Battery health vs cycle count — which one actually matters?

Updated 18 August 2026 · every figure below was read on 18 August 2026 from one Apple silicon MacBook Pro (Mac15,10, M3 Max) — from the battery controller's own registry entry and from that Mac's recorded history

Short version: neither, because they are answers to two different questions. Cycle count says how much of the pack's rated life has been spent. Health says how much charge the pack holds today. Asking which one matters more is like asking whether a car's odometer matters more than the size of its fuel tank — one tells you what's been used up, the other tells you what you have.

But the question is usually standing in for a sharper one: is my cycle count too high? That has an answer, and it isn't the answer the search results give you.

The two numbers, side by side

The second column has a wrinkle worth knowing before you compare any two apps: the controller publishes two capacity figures, and software picks one. Why your Mac reports two different health numbers takes that apart; this page assumes it and moves on.

"80% after 1000 cycles" is a design target, not a threshold

This is the one line everybody on the subject repeats, and almost every version of it is worded wrong. Here it is in the manufacturer's own words, from Apple's battery page — a MacBook battery is "designed to retain up to 80% of its original capacity at 1000 complete charge cycles" (apple.com/batteries/service-and-recycling, MacBook Owners section, read 18 August 2026). Read it for what it is: a design commitment about how the pack is built. It is not a failure line, it is not a service recommendation, and — the part that matters here — it is not a forecast for your particular battery.

Two things follow from that, and the second one is uncomfortable.

First: the thousand is not a constant. The battery states its own rated cycle count, and the Mac can read it. On the machine used for this page the controller publishes DesignCycleCount9C, and for this pack it says 1000. That number came from the battery, not from an article about batteries. We printed "of about 1000" as a hardcoded constant ourselves until we started reading the key — and because this pack publishes exactly 1000, the swap changed nothing visible on this machine, which is why a test feeds a different rated count rather than leaving it to the eye. On a pack that publishes no key at all, WattMate drops the section entirely rather than print a denominator it was never told.

Second: on this pack, capacity is already down near the figure the target names for the end. Read on 18 August 2026 — the capacity and the cycle count from that morning's recorded measurement at 10:28, the design figures and the operating time from the controller's registry entry the same day:

The clock time on the capacity row is not decoration. Capacity is an estimate, it moves during a day, and the controller publishes more than one version of it — which is what the page on the two health numbers is for. Quote a capacity figure without saying when it was taken and you have quoted a moving target.

538 out of 1000 is 53.8% of the rated cycles spent — a bit past halfway. And capacity is already at 82.22%, close to the figure the design target names for the end of that thousand. That is a comparison of two states, not a position on a schedule: we have no curve running from the first cycle to the thousandth, so "ahead" and "behind" are words this page has nothing to attach to.

Read that carefully, because it is easy to over-read. It does not predict where this pack lands at a thousand cycles — we have no curve, and the next section explains why we refuse to draw one. It does not say MacBook batteries wear faster than Apple claims: this is one pack, one reading, on one machine, and one of anything is an anecdote with a decimal point in it. What it does establish is enough on its own: the norm is a promise about the design, not a prediction about an individual battery. Anyone who tells you where your capacity will be at some future cycle count is reading a promise as a forecast.

So how many cycles is too many?

The count on its own has no such line, and we are not going to invent one for you. What the count is good for is being the denominator: 538 of 1000 says a bit over half of this pack's rated life is gone. That is a fact about the battery's age, not about your afternoon.

What you'd act on is capacity — and even there, 80% is a design figure rather than a service threshold. A pack below it works; it holds less. Whether "less" is a problem depends entirely on how much your Mac spends per hour, which is a separate measurement with nothing to do with either of these numbers.

Where to look, exactly, with nothing installed: hold Option and click the Apple menu → System Information → Hardware → Power. Under "Health Information" macOS prints Cycle Count, Condition and Maximum Capacity — three numbers on one screen.

Two points are a line segment, not a curve

The tempting next move is to write both numbers down, come back later, and work out how fast you're wearing the battery out. Here is what that actually looks like with real readings from the machine above — and why we will not print a rate from it.

