Anandtech's new battery life ratings

[Pataling]

Could someone please explain Anandtech's new tests for phone and table ratings. I bought an iPhone 5 in large part because it was topping the Anandtech charts but I'm not getting anywhere near 8 hrs on LTE in my regular usage and indeed I found 3G to be much better on my phone. Overall it still didn't match the battery life of the iPhone 4 I had previously ... I returned the iPhone5 and I'm now trying to figure out how to read their ratings to determine which phone to get. Any tips?

 

[God Mode]

Results may vary with usage and signal quality. Even if it shows all the bars, I get different standby operation times with each different location. The worst is if you take a subway for extended periods and the phone repeatedly tries to gain a signal.

 

[Skurge]

AT battery life tests load a new page at a fixed interval. So since LTE loads faster, the phone goes into idle state sooner and thus has longer battery life.

If you load a new page as soon as the current one is done loading, then AT's battery tests don't relate to your usage scenario and you will get much worse life than is stated on their charts.

 

[TuxDave]

Could someone please explain Anandtech's new tests for phone and table ratings. I bought an iPhone 5 in large part because it was topping the Anandtech charts but I'm not getting anywhere near 8 hrs on LTE in my regular usage and indeed I found 3G to be much better on my phone. Overall it still didn't match the battery life of the iPhone 4 I had previously ... I returned the iPhone5 and I'm now trying to figure out how to read their ratings to determine which phone to get. Any tips?

How much life are you actually getting vs iP4? Is your usage rate different w/ LTE? We need more information because we can't help you.

 

[kpkp]

AT battery life tests load a new page at a fixed interval. So since LTE loads faster, the phone goes into idle state sooner and thus has longer battery life.

Could this explain the dominance of the iPhone 5 in this test, since it seems that Safari javascript engine is better optimized then the one found in Android (btw do the new Nexuses use the Chrome javascript engine?), giving the A6 longer idle times. And in general I have the feeling that the test in question is not really proportional to the battery life in a day to day use and therefore makes the Apple hardware shine a bit more then it deserves?

 

[jersiq]

I only read one of the reviews showing the testing methodology, and here's where there can be a disconnect. It looks like they are emulating RRC connected and idle states. In idle, the DRX paging cycle index is king. If it's to aggressive, battery life goes down the tube. Too lenient, and there is an opportunity to miss down link data and force TCP re transmits.

So, without knowing what the DRX paging cycle is in comparison to a network operators DRX, there can be some variability.

Although, I may be misunderstanding how they explained the test methodology.

 

[BenSkywalker]

Could this explain the dominance of the iPhone 5 in this test

The staff at AT rigged the results for the benefit of Apple. They went out of their way, compromising their integrity in the process, to make Apple look good. We have a couple of devices that set the standard for battery life, I'm not sure if they were given instructions to, or simply of their own accord decided to exclude them to make Apple look better then they would otherwise. At this point AT has fallen down to applewtfpwnzjoo.com levels of integrity when it comes to reviewing mobile devices of any sort. They are clearly lap dogs doing Apple's bidding(GLBench exclusively, let's not have the Maxx on battery tests, lets not use GeekBench or LinPack).

 

[ImDonly1]

OP if you are streaming non-stop, then LTE is on the whole time and doesn't do the whole power on/off thing and drains battery faster?

If you are reading a webpage and it loads it fast then goes to sleep, better battery life with LTE, in theory.

 

[Mopetar]

The staff at AT rigged the results for the benefit of Apple. They went out of their way, compromising their integrity in the process, to make Apple look good. We have a couple of devices that set the standard for battery life, I'm not sure if they were given instructions to, or simply of their own accord decided to exclude them to make Apple look better then they would otherwise. At this point AT has fallen down to applewtfpwnzjoo.com levels of integrity when it comes to reviewing mobile devices of any sort. They are clearly lap dogs doing Apple's bidding(GLBench exclusively, let's not have the Maxx on battery tests, lets not use GeekBench or LinPack).

You want a tinfoil hat to go along with that post?

 

[Red Storm]

I don't think AT is fudging numbers or trying to make Apple look better. However, their new test results do not correlate to what I have seen and experienced in real life.

