Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
You have to know how they count cycles.
What do you mean with that? A battery cycle is equal to 100% of discharge, whether this is 5 times 20% or 100% directly, no matter recharging the battery in the mean time. So 45 battery cycles would mean a sum of discharges equal to 45 times 100% battery.
 
I'll say this. I was getting ~5.5 hours on a full charge (on a new 15" MacBook pro touchbar), the update for Sierra came out and I was getting about the same. So I started looking at background applications and processes using Activity Monitor and I turned off about 4 or 5 Adobe Creative Cloud apps that were running in the background. None of them showed any significant battery drain at all, but that was a lie... shutting them down and voila, I was getting 10-14 hours of battery life on a full charge. After 2 hours it was still at 86% battery (doing Skype for business, Outlook, safari and iMessage at 60% brightness).

I have 2 other colleagues at my work that purchased 13" MacBook pros with touchbar, all had Adobe CC on as well and all had the same 5-6 hours of battery. We shut down the background apps on their MacBooks as well and they also started getting over 10 hours.

So, the update didn't seem to do squat, but the culprit in our case was Adobe CC. Food for thought.
 
If Apple's position is that battery life estimates are so inconsistent as to remove that feature from OS X, how can we rely on Coconut Battery's estimates as evidence of longer battery life?
 
Good news for a change! If you want to verify the battery time remaining you can still use Activity Monitor, even though it isn't as convenient as the old Battery Bar Estimate.
 
Rubbish! Utter rubbish!
Averaging out over a longer period will average out these power spikes and give you the mean power drain over time. There is no reason this won't work accurately. And be useful.

That will just tell you a stupid average. As I said, laptops in the past used to have a fairly consistent power draw regardless of use. This means a low standard deviation. So predicting future power usage was fairly decent. Today, laptops have very good power efficiency features where the power draw is not consistent at all. It goes from using very little power at idle or at performing system-level tasks to using a lot of power when actually interacting with an app. Even something as simple as browsing a website, the CPU goes from idle to full throttle back to idle many times depending on what scripts the site uses and how much dynamic content there is. Sure you can average it over a longer period of time, but it will be useless if the standard deviation of that average is high. In stats terms, the r-coefficient of your prediction will be low.

The prediction could be accurate to the extent your usage from day to day is consistent, but how is that a useful prediction? If your usage is consistent, you don't need software to tell you the battery will last today for about the same amount of time as it did yesterday.

This is why it's so hard to design a proper battery life test, much less design a good way to predict future battery life.
 
  • Like
Reactions: moonjelly
I've noticed that using an external monitor in clamshell mode (just 1920x1200@60Hz, USB-C to DisplayPort) causes the MBP to use 'High Performance' graphics according to Activity Monitor. There are no running processes that require it. This is a bug, surely? It's a lot less pixels than when driving the built in screen.
 
  • Like
Reactions: HurryKayne
Hmm. Still getting ~4 hours of actual use on my 2016 15" - I do have a few things running in the background / menu bar, but nothing on activity monitor seems to jump out. 460 but it switches to integrated fine. My idle draw is 8-11w. :/ Wifi on, bluetooth off, no apps open in foreground, screen dimmed. Any suggestions? It's better than the ~2 with my early 2011 (ssd + platter drives via owc, upgraded RAM)

View attachment 678055


Everything works fine aside from headphones while bootcamped in Windows 10 (and bootcamp assistant didn't copy over files to internal partition, had to do some jiggling with downloading drivers for my machine on top of an install made from a USB drive on someone else's).
Am I reading this wrong? It says you have been using it for 3.5 hours on battery and still have 72% left and have 4:48 hours to go. A total of 8+ hours?

Also, duet and Pathfinder seem to be power hungry and you have a lot of things running. See if you can unclutter some programs that you no longer use.
 
