Following widespread complaints that the remaining battery life estimate was wildly inaccurate on the new MacBook Pro models, Apple acted swiftly to resolve the issue. Unfortunately, that action was not to fix the estimate, but rather to remove it altogether. As John Gruber put it, ‘this is like being late for work and fixing it by breaking your watch.’
I do appreciate the challenge Apple faces with providing accurate battery-life estimates with the 2016 MacBook Pro. I’ve heard from multiple sources that the Skylake CPUs used in the new machines are extremely responsive in the way they ramp Turbo Boost up and down to provide the optimum balance of power and energy-efficiency for each given task.
That means that the amount of battery power being consumed can vary dramatically within the course of just a few seconds. Given that macOS seemingly estimates battery life on a second-by-second basis, the combination of the two things is clearly going to result in the kind of wild fluctuations and inaccurate estimates we’ve seen.
But it’s surely not beyond the wit of Apple’s software engineers to change the way the estimate is calculated … ?
There are those pointing out that, in the first few days of usage, a new Mac is busy doing things like Spotlight indexing. This may mean that both estimates and actual battery life are atypical. This is fair comment, but it’s not like this problem only affects new machines. As my colleague Benjamin Mayo observed, the estimates have always been inaccurate, no matter which Mac you’re using.
One solution would seem to be to take a more intelligent approach to calculating the estimate. Instead of simply dividing the remaining battery power by the amount being used right now, Apple could use a moving average over say the past hour with some historical learning thrown in.
That would not, of course, be a perfect solution. You would, for example, get no help at all during the first hour of usage. But then, Apple’s present solution is to give us no estimate at all period so anything is really better than that.
It’s also true that a moving average across an hour could still be inaccurate. For example, if you spent the first hour doing something undemanding like web-browsing and then started something processor-intensive like video editing, the estimate could still be wildly out. But for most people most of the time, it would at least be more accurate than the current approach.
Marco Arment makes one argument against my suggestion:
Having used Apple laptops for over a decade, I’ve always found the time-remaining estimate to also be a good indicator of how much power I’m burning with my current activities so I can “budget” my battery usage when I’m going to need it.
In other words, the power usage right now approach might be useless for estimating remaining life, but at least if you see it drop dramatically, you know you’re using a lot of power.
But that would be simple to address: simply include an icon next to the battery indicator that appears if you are doing something power-intensive. You could then click on it as now to see which apps are using significant power.
9to5Mac reader Henk van Ess made two further suggestions in a recent email to Phil Schiller. The first is to make battery estimates more personalized.
Monitor workflow of an user say for 30 days and incorporate the daily user routine in some sort of algorithm that influences your battery life prediction.
So, sure, you might be using Photoshop right now, but macOS would know from your typical usage patterns that you only do so for an average of half an hour a day, while the rest of your usage is writing and web-browsing. That knowledge would be factored in to the estimate given to you.
Second, use findings from academic research to develop a much more sophisticated model.
Based on the relationship between the battery capacity, Arrhenius formula, temperature accelerated stress and charge-discharge current accelerated stress, a fitting formula is obtained to predict the battery capacity fading rate and battery charge-discharge cycle numbers. It is shown in the lithium-ion cell tests that the proposed formula can accurately reflect the tendency of battery capacity fading, and the estimation error is less than 5%.
Both strike me as excellent ideas.
Of course, battery life estimates are one thing, and actual battery life is another. So far, the complaints span both issues, and you can count me among those unhappy with the actual life I’m seeing so far. I’ve repeated the test several times since, and never seen more than seven hours of even very undemanding use.
It seems to me that – by claiming 10 hours and dropping MagSafe – Apple’s vision is to turn MacBooks into devices we use like an iPhone or iPad: charge them up overnight, then use them on battery power. For most of us, however, we’re nowhere remotely close to that vision.
Perhaps there are software fixes there too that would improve the power-efficiency of the machines. I certainly hope so. But in the meantime, more accurate estimates would at least mean we know what we’re working with.
Subscribe to our YouTube channel for more videos
How are you finding both estimates of actual battery life? Please take part in our poll, and let us know whether you’d like to see Apple implement one or more of the above ideas.