Consumer Reports Reverses Course, Recommends MacBook Pro Following New Testing After Apple Bug Fix

[MacsRuleOthersDrool]

I don't know, those battery numbers still seem way off. Especially when you have real use reports from actual users that are getting less it really makes this data to be subjective.

And others, like a friend of mine with a tb 15, that are getting stellar battery life.
[doublepost=1484281270][/doublepost]

Shouldn't they wait to change their recommendation until after the fix is actually released? Average Joe who goes and buys one based on this report will then wonder why it's still an issue for them. No recommendation should be made based on beta software.

No. It is Average Joe WEB DEVELOPER who:

1. Is living in a cave, and hasn't heard of all this.

2. Disables his Safari Browser Cache.

3. Develops with Safari.
[doublepost=1484281749][/doublepost]

CR refrained from its recommendation on grounds of battery life.

As for best laptop.. we both know that's not true.

It's the best Laptop that Apple has ever produced. And the best Laptop that can legally run macOS.

So, both of those things alone make it the best Laptop. Seriously.

But it also happens to be one of the best Skylake-based laptops around, if not THE best.

Highest (by far!) Raw I/O capability, highest ability to intrinsically drive the most and highest-resolution displays, highest resolution display (I think), unique, multifunctional, non-intrusive, graphical input device.
[doublepost=1484282320][/doublepost]

If you remember the initial results, they also included 15+ hours of battery life in some test runs. So what you see here is the removal of the bug that intermittently reduced the battery life significantly from 15+ hours under a light load.

EXACTLY!

The deeper message here is that Intel and Apple have managed to reduce battery consumption under idle or light load very significantly. Essentially, when not needed various parts of the CPU, motherboard and OS can go into something like a sleep mode. The problem with these advances is that the power consumption in idle mode and under full throttle now probably exceeds a more than 10x differential. This means that actual power consumption will vary dramatically from user to user and application to application. And that any bug that disturbs this 'sleep' has very large consequences. And one might ask whether any real system with many applications and with minor problems accumulating will be highly unlikely to allow these advanced power-saving measures to become active.

Again, Exactly!

 

[shareef777]

Shouldn't they wait to change their recommendation until after the fix is actually released? Average Joe who goes and buys one based on this report will then wonder why it's still an issue for them. No recommendation should be made based on beta software.

The check they wrote isn't beta.

 

[Pike R. Alpha]

Your post is a great example about how facts are bended through incompetent journalism and careless retelling by forum posters who didn't bother to read the entire story. Facts: CR was using a non-default, debugging configuration of Safari that can only be activated via a hidden developer menu. Virtually no users use that configuration. Yes, its Apple's bug in the end, but this bug only affects a very particular, low-profile operation mode and is therefore much less tested than the configuration a normal user would use. CR should know that using non-default settings of a browser cannot be representative for default operation. They have enabled that setting to emulate a particular scenario — and that made sense — however, they should have implemented that scenario at their server's end instead. Their mistake, one that I consider to be very crude for such a well-known organisation — is that they used a non-default configuration of the browser while not mentioning this fact in their original review(!!), observed some conflicting and overall clearly weird results (they even say it themselves!), and with all that, still proceeded to publish the article. I come from the scientific community, and thing like these are considered gross negligence and unprofessionalism. If you get non-systematic results in your experiment, which also conflicts with other related experiments, the only conclusion you can make is that your test is obviously not working properly.

Wait no! You are completely wrong!

The way Consumer Reports did their testing actually triggered the bug, and it was a bug that may have been there for many more years without the way CR did their testing (the flawed code was first developed like 14 years ago)

Ergo. Your assumption that CR should have implemented it at the server end, is a bogus solution. What that would have done is hide it, and then Apple would not have known about it, and not fixed it.

 

[macmediausa]

I never understood how anyone could get 10 or even 7 hrs on a charge. I have NEVER gotten more than 5 or so hours. Right now with my MBP mid 2015, I can barely get 3.5 hrs on a charge doing strictly Safari (I do have a few windows open with many tabs). Only 68 battery cycles too. The backlight is set for about 50%
Ditto for every other MBP and air that I own (I have 9 mac laptops and all last for about the same amount of time)

 

[nofear1az]

And others, like a friend of mine with a tb 15, that are getting stellar battery life.
[doublepost=1484281270][/doublepost]

Prove it! (As you said to another poster getting the opposite.)

You are so BIASED in your opinions that you don't want to hear or see anything that does not go along with your beliefs. You have total fanboyism. Apple can never do wrong in your biased views so there's no reason to refute or even to have a real discussion with you. Millennials!