That will just tell you a stupid average. As I said, laptops in the past used to have a fairly consistent power draw regardless of use. This means a low standard deviation. So predicting future power usage was fairly decent. Today, laptops have very good power efficiency features where the power draw is not consistent at all. It goes from using very little power at idle or at performing system-level tasks to using a lot of power when actually interacting with an app. Even something as simple as browsing a website, the CPU goes from idle to full throttle back to idle many times depending on what scripts the site uses and how much dynamic content there is. Sure you can average it over a longer period of time, but it will be useless if the standard deviation of that average is high. In stats terms, the r-coefficient of your prediction will be low.

The prediction could be accurate to the extent your usage from day to day is consistent, but how is that a useful prediction? If your usage is consistent, you don't need software to tell you the battery will last today for about the same amount of time as it did yesterday.

This is why it's so hard to design a proper battery life test, much less design a good way to predict future battery life.
Quite simply: you are wrong. The variation in power consumption is irrelevant if you take and use the averaged power draw over a 10-15 minute window.
 
Quite simply: you are wrong. The variation in power consumption is irrelevant if you take and use the averaged power draw over a 10-15 minute window.

Quite simply: You don't know what you're talking about. Average is not necessarily accurate, or even useful. There are ways, in statistics, of determining the quality of a prediction based on averages. The input data (e.g., variation in power consumption) does matter. In this case, the average is useless.
 
  • Like
Reactions: moonjelly
My battery consumption on sleep used to be horrible. I would get 17% battery usage on sleep overnight, now it only uses 1%. I feel battery life when using the laptop is still horrible and nowhere near my old my MacBook Pro 2013. Apple needs to get their **** together and fix these issues.
 
  • Like
Reactions: HurryKayne
I have updated and seen zero improvement in battery life. Fully charged at 8.55am, dead at 1.30pm - Just using Safari and Excel.
make sure you do a full SMC and PRAM reset, if that does not work, seriously, bring it back, as its broken, I get over 11hours doing ALL kind of things. (except chrome)
[doublepost=1481819905][/doublepost]
Good news for a change! If you want to verify the battery time remaining you can still use Activity Monitor, even though it isn't as convenient as the old Battery Bar Estimate.
or you can download "battery Monitor" for free, which gives you the reading in the icon bar with all kind of info directly from the activity monitor. right now I have the exact same functionality I used to have with time remaining..
 
Ironic that you commented in the same box about someone else not having a sense of humour. Did you THINK about that? LOL. Here's another word for you - hypocrite.

Actually, he wasn't joking and he knows it. For the lack of anything else intelligent to say, he just became a parrot and repeated what I said. It's really sad. ;)
 
  • Like
Reactions: kahkityoong
I can ask you the same thing - how little do you actually know?

The estimate updated every minute. This forum is full of people describing massive swings in the estimate. Indeed, it was one of the main complaints with the new MBPs.

Also, my claim was a paraphrasing of Apple's own support article:
View attachment 678058
https://support.apple.com/en-us/HT204054

Ok, so we're maybe talking about a 1-minute average which is a lot different to an instantaneous estimate. So if you're doing a lot in that minute it'll go down to reflect that and vice versa. People know this, it doesn't mean it's not useful.

I get 3-4 hours of moderate to heavy use and 8-9 hours of light use on my 2015 model with the average falling somewhere in between (6-7 hours of normal use). I've never seen estimates outside of this range. The estimates have always been quite good.

So Apple's justification for removing it is: we can't predict exactly to the nearest minute what battery life will be or predict your future usage so therefore it's not worth providing an estimate at all! I don't buy it.

Apple removed it from the menu bar but kept it in Activity Monitor. So it's good enough for Activity Monitor apparently but not good enough for the menu bar and has to be removed for fear of people not understanding it? Ok I guess? It's not like every other laptop made in the past 20 years hasn't had time remaining estimates. There evidently hasn't been a need to protect the public from such dangerous information until now! ;)
 
Ok, so we're maybe talking about a 1-minute average which is a lot different to an instantaneous estimate. So if you're doing a lot in that minute it'll go down to reflect that and vice versa. People know this, it doesn't mean it's not useful.