 

[lopri]

AT also don't include numbers from the original iPad, iPhone 4 (but includes 4S), or Nexus 7 (but does Nexus 4/10?).

You don't need to "fudge" numbers to make certain products look good on the charts. You just need the "right" numbers from "right" tools. ^^

 

[TuxDave]

OP if you are streaming non-stop, then LTE is on the whole time and doesn't do the whole power on/off thing and drains battery faster?

If you are reading a webpage and it loads it fast then goes to sleep, better battery life with LTE, in theory.

Heh I wonder if the OP is coming back or is this just a one shot post.

 

[Suspicious-Teach8788]

OP if you are streaming non-stop, then LTE is on the whole time and doesn't do the whole power on/off thing and drains battery faster?

If you are reading a webpage and it loads it fast then goes to sleep, better battery life with LTE, in theory.

But it depends how heavy of a website right? Anand is testing with heavy desktop websites... say like theverge. But if you load CNN mobile edition, it's a matter of 1.5 sec vs 3 sec on LTE vs HSPA. I think in that case 3G might win out, but if you load heavy ass websites, yeah I can see 3G taking much longer.

That's why I question the usefulness of this test. When you do surf tests, is it reasonable that you load a page every 10 seconds or every minute? I think they need to simulate 3G/4G use that accurately matches a user's surfing patterns. Does it really make sense that 3G use uses more battery or 4G in regular surfing? I don't know.

 

[BenSkywalker]

You want a tinfoil hat to go along with that post?

You test a new high end video card, what is the first part you have in your lineup to compare it to? The existing high end video card.

You test a new high end processor, what is the first part you have in your lineup to compare it to? Existing high end processors.

You test battery life on cell phones what is the first party you make sure you exclude? Anything that makes Apple look bad.....?

That isn't tinfoil hat level. Clearly AT set out to make Apple look good, why else would you exclude *the* bar setter for battery life? And not just one of them, both the original Maxx and the Maxx HD? If AT released a review comparing the GTX 680 to the AMD 5870 excluding the existing and well established 7970 people would be outraged at what would be a very explicit display of extreme bias. But when they do it to make Apple look good, it is perfectly fine. AT has removed themselves from a viable source for review information for anyone who doesn't worship Apple.

 

[Suspicious-Teach8788]

You test a new high end video card, what is the first part you have in your lineup to compare it to? The existing high end video card.

You test a new high end processor, what is the first part you have in your lineup to compare it to? Existing high end processors.

You test battery life on cell phones what is the first party you make sure you exclude? Anything that makes Apple look bad.....?

That isn't tinfoil hat level. Clearly AT set out to make Apple look good, why else would you exclude *the* bar setter for battery life? And not just one of them, both the original Maxx and the Maxx HD? If AT released a review comparing the GTX 680 to the AMD 5870 excluding the existing and well established 7970 people would be outraged at what would be a very explicit display of extreme bias. But when they do it to make Apple look good, it is perfectly fine. AT has removed themselves from a viable source for review information for anyone who doesn't worship Apple.

Dude whatever man. I'm not saying Anand isn't perfect. He makes plenty of mistakes left and right, but to insinuate that these benchmarks are a conspiracy designed to cover up Apple and to pimp Apple up is ridiculous. Unless you have solid evidence he's getting paid off by Apple, I suggest you shut up. I can point out plenty of his graphics cards reviews where it doesn't make sense at all his comparisons. There's more than a few reviews where he doesn't look at previous generation cards or neglects logical comparisons. Who knows why he doesn't have the Maxx or Maxx HD. Remember they don't just hold onto all these phones in a single lab. I think some people in AT actually use the phones themselves or it's their phone. You might not like the fact that its missing a bunch of phones, but it has the iPhone 4S, iPhone 4, and iPhone 5. It compares against previous Apple phones. It also compares against flagship Android phones like the SGS3, HTC One X. If I was a WinMo fanboi, should I also complain and cry Anand didn't include Windows Phone to make the platform look bad? Jeez.

Look, one day smartphones will be huge on AT that perhaps Anand will have a smartphone comparator on the bench. Maybe not. But seriously I think you're taking it too far.

And really I think we're arguing about the testing methodology here about 3G/4G web browsing and how the browsing battery life test disagrees with its previous testing results, and in general the results that other reviewers get, not about some ridiculous Apple conspiracy.