 

[MacsRuleOthersDrool]

15.25 to 17.75 hours? That comports with the experiences of exactly zero real world users that I've talked to or read about.

Probably because they don't just sit there and load web pages, and so have an AGGREGATE higher-current consumption.

I would guess that is why Apple's test is probably designed to be more "real world", and thus gets lower battery-life results than CR's test.

With these highly-power-optimized CPUs, combined with Apple's highly-power-optimized OS, seemingly minor variations in minute-to-minute application load likely can make dramatic differences in the minute-to-minute battery drain.

 

[okieiam]

foolish, will the bug fix improve older hardware too?

 

[lowkey]

Lol, looks like Apple hit them on the head to get this result. They shouldn't recommend a laptop that is aimed at the pro market, yet Apple has prioritised features that belong to the ultra portable market.

Yeah, like only having 4 full bandwidth TB3 ports that can run multiple 5k monitors and have enough spare for multiple raid arrays, i7 chips with the fastest SSDs in town.

Oh, but by professional you obviously mean "someone who can't afford new cable" LOL
[doublepost=1484283021][/doublepost]

I never understood how anyone could get 10 or even 7 hrs on a charge. I have NEVER gotten more than 5 or so hours. Right now with my MBP mid 2015, I can barely get 3.5 hrs on a charge doing strictly Safari (I do have a few windows open with many tabs). Only 68 battery cycles too. The backlight is set for about 50%
Ditto for every other MBP and air that I own (I have 9 mac laptops and all last for about the same amount of time)

Really?
My MacBook gets through the workday with at least 25% left at the end.

 

[MacsRuleOthersDrool]

Curious what the technical explanation is because the variance is huge for something as simple as browser cache? I'm guessing a ram disk cache is utilized to keep SSD and WIFI in a low power state but those hardly draw any power to begin with. Perhaps CPU is also kept in a low power state.

It wasn't the CACHE, it was the BUG that was "ENABLED" by Disabling the CACHE.

According to Apple, that Bug caused Safari to INTERMITTENTLY fall into a LOOP, CONTINUOUSLY RELOADING certain Assets on the webpage. THAT's what burned the battery-charge.
[doublepost=1484283604][/doublepost]

where Apple's published specifications list up to 10 hours use, in test scenarios that were much less stringent.

Really? I would assume that Apple also disables the Browser Cache when looping through its 25 website-test-set.

 

[julesme]

Nobody here likes the fact that Apple actually fixed a bug that was exposed by CR's testing methods? I certainly do.

To me it doesn't matter if you value Consumer Reports, or not, but come on folks. Some of you here are really acting like a five year old child. You may find CR stupid, or whatever, that is your good right, but nobody here should have to read baseless insulting rants from anyone. Now. Grow some balls and dump your silly sh*tload of pulp fiction to /dev/null

No, it's about the unprofessional way that CR is approaching the issue and a general recognition by the public that CR is no longer relevant as it was many years ago. First of all, this is a completely inconsequential bug. Who turns off caching on their browser? 4 or 5, maybe half a dozen, out of a million people? CR is only distracting from real issues, bugs, neglected product lines, etc. and other problems facing Apple, many of which are addressed by MacRumors users on a daily basis right here on these forums. CR also comes across as promoting clickbait and somewhat desperate for attention. Just look at the Youtube video above with the giant text superimposed: "Battery issue FIXED!" It wasn't even a battery issue. It was a Safari issue.

 

[oldmacs]

Yeah, like only having 4 full bandwidth TB3 ports that can run multiple 5k monitors and have enough spare for multiple raid arrays, i7 chips with the fastest SSDs in town.

Oh, but by professional you obviously mean "someone who can't afford new cable" LOL

By professional I mean not shrinking the battery and making other compromises for the sake of making it thinner.

 

[MacsRuleOthersDrool]

This kind of reminds of when you need AA batteries for your flashlight and you decide on said Name Brand or those generic Heavy Duty batteries over there. You always pick up the Name Brand for high drain devices and pick the generic battery for low drain devices.

We all know the MacBook Pro has great -no awesome power saving features which seems to make it last a long time but once the system is tagged it drains much too quickly - see reference above. Just my opinion.

I think that Apple is guessing that most people who are using high-demand applications will do so while connected to AC power.

And in real life, that's a pretty good guess, at least for people not looking to "find fault".

The ONLY thing that is relevant is how long the essentially-equivalent 2015 MBP lasts, doing EXACTLY (not kind-of) the same things, and running the SAME version of MacOS.
[doublepost=1484284655][/doublepost]

By professional I mean not shrinking the battery and making other compromises for the sake of making it thinner.