I get 3-4 hours of moderate to heavy use and 8-9 hours of light use on my 2015 model with the average falling somewhere in between (6-7 hours of normal use). I've never seen estimates outside of this range. The estimates have always been quite good.

So Apple's justification for removing it is: we can't predict exactly to the nearest minute what battery life will be or predict your future usage so therefore it's not worth providing an estimate at all! I don't buy it.

Apple removed it from the menu bar but kept it in Activity Monitor. So it's good enough for Activity Monitor apparently but not good enough for the menu bar and has to be removed for fear of people not understanding it? Ok I guess? It's not like every other laptop made in the past 20 years hasn't had time remaining estimates. There evidently hasn't been a need to protect the public from such dangerous information until now! ;)

And when they remove it from Activity Monitor you will still be able to get it through a terminal command. I don't understand what the big deal is about.

Clearly a bunch of people complained about a legitimate battery issue, and cited an unreliable battery life estimate in the UI as proof. So a real issue supported by bad evidence. The Apple engineers agreed there was an issue with battery life and found other evidence in the software to prove it, but also learned that far more people were relying on this semi-hidden time estimate that the engineers know to be unreliable. So they further removed it from prominence to de-emphasize it, while at the same time fixing the actual software issue with the battery life. What's so wrong with this?

As I said to someone else above, if your usage is consistent then your battery life today will be about the same as it was yesterday. No time remaining estimate needed. If you usage changes greatly day to day, or even hour to hour, then the time remaining estimate will be useless to you as it won't be reliable. In either case, nobody will miss it.
 
Am I reading this wrong? It says you have been using it for 3.5 hours on battery and still have 72% left and have 4:48 hours to go. A total of 8+ hours?

Also, duet and Pathfinder seem to be power hungry and you have a lot of things running. See if you can unclutter some programs that you no longer use.

You are, that's total time on battery - I was on it for maybe an hour at a cafe. I've dropped Duet from auto-starting on load due to energy use and the fact that it's not something I really use when on battery anyways. Pathfinder is something I'd prefer to keep. I didn't have any non-background process running when testing that load - if I'm expected to close pathfidner, istat, dropbox, alfred, flux, etc every time I unplug because they're halving my battery life I don't see how people are supposed to do work on this for 10 hours. :/

Duet and Pathfinder? Those are known energy burners, and it does show there. What's your power draw without those?

Skimming activity monitor while real applications are open Coda 2 has an energy impact of 12 to 118 with a single gulp task running in a terminal tab (though that spike wouldn't happen often). Sketch uses 28-75 (and pops on my dGPU). Safari is generally on 10-16. Having some ~.5 and one ~3-4 energy impact process (pathfinder) should be within acceptable use.

Included Activity Monitor screenshot below of processes with a low of 7.7w when activity monitor wasn't running.

Screen Shot 2016-12-15 at 9.57.09 AM.png


Killing everything except istat menus (on slow refresh, quit just closes preferences) and backblaze (bluetooth off, wifi on, screen ~75%) dropped me to 7.7w with an estimated little over 4 hours of battery life, using up around 5% of my battery in 15min closing applications, checking activity monitor, and firing up safari occasionally.

OP had > 9 hours of estimated use, and that's with Safari running (and presumably some other things like an email client, etc). I'm guessing there's some hardware flaw with my machine.
 
Last edited:
I've previously speculated that due its downsized battery the 2016 MBP would inherently suffer less battery life than previous MBPs when used in more demanding tasks. Nor see any reason to alter that supposition.

If some customers are reporting improved battery life now, then at least some of them likely due the placebo effect of thinking all better and now having no way of quickly seeing otherwise—thanks to Apple making OSX all the more iOS.