Before reading a direction into those two rows: this series is kept under a rule that a recorded health figure does not climb back up, as long as the cycle count says the pack was not replaced. So the direction from the first row to the second is set by the rule, not observed. What the two rows observe is the size of the gap.

17.5 days apart: 29 cycles added, 1.79 percentage points lost. And that is the whole of it. Two points define a line, and a line through two points fits them perfectly every time — which is exactly why the fit proves nothing. Worse: capacity is an estimate the controller recomputes, and on this same Mac that estimate once moved the other way — up 1.78 points in two days, which is the measurement that made us build the rule above. A gap of 1.79 points is inside noise we have already seen on this machine, so how much of it is aging, we do not know. We do not have a wear curve for this machine. We won't draw one through two dots, and any page that hands you "your battery reaches 80% in N months" is drawing one through about that many.

What would change that is months of readings, from more than one pack. We have not found anywhere macOS keeps past capacity values — it prints today's figure, and we could not find yesterday's — which is why the app records its own measurement once a month, and only when you open the window, and why the direction arrives later rather than on day one. Until it does, this page has a segment and says so.

What neither number will tell you

Both are properties of the pack. Neither says anything about the other half of the arithmetic — what is being pulled out of it. What separates a good day from a bad one is how much the machine is spending per hour, and that is the one quantity neither of these two numbers reports. If that's the shape of your problem, health is fine but the Mac only lasts four hours is the page for it.

This is where a measuring tool earns its place, and it isn't by printing the health percentage prettier. WattMate shows the cycles against the count this pack publishes for itself, and the capacity as the raw reading held to the lowest value honestly recorded: once a lower number has been recorded, and the cycle count says the pack was not replaced, the lower one stays. A health number that walks upward is noise, not a battery growing younger. Then it measures the part the registry has no key for: watts, per app, on your machine under your load, and what comes back when you close something. Wear is what the pack lost. Watts are what you are spending the rest of it on, and only one of the two is still under your control today.

Questions people ask

How many cycles is too many for a MacBook?

There is no such count, which is why nobody credible will name one. The pack measured here has spent a bit over half of what it is rated for, and that is all the count says on its own. That Mac gives three separate readings — 538 cycles and 82.22% of the original capacity from the 10:28 measurement, 24,681 hours of total operating time from the controller's registry entry — and they are three readings, not a chain: none of them is computed from another, and we have not measured how any one of them converts into the others. Another pack at the same count could be somewhere else entirely. The count is the odometer, not the diagnosis.

Should I replace the battery when health hits 80%?

80% is the capacity the pack is designed to still hold after its rated cycles — a target the manufacturer builds to, not a point at which the battery stops working. Below it the battery holds less; whether that ends your working day depends on what your Mac draws, not on the percentage.

My cycle count is low but health dropped fast. Is that normal?

We can't tell you from a number, and honestly neither can anyone else holding one reading. What we can say is that the design norm is not a per-battery promise: the pack on this page has spent 53.8% of its rated cycles and is at 82.22%. Shutdowns well before empty, or a Condition other than Normal in System Information, is what turns this from a curiosity into a service appointment.

Which number does WattMate show?

Both, with their sources named. Cycles are shown against the rated count the pack publishes for itself rather than a constant, so the denominator is true on your machine or absent. Capacity is the raw controller reading, held to the lowest value honestly recorded.

Does a partial top-up add a cycle?

Not one each time. A cycle is a full charge in total, not one trip to the outlet: two half charges also make one. That is why a Mac that lives on a desk can collect cycles slowly despite being plugged and unplugged constantly.

What these numbers do not prove

• Nothing about MacBooks in general. One pack, one machine, one set of readings taken on 18 August 2026. The direction of the finding is a property of this battery, and yours can sit on either side of it.
• No wear rate. Two health readings 17.5 days apart are a segment. A rate needs a curve, a curve needs months, and we don't have them yet.
• No link between cycles and hours. We have not measured how a cycle count translates into runtime on a given machine, so we are not going to put a number on it.
• No replacement advice. That decision has your budget and your workload in it. We can supply one of the inputs, measured, and say plainly where it stops being informative.

Related: why your Mac reports two different health numbers · health is fine but the Mac only lasts four hours · how we count watts, minutes and battery health