But it appears that they actually DID shrink the battery and make it thinner AND it did NOT compromise anything, due to the hard work of both Apple and Intel.
[doublepost=1484285051][/doublepost]

COOL ... when CR reached out to Apple, what was Phil the Shill's response "if anyone is having a problem they should contact AppleCare" ... the YOKE is all over Phil and his cronies at Apple.

Take another gulp of that Hatorade. Hope you like the taste of WRONG!

Apple admitted it was a Bug in Safari. They fixed the Bug. CR Retested. Case Closed.

Sorry to disappoint you; but not EVERYTHING is a Conspiracy.
[doublepost=1484285233][/doublepost]

"obscure" bug... I see.

Spoken like someone who has never written a line of production-code in their life.

 

[JamesPDX]

I am not sure what this means? You mean the MacBooks or the nTB MacBook Pros?

nTB=non-Thunderbolt?
[doublepost=1484285443][/doublepost]

It wasn't the CACHE, it was the BUG that was "ENABLED" by Disabling the CACHE.

According to Apple, that Bug caused Safari to INTERMITTENTLY fall into a LOOP, CONTINUOUSLY RELOADING certain Assets on the webpage. THAT's what burned the battery-charge.
[doublepost=1484283604][/doublepost]
Really? I would assume that Apple also disables the Browser Cache when looping through its 25 website-test-set.

Not a fan of cache chewing through the flash/SSD. Firefox and Calomel SSL validation .xpi for me...

 

[MacsRuleOthersDrool]

I've been following threads about battery life, and never seen anyone mention 15-18 battery life. Those figures are ridiculous, something must be wildly unrealistic about those tests.

They ARE wildly unrealistic. That's why Apple's own tests actually show LOWER battery life. But CR's test is obviously still allowing macOS to "nap" some parts of the system, even though Safari is having to re-load and re-render the Pages.

But you will note that the figures that CR is getting are NOT out-of-line with the "higher result" tests from the initial test-set. So, those are not necessarily out of line for THEIR test methods AND Apple's combination of power-optimized CPU and power-optimized OS.
[doublepost=1484285682][/doublepost]

This is a confusing article stating an "average of 18.75 hours" with no clarification of usage.
I'd imaging people editing video and pictures or any other intensive programs will see nothing like 18 hours.

No fooling. That's why it is getting VERY important to know EXACTLY the Test Conditions and Methods.
[doublepost=1484285741][/doublepost]

BRAVO Consumer Reports ... great that you finally exposed Apple's arrogance ... "you're not testing it right" ... ONLY to find out that Apple had another buggy operating system in the wild.

That's just ridiculous.
[doublepost=1484285853][/doublepost]

wow.

Can't believe CR caved and retested.

You can do all the retests in the world and spin your PR till the cows come home, Apple. It's not going to change the fact that the new MacBook 'Pro' is a turd. Polish it all you want. It's still a turd.

If Apple had put 1/2 the effort they did in designing new emoji's into developing these new machines perhaps we'd have a better product.

They did; but you just don't want to see it. Better to just be snarky on some internet forum.

 

[iriejedi]

How many legs were broken for this to happen? This is a really dirty move for both of them.

Broken legs is so Obama.. cut off fingers... get with current times....
[doublepost=1484286051][/doublepost]

They ARE wildly unrealistic. That's why Apple's own tests actually show LOWER battery life. But CR's test is obviously still allowing macOS to "nap" some parts of the system, even though Safari is having to re-load and re-render the Pages.

But you will note that the figures that CR is getting are NOT out-of-line with the "higher result" tests from the initial test-set. So, those are not necessarily out of line for THEIR test methods AND Apple's combination of power-optimized CPU and power-optimized OS.
[doublepost=1484285682][/doublepost]
No fooling. That's why it is getting VERY important to know EXACTLY the Test Conditions and Methods.
[doublepost=1484285741][/doublepost]
That's just ridiculous.
[doublepost=1484285853][/doublepost]
They did; but you just don't want to see it. Better to just be snarky on some internet forum.

the test is rigged!

 

[carrrrrlos]

Looks like Consumer Reports hired the engineers who worked on ‘solving’ the VW diesel emissions issues.

 

[MacsRuleOthersDrool]

May I ask - where do you have the brightness set - thinking that is their issue? I bought 2 touch bar models for my kids (13" added 16 GB). They get about 20% less battery life than my oldest daughters machine (2015 13" rMBP - also with 16GB). My kids use theirs on full bright and usually get around 6 hours for school (includes videos). They got 6 1/2 hours watching Netflix movies (again full brightness). My kids love the new models and have no complaints but I'm wondering if they are doing something wrong to get so much less battery life. Apple told us to expect 20% less battery life on these (compared to my daughter's 2015 model). They also told me that using full brightness uses a LOT more battery. They use Safari. Thanks

You answered your own question.

