of Choice Overload](https://scheibehenne.com/ScheibehenneGreifenederTodd2010.pdf)
- [Danziger, Levav & Avnaim-Pesso (2011): Extraneous Factors in Judicial Decisions](https://pubmed.ncbi.nlm.nih.gov/21482790/)
- [Glöckner (2016): The Irrational Hungry Judge Effect Revisited](https://www.cambridge.org/core/journals/judgment-and-decision-making/article/irrational-hungry-judge-effect-revisited-simulations-reveal-that-the-magnitude-of-the-effect-is-overestimated/61CE825D4DC137675BB9CAD04571AE58)
- [Hagger et al. (2016): A Multilab Preregistered Replication of the Ego-Depletion Effect](https://pubmed.ncbi.nlm.nih.gov/27474142/)
- [Elnaem et al. (2020): Impact of Medication Regimen Simplification on Medication Adherence and Clinical Outcomes](https://pubmed.ncbi.nlm.nih.gov/33173282/)
- [Picorelli et al. (2014): Adherence to Exercise Programs for Older People Is Influenced by Program Characteristics and Personal Factors](https://pubmed.ncbi.nlm.nih.gov/25092418/)

---

*Scott Shumaker, maker of Compound. Builds Compound, an iPhone and Apple Watch app that runs one lifting program and does the bookkeeping. Trains on the pull, push, legs and auxiliary split, and wrote the progression engine described here.*

*Compound: A workout app with exactly one workout in it. Four days, automatic double progression, warm-up and mobility built in. No account, no cloud. https://trycompound.cc*