I think there's plenty of criticism I have for AT's smartphone reviews, but I'm pretty sure it's not a freaking conspiracy.

 

[Bman123]

The numbers are way off thats for damn sure. How can a phone get double the battery life on lte then it gets on 3g. I saw the chart and laughed that shit is way off. I never use AT for anything review related. The phone chart is a prime reason why...

 

[Suspicious-Teach8788]

The numbers are way off thats for damn sure. How can a phone get double the battery life on lte then it gets on 3g. I saw the chart and laughed that shit is way off. I never use AT for anything review related. The phone chart is a prime reason why...

Ok, I think it's clear depending on how you test a phone, battery life can vary.

If you leave a phone on downloading the whole time on battery, it's very possible that LTE uses more power and dies first compared to HSPA. Howerver surfing tests a fixed amount of bandwidth being loaded, and since LTE is faster, the phone spends more time in idle.

It's the same argument when we went to high power chips. In moving from ARM11 to Cortex A8, the iPhone 3GS was able to maintain its battery life. That's not because the A8 uses less power. It uses more under load, but in general if you can get tasks done quicker, you can get to idle faster.

So obviously if you leave the phone on running at 100% load, The 3GS dies faster.

Bottom line is I think the battery life depends on the tests you impose. I'd like to get a better technical understanding of HOW these tests are run specifically and what changed compared to the old tests. That way I think we can better understand how the results changed.

 

[Mopetar]

That isn't tinfoil hat level. Clearly AT set out to make Apple look good, why else would you exclude *the* bar setter for battery life? And not just one of them, both the original Maxx and the Maxx HD?

If I had to guess, it would be because they don't have those devices around and haven't been able to test either since they've changed their test.

But when they do it to make Apple look good, it is perfectly fine. AT has removed themselves from a viable source for review information for anyone who doesn't worship Apple.

What a load of crap. If you posted something like this targeted at another forum user you'd get an infraction. Have you tried actually emailing any of the team at AT and asking them for a clarification about why they MAXX isn't in their benchmarks or when we can expect it to be included?

No, you probably haven't. Instead you go off on a rant about how AT is clearly biased and is intentionally trying to make Apple look good. This doesn't even have anything to do with Apple. You could have said the same for any other company and I'd still point out that it's a terrible post, with no evidence, that jumps to conclusions.

The numbers are way off thats for damn sure. How can a phone get double the battery life on lte then it gets on 3g. I saw the chart and laughed that shit is way off. I never use AT for anything review related. The phone chart is a prime reason why...

Did you even read the review, where it's clearly explained why this happens? With LTE the page is downloaded more quickly and the radio is able to shut off. With 3G it takes longer and so the radio has to stay on longer, drawing more power. It's the same idea behind "Hurry up and wait" or "Hurry to rest" that we saw in CPUs a while back where it was often deemed to be much more efficient to quickly complete a task and get to an idle or low power state.

If anything you can argue that this might not be indicative of actual usage, and you may be right, but it's still a methodological approach to measuring battery life. Honestly, about the only thing I use for reviews any more is AT because they actually use a systematic approach for nearly every aspect of their review and usually describe the methodology enough where I can decide if it's good or not. Other sites just go with a lot of gut feeling, which doesn't always line up with reality.

 

[3chordcharlie]

'Clear bias in reviews' not found.

'Questionable usefulness of reviews' found in spades.

I strongly suspect BSW is wrong about motivation. I certainly trust and hope that he is.

He's not wrong about the large hole in the dataset. For the moment, I'm willing to chalk that up to happenstance.

 

[Pataling]

Results may vary with usage and signal quality. Even if it shows all the bars, I get different standby operation times with each different location. The worst is if you take a subway for extended periods and the phone repeatedly tries to gain a signal.

AT battery life tests load a new page at a fixed interval. So since LTE loads faster, the phone goes into idle state sooner and thus has longer battery life.

If you load a new page as soon as the current one is done loading, then AT's battery tests don't relate to your usage scenario and you will get much worse life than is stated on their charts.