The backlight in a laptop display is one of the biggest, if not THE biggest, current-hogs. That's why Apple specifies what the Screen Brightness setting is in their battery-life testing (12 clicks from the bottom, or 75%, with auto-brightness OFF). Add to this the fact that the 4k display is having to drive significantly more pixels than the earlier model. This is relevant, because what causes an LCD to consume power is when pixels CHANGE. So, the more pixels, the more pixels that change, and thus a higher-resolution display just naturally consumes more power than a lower-resolution one. That's just the physics of LCDs.

There is an interesting thing about brightness: (Within reason), your eyes automatically-compensate for varying brightness levels. I'd suggest to your daughters TRY to back-down the brightness, and give it a shot. After all, the new MBP displays are actually about 25% BRIGHTER than their predecessors; so they SHOULD be able to back-down about 25% and achieve the same ACTUAL brightness. This will make a BIG difference (probably about 2 hours overall) in battery life, given the same types of usage in both cases. But I think they have just gotten used to "flooring" the brightness-control, and really haven't stopped to actually think about how bright the screen really is.
[doublepost=1484286979][/doublepost]

It sounds like there is about a 6x difference in power consumption between the machine idle at low brightness, and the machine at max CPU and GPU usage at full brightness. So someone might get 20 hours with the machine completely idle, or a bit over 3 hours running something like Creative Suite.

Which is about what you'd expect to get under the same conditions with other, similarly-equipped, similar battery-sized laptops like the HP 15" Spectre X360.

 

[idark77]

The 13-inch model without a Touch Bar had an average battery life of 18.75 hours, the 13-inch model with a Touch Bar lasted for 15.25 hours on average, and the 15-inch MacBook Pro with Touch Bar had an average battery life of 17.25 hours.

4781mA / 18.75h = 255mA average or P= 54.5Wh/18.75h= 2.9W average
4310mA / 15.25h = 283mA average or P= 49.2Wh/15.25h= 3.23W average
6667mA / 17.25h = 386.5mA average or P= 76Wh/17.24h= 4.41W average

Very very low power consumption...

 

[leman]

So CR was testing it wrong? If CR used random data for the web pages, they wouldn't need to disable cache. Am I Right?

CR should have implemented randomized URLs on the server end. That would effectively disable the cache but did in a way that a normal user could experience, without messing with Safari's settings.

 

[Mainsail]

It's like Consumer Reports is the James Comey of product reviewers.

 

[Jim Lahey]

Some people just actively search for a negative in absolutely everything. That must be a very sad and hollow existence.

Life isn't black and white. There are nuances to everything. Smile, the sun is always shining above the clouds

 

[nofear1az]

I think that Apple is guessing that most people who are using high-demand applications will do so while connected to AC power.

And in real life, that's a pretty good guess, at least for people not looking to "find fault".

The ONLY thing that is relevant is how long the essentially-equivalent 2015 MBP lasts, doing EXACTLY (not kind-of) the same things, and running the SAME version of MacOS.
[doublepost=1484284655][/doublepost]
But it appears that they actually DID shrink the battery and make it thinner AND it did NOT compromise anything, due to the hard work of both Apple and Intel.
[doublepost=1484285051][/doublepost]
Take another gulp of that Hatorade. Hope you like the taste of WRONG!

Apple admitted it was a Bug in Safari. They fixed the Bug. CR Retested. Case Closed.

Sorry to disappoint you; but not EVERYTHING is a Conspiracy.
[doublepost=1484285233][/doublepost]
Spoken like someone who has never written a line of production-code in their life.

One thing I agree with - is we need to see the results from a previous gen with this gen with the exact same scenario and software.

In no way is this case closed yet, too many vocal people are seeing different results that we still need more tests and verification on. You can defend Apple all you want, your crush on Apple is way too strong to accept clashing opinions.

 

[oldmacs]

But it appears that they actually DID shrink the battery and make it thinner AND it did NOT compromise anything, due to the hard work of both Apple and Intel.

Yet it appears that people with a 2015 and 2016 MacBook notice a difference - less battery life for similar tasks.

Battery life should have gotten better, not worse. Apple sacrificed battery and also performance improvements to unecesarily shave a little more off the thickness.

 

[leman]

Wait no! You are completely wrong!