However it has also been pointed out by others that background tasks could be a factor in this. Not just the initial indexing which may have confused some with its system demands, but ongoing. Apparently Sierra is the first version of OSX which really isn't, or rather a hybrid between between that and iOS. This going far beyond just the exclusion of remaining battery time to only percentage as common with iOS. Among other things is now the symbiosis with iCloud, which I believe is the default option unless manually turned off. Apple has alluded to such background operations which a customer may not be aware of, or want—and which would but additional demands on a battery.

If Apple did make any adjustments towards improved battery life in this update, then perhaps in exactly this, to ramp down the demand of background operations which many customers might not be aware of or account for.
 
Quite simply: You don't know what you're talking about. Average is not necessarily accurate, or even useful. There are ways, in statistics, of determining the quality of a prediction based on averages. The input data (e.g., variation in power consumption) does matter. In this case, the average is useless.
The variation doesn't matter at all. All that needs to be measured is the amount of power that is used over a 10 or 15 minute sliding window to remove the impact of any short but high intensity task.

Consider equal time intervals of 1 minute each recording power drops of:
1,1,1,1,1,20,1,1,1,1

At any moment the estimated time left may be somewhat innacurate, but if you average it over the 10 minute window you get an average power drop/usage of 29/10 = 2.9 per minute
 
You are, that's total time on battery - I was on it for maybe an hour at a cafe. I've dropped Duet from auto-starting on load due to energy use and the fact that it's not something I really use when on battery anyways. Pathfinder is something I'd prefer to keep. I didn't have any non-background process running when testing that load - if I'm expected to close pathfidner, istat, dropbox, alfred, flux, etc every time I unplug because they're halving my battery life I don't see how people are supposed to do work on this for 10 hours. :/



Skimming activity monitor while real applications are open Coda 2 has an energy impact of 12 to 118 with a single gulp task running in a terminal tab (though that spike wouldn't happen often). Sketch uses 28-75 (and pops on my dGPU). Safari is generally on 10-16. Having some ~.5 and one ~3-4 energy impact process (pathfinder) should be within acceptable use.

Included Activity Monitor screenshot below of processes with a low of 7.7w when activity monitor wasn't running.

View attachment 678176

Killing everything except istat menus (on slow refresh, quit just closes preferences) and backblaze (bluetooth off, wifi on, screen ~75%) dropped me to 7.7w with an estimated little over 4 hours of battery life, using up around 5% of my battery in 15min closing applications, checking activity monitor, and firing up safari occasionally.

OP had > 9 hours of estimated use, and that's with Safari running (and presumably some other things like an email client, etc). I'm guessing there's some hardware flaw with my machine.
Your last screenshot was different, this one makes more sense.

Can you please post a screenshot with the Average Energy Impact column sorted from highest to lowest? The time remaining will refresh after a while of doing the same tasks and not instantly. Seems like your Safari was drawing an average of 10+
[doublepost=1481827817][/doublepost]
Your last screenshot was different, this one makes more sense.

Can you please post a screenshot with the Average Energy Impact column sorted from highest to lowest? The time remaining will refresh after a while of doing the same tasks and not instantly. Seems like your Safari was drawing an average of 10+
You are, that's total time on battery - I was on it for maybe an hour at a cafe. I've dropped Duet from auto-starting on load due to energy use and the fact that it's not something I really use when on battery anyways. Pathfinder is something I'd prefer to keep. I didn't have any non-background process running when testing that load - if I'm expected to close pathfidner, istat, dropbox, alfred, flux, etc every time I unplug because they're halving my battery life I don't see how people are supposed to do work on this for 10 hours. :/



Skimming activity monitor while real applications are open Coda 2 has an energy impact of 12 to 118 with a single gulp task running in a terminal tab (though that spike wouldn't happen often). Sketch uses 28-75 (and pops on my dGPU). Safari is generally on 10-16. Having some ~.5 and one ~3-4 energy impact process (pathfinder) should be within acceptable use.

