Cat Track

What do my cats actually do all day?

Toast and Peach are very good at looking busy doing nothing. I made them a pair of activity-tracking collars to find out what happens between naps.

Follow the data
Toast and Peach stretched out on their backs, sharing a couch cushion
The subjects, in their natural state.
The storyThe activityThe napsWhat’s nextThe method

26%

More average movement from Toast

70%

Of each cat’s daytime rest overlaps

~6h

Of quiet rest in each eight-hour night

01 / The story

I’ve had cats my whole life.

Michael, Michael Todd, Micky, Muffy, Church, Gypsy, Slash, Saber. And now, Toast and Peach.

Toast also answers to Toaster, Toast Malone, Cinnamon Toast Crunch, and Toastie the Mostie. Or, more accurately, is occasionally addressed by those names.

Toast and Peach sleep a lot. They’re also probably the fattest cats I’ve had. So I got curious: just how active are they, really?

I made some collars using the Axivity AX3, a tiny accelerometer that can log movement for a couple of weeks before I download the data. Then I let the cats get on with being cats.

An Axivity AX3 accelerometer held between two fingers, showing its small size
The setup

Two collar-mounted AX3s. Three axes of movement, sampled 25 times a second. About a week later: 30.2 million readings.

02 / The activity

Toast moves more. But the hour matters.

Across six complete days, Toast recorded 26% more average collar movement. The difference is 32% during the day and just 3% at night. Their nighttime averages are almost the same.

A day in the life

Average of August 31–September 5, 2026

PeachToastNight · 10 p.m.–6 a.m.Meals · usual routine

Bold lines: trailing 5-minute average. Thin, 50% transparent lines: selected detail. The running average needs five complete minutes and stays blank across recording gaps.

Zoom, then scroll sideways. Scale adjusts to day and detail.
Move across the chart or use the time slider.

Movement is average acceleration variability, in mg. It includes head movements and grooming as well as getting up and going somewhere.

Peach is busier before dawn.

From 4–7 a.m., Peach records about 30% more movement than Toast. The pattern holds on five of the six days.

Toast keeps going later.

From 10 p.m. to midnight, Toast records 55% more movement. That also holds on five of six days. By 7–9 a.m., they’re nearly tied.

Time adds up

Toast spends more time moving.

Toast accumulates about 84 extra minutes of movement each day. During those active intervals, average movement intensity is only about 7% higher.

Peach
6.5 h
Toast
7.9 h

Five-second intervals above the same 20 mg cutoff, added together. The full bar represents 24 hours. These are accumulated movements, not hours of continuous exercise.

A fluffy cat sitting in a small wooden toy shopping cart
Why walk when there’s a perfectly good cart?

A surprisingly rigorous investigation into doing very little.

The collars stayed on continuously. The cats kept their usual routines.

A cream-colored cat reclining on a chair beneath a row of coats
Reserved seating.
03 / The naps

The night shift is mostly sleep.

Both spend about three-quarters of the night in sustained quiet rest, compared with about two-fifths of the day. Daytime is twice as long, so it still adds up to slightly more resting hours.

Here, “sleep-like rest” means sustained stillness. Quiet wakefulness can count too.

Nighttime

10 p.m.–6 a.m. · 8 hours

Peach
5.9 h74% of the night
Toast
6.1 h77% of the night

Daytime

6 a.m.–10 p.m. · 16 hours

Peach
6.5 h40% of the day
Toast
6.6 h41% of the day
Per 24 hours

Peach 12.4 hToast 12.7 h

Estimates, not a sleep competition.
70%

They even rest on the same schedule.

About 70% of each cat’s daytime rest happens while the other is resting too: an average of 4.6 shared hours a day. Most of that overlap falls between 11 a.m. and 6 p.m.

Six days of shared downtime

Each row spans midnight to midnight. Toast is on top, Peach below; color marks each cat’s detected rest.

Toast · topPeach · bottomNo rest detectedMissing dataHuman sleep time · 10 p.m.–6 a.m.Meals · usual routine
Darker = longer continuous rest; full color at 10 minutes. Short blocks stay at least 45% opaque.

Loading the rest intervals…

Choose a day below to inspect its five-second intervals.

Totals at right show shared rest across all 24 hours; the 70% figure above refers to daytime only.

There’s solo downtime, too.

Each cat has about two additional hours of daytime rest while the other isn’t classified as resting. Shared timing doesn’t mean they’re sleeping in the same place.

More than a similar daily routine.

Pairing different dates at the same clock times gives 3.7 shared hours, compared with 4.6 on matching dates. That’s about 54 extra minutes of overlap on the actual day.

04 / What’s next

A little less dinner. Maybe a laser robot.

At a recent vet check-in, we determined that Peach needs to lose about three pounds and Toast about one. The plan is mostly measured food portions, with gradual loss of about 1% of body weight per week.

I’ve also considered an automated cat toy—something like a motion-activated turret arm with a laser pointer attached. I’m skeptical they’d engage. But I am very tempted to find out.

For now, this week gives us a starting point. The next recording will tell us what changes.

05 / The method

A small experiment. A useful baseline.

The files cover August 30–September 6, 2026. For a fair comparison, I used the six complete calendar days both collars recorded: August 31–September 5.

2AX3 collars
25 HzConfigured sampling
6Matched full days
99.985%Paired time coverage

Measuring movement

I summarized acceleration changes in five-second intervals, using all three axes. This reduces the effect of static collar orientation. Higher values mean more variable collar movement; they aren’t steps, calories, or distance.

Toast’s recording has a 77-second interruption. The 80 seconds of five-second bins overlapping it were excluded from both cats. The device clocks are treated as local Pacific time.

Estimating rest

An interval qualifies as quiet when movement is below a shared cutoff and the collar’s orientation stays steady. Rest periods need to last at least five minutes; brief interruptions can connect a period but don’t add to its rest time.

Night is 10 p.m.–6 a.m. Day is 6 a.m.–10 p.m. Comparing percentages matters because those windows are different lengths.

How certain are the sleep estimates?

Not certain enough to declare a winner. Across plausible stillness cutoffs and five- or ten-minute minimum rest periods, Peach’s estimated rest ranges from 12.0–13.3 hours per day and Toast’s from 12.2–13.0. The ranking changes with the cutoff.

Peach 12.37 hToast 12.69 h

Totals use a five-minute minimum period. These are method sensitivity ranges, not confidence intervals or bounds on actual sleep.

Quiet wakefulness can look like sleep, and movement during real sleep can break a rest period. We don’t have labeled video for the recording. This is a description of sleep-like rest, not a validated feline sleep measurement.

Data, definitions, and sources

Movement variability is 1,000 × √(var(x) + var(y) + var(z)), calculated within each five-second interval from acceleration in g. The shared rest cutoff is 20 mg, with less than 5° of orientation change and at least 90% quiet intervals in a qualifying period. Missing data never connects rest periods.

Their morning and evening activity peaks resemble patterns reported in other housed cats. But age and outdoor access affect activity, and a fair population comparison needs the same measurement method. A week can reveal routines, not permanent personality traits. These recordings can’t identify specific behaviors, diagnose health, measure weight loss, or establish why the cats’ schedules align.