I got the iPhone5, activated the sim, restored from my iPhone4 backup and continued on with my day. I use the phone most often for browsing during my commute and I don't think my usage was different. Perhaps the bigger screen meant I read pages faster ? I tried restoring to factory settings and also tried turning off LTE so with just 3G it should be the same conditions as the iPhone4. Maybe I got a broken unit. Anyway thanks for the tips; seems Anandtech's results can significantly diverge from real world usage

Heh I wonder if the OP is coming back or is this just a one shot post.

The Apple conspiracy theories almost scared me away. I guess Anandtech reviewers use Apple products as their personal devices but then I've usually found the reviews of competing products to be quite rigorous whenever Anandtech gets around to putting them up.

 

[BenSkywalker]

Have you tried actually emailing any of the team at AT and asking them for a clarification about why they MAXX isn't in their benchmarks or when we can expect it to be included?

Let's cover some of the *massive* flaws in their testing methodology.

Let's look at their CPU benchmarks...

Oh wait,

That's right,

They don't run any at all anymore. Zero CPU benches. Zip. Zilch. Nada. Why? They call it 'non represenative'. LinPack, Pi, GeekBench, plenty of options for them to use. If you are reviewing a computing device, how can you possibly justify *not* having a *single* CPU benchmark? None, at all. Zero. It is absurd to a rather astonishing degree.

Then we get to the GPU benchmark. Where is Basemark, or Antutu? Antutu is available not only on Android and iOS, it is also available for WP. Instead, we get PowerVR's press vehicle and *nothing* else. If AT ever released a review and only used a single game that everyone knew made one GPU maker look incredibly good, it was the only bench they would ever use refusing to use anything else that painted any other picture people would go nuclear over the insane levels of bias. Somehow, this is supposed to be OK for mobile devices? To call that a stretch of logic would be a *bit* of an understatement.

I brought up the Maxx as that was extremely clear cut. We have a bar setting device for battery life. Everyone knows this. It wasn't included with no explanation whatsoever. Given that they have removed all but one suspect bench for anything but browsers already, does it really take much to figure out what is going on?

Have the massive flaws in their testing been pointed out to them? Repeatedly. Open the comments in the review thread. They don't even try to defend themselves anymore. I think that most people have just accepted that when it comes to AT, if you aren't a die hard Apple fan, you just shouldn't bother reading their carefully worded press releases.

 

[at80eighty]

damn, you have to be hardcore playerhater to rag on the AT crew. fwiw im itching to toss my iphone away; so no real love for Apple - but reading AT (& just sometimes, The Verge) is about the closest you can get to objectivity in this day & age.

 

[cheezy321]

Conspiracy theorists always make me LOL.

AT is one of the best places to get reviews of any device period. I do think the omission of the MAXX is questionable but I doubt it was due to some mandate brought down from apple. You have to be a little paranoid to think that.

 

[A5]

It's weird, considering the AT review is usually one or two weeks behind the other sites out there. You'd think they would have time to run more benchmarks, but they seem to spend it all doing deep-dives on architecture stuff.

 

[BenSkywalker]

damn, you have to be hardcore playerhater to rag on the AT crew. fwiw im itching to toss my iphone away; so no real love for Apple - but reading AT (& just sometimes, The Verge) is about the closest you can get to objectivity in this day & age.

Zero CPU benches, one highly biased GPU bench and you can't very easily find better? Try 'fanboy blogger rant'- probably be a more telling review then what they post at AT for mobile devices.

Conspiracy theorists always make me LOL.

It has nothing to do with a conspiracy theory. Zero CPU benches. That is a point of fact. They have a bunch of browser benchmarks and one highly slanted GPU bench. That is it. That isn't looking for a conspiracy- they suck at reviewing mobile devices, period. This has been pointed out to them many times over, their mobile reviews keep getting worse and less useful. They will go in depth if say, Apple releases a high PPI device, then all of a sudden that is important and that becomes their focus. Apple releases an updated browser engine? Drop all the CPU benches and only run browser benches. Apple has a highly calibrated display? Start running a full page on display calibration and drop the majority of the GPU benches. That isn't conspiracy- that is what happened. Conspiracy would be me saying that Apple is paying them to do it, I don't think that for an instant. They just *suck* at reviewing anything mobile that doesn't have "Apple" in the name.