Included Activity Monitor screenshot below of processes with a low of 7.7w when activity monitor wasn't running.

View attachment 678176

Killing everything except istat menus (on slow refresh, quit just closes preferences) and backblaze (bluetooth off, wifi on, screen ~75%) dropped me to 7.7w with an estimated little over 4 hours of battery life, using up around 5% of my battery in 15min closing applications, checking activity monitor, and firing up safari occasionally.

OP had > 9 hours of estimated use, and that's with Safari running (and presumably some other things like an email client, etc). I'm guessing there's some hardware flaw with my machine.
Also kill Adobe CC apps, they seem to not appear in the Energy usage but are drawing a lot
 
Your last screenshot was different, this one makes more sense.

Can you please post a screenshot with the Average Energy Impact column sorted from highest to lowest? The time remaining will refresh after a while of doing the same tasks and not instantly. Seems like your Safari was drawing an average of 10+
[doublepost=1481827817][/doublepost]

Also kill Adobe CC apps, they seem to not appear in the Energy usage but are drawing a lot

No CC apps running (the only adobe product I use in LR standalone, though the photography plan for RAW in iOS is tempting). With the 7.7w I had nothing in the foreground (e.g. force quittable) and hardly anything in the background, and that was waiting a minute or two for usage to drop. 18-23w load was with normal background applications sans duet and only safari in the foreground. It had been shifting slightly within a ~20 minute range total during those 15 minutes of use (3:55 to 4:15 total time).

I now have all my standard background stuff running and a decent amount of normal dock applications and I'm drawing 13.8w with a total time of 6:15 hours with my 460 on (because sketch is app napping I assume). Which seems more reasonable. Both iStat and Battery Health seem to poll faster than activity monitor a 5:59 remaining lingered there well after the two apps had updated their expected life. Now up to 14.7w.

Completely baffled as to how I'm getting ETA 50% more battery life with a bunch of stuff open (a single app with a power draw of 10 should be expected if you're actually using it imo). I'll post again later in the day when average energy impact has had a chance to accumulate stats.
Screen Shot 2016-12-15 at 11.12.14 AM.png
 
The variation doesn't matter at all. All that needs to be measured is the amount of power that is used over a 10 or 15 minute sliding window to remove the impact of any short but high intensity task.

Consider equal time intervals of 1 minute each recording power drops of:
1,1,1,1,1,20,1,1,1,1

At any moment the estimated time left may be somewhat innacurate, but if you average it over the 10 minute window you get an average power drop/usage of 29/10 = 2.9 per minute

Ok, let's go with your example sequence. The average over 10 minutes is 2.9 per minute, and the standard deviation is 6. That's a huge standard deviation. This means, we can say that in 68% of circumstances, the 11th minute's usage will be somewhere between -3.1 and 8.9. Obviously it can't be less than zero, but this demonstrates that your example sequence gives a really bad prediction. That's not a very tight prediction.

Let's test the r-squared value of your prediction. First, we have to map the usage. Under your sequence the total used energy by minute is 1, 2, 3, 4, 5, 25, 26, 27, 28, 29. Right? Under your 2.9/minute prediction, the next 10 minutes will be 2.9, 5.8, 8.7, 11.6, 14.5, 17.4, 20.3, 23.2, 26.1, 29. Same total usage at the end of the 10 minutes. The r-squared of this is 0.8. Not terrible, but not great either.

How about this sequence for 10 minutes: 20,1,1,1,1,1,1,1,1,20. Average is 4.8/minute. Standard deviation is 8! R-squared is 0.6 - pretty freakin' bad.

Compare your sequence, and my other sequence, to a tighter 10-minute sequence: 17,17,17,17,17,20,17,17,17,17. Average is 17.3/minute. Standard deviation is 0.9. R-squared is 0.999. Now that is something from which you can make a good prediction.

Variation matters!
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.