The way Consumer Reports did their testing actually triggered the bug, and it was a bug that may have been there for many more years without the way CR did their testing (the flawed code was first developed like 14 years ago)

I am not really sure how to make it more clear. So let me try with an analogy:

Say you want to test how some person is affected by being in high altitude (mountains). Basically, you want to have them do some exercises in a high altitude, low oxygen environment and see how they manage. However, you don't the time to actually bring them to the mountains, so you decide to simulate the entire experience by injecting them with a drug that constricts the respiratory tract. The person gets massive asthma attack and collapses. You conclude that their physical condition is crap and they should avoid going to the mountains.

The only problem is: they had an allergic reaction to the drug. They would have done just fine by being in the mountains.

Ergo. Your assumption that CR should have implemented it at the server end, is a bogus solution. What that would have done is hide it, and then Apple would not have known about it, and not fixed it.

If you want to test "X", test "X" instead of testing "Y" in assumption that you are testing "X". Again, CR failed by a) not disclosing the fact that they uses a debug configuration and b) by not investigating the issue further, but releasing an official recommendation in a situation where the results were clearly inconclusive. Yes, in the end it was Apple's bug and I guess its the only good thing in this story that this bug has being fixed. But that bug would have never affected a normal user in the first place! And CR does its testing for the normal use case, not for the web developer.

 

[MacsRuleOthersDrool]

Prove it! (As you said to another poster getting the opposite.)

You are so BIASED in your opinions that you don't want to hear or see anything that does not go along with your beliefs. You have total fanboyism. Apple can never do wrong in your biased views so there's no reason to refute or even to have a real discussion with you. Millennials!

I'm 60 years old.

As an embedded developer (hardware and software) with several decades of experience, and several industrial embedded products under my belt, and several dozen (at least!) product spec-testing sessions I have conducted and/or participated in, I know bad test methodology when I see it, and I definitely saw it when CR got insanely disparate results in the testing, and then even suspected that Safari was involved; but STILL just "ran with it", rather than asking Apple if that seemed odd to them.

As for your "Prove it!" challenge, I offer this anecdote: The only person I know who actually owns a 2016 MBP, has a tb 15" model. He hadn't been following any battery complaints, and I hadn't discussed same with him. He is a degreed Electrical Engineer, BTW. So, when I casually inquired as to the battery-life he was getting, here is the snippet of the email exchange (the rest of the conversation was about GarageBand and Logic Pro X) :

ME: BTW, on an unrelated subject, how has your MBP's battery life been?

FRIEND: I can sit and read/write/surf for a couple of hours and it drops to around 85%. Of course, that's not heavy use, but it does keep the display lit full time - which is probably a bigger drain on the battery than some things. It charges very slowly if you use an adapter not rated for the MBP - no surprise there

Then, in my Reply to that email, I explained a bit why I was asking, and asked a couple of other things, and he subsequently responded:

ME: But you're not seeing sh** like run times of just a couple of hours TOTAL, right?

FRIEND: No, I've never pushed it past a couple of hours, but it has never acted up at all. In fact, I'll usually make it sleep instead of powering down, and 8-10 hours later I'll wake it to show about the same power level as when I put it down and I'll get even more time out of it before I set it aside and plug it in to recharge. I know you're supposed to run it low about once a month, but I haven't gotten around to trying that yet. When I do, I'll probably see just how long a charge lasts.

ME: 2 hours for 15% battery usage is right on-track for Apple's 10 hours for that sort of activity.

ME: What Browser?

FRIEND: Firefox - which, incidentally HATES being put to sleep and crashes. No data loss, but ouch.

And that was the end of that. I agree it is anecdotal "evidence"; but that's all we're getting from these other "I only get [x] hours" posts, anyway, right?

And at least I have tried to actually "Prove it". Not scientific; but at least I have tried...
[doublepost=1484289750][/doublepost]

How are they getting 18 hours!?

Even Apple advertise it as 9-10 hrs best.

Crappy test that is still not "real-world" enough for a highly power-optimized CPU like Skylake, coupled with a highly power-optimized OS like macOS Sierra. I suspect that, even with Caching turned off on Safari, the repetitive and predictive nature of CR's testing was being predicted by macOS, and it was actually able to take measures to "sleep" some parts of the CPU, etc. that Apple's test did not allow (or allow as much).
[doublepost=1484290030][/doublepost]

I'm not sure how this computer lasts 17 hours, but I'll give at a try...

Don't be surprised if you get more like what Apple suggests (8-10 hours surfing); because IMHO, CR's test is still not "Real-World" enough, cache or no cache.

And the fact that CR's initial testing SOMETIMES achieved this suggests that, when they weren't accidentally triggering the Safari bug, THEIR test methods WERE actually achieving those numbers.