Verizon’s 3G is the fastest, while AT&T's one is the slowest, says Wired

A test has been conducted by Wired to measure the performance of the 3G networks of four major U.S. carriers…
This is a discussion for a news article. To read the whole news, click here

45 Comments

13. PapaJay224

Posts: 866; Member since: Feb 08, 2009

We've been saying this all along...it sounds like people are too use to being lied to in the commercials (cough) At&t

15. rwolf1984

Posts: 536; Member since: Jun 06, 2009

3G CDMA is not faster than 3G GSM 3.6 let alone 7.2. . .Consider these tests. . .Conducted in Northern Ohio. Both phones present on site. 1. VZW EV-DO (3G CDMA) Curve 8330 vs AT&T EDGE (2G GSM) Curve 8320\ AT&T Curve loaded 3 test websites faster. Both phones had a cleared web cache and full were in full network coverage. 2. VZW EV-DO (3G CDMA) Curve 8330 vs AT&T Blackjack II (3G GSM) AT&T Blackjack II loaded youtube videos faster via mobile youtube 3/3 times. 3. T-Mobile 8900 EDGE (2G GSM) vs AT&T EDGE (2G GSM) Curve 8320 T-Mobile 8900 could not acquire EDGE service, AT&T won again loading 3 test sites. Conclusion: AT&T and Verizon have the best coverage at the test site. . .AT&T was faster. Location of Test Site(s), Availibility, Device, Processor Speed, RAM need to be taken into consideration. The Wired test had too many variables. Who participated? What kind of users were they? Too many questions to answer, scrap this and get more scientific!!

21. BlackberryUser

Posts: 609; Member since: Jun 26, 2009

It's my understanding that GSM has more available bandwidth then CDMA, hence the push for 4g on CDMA. That said, if a wired network connection at home is 100mb, that doesn't mean your internet is faster then someone on wireless with a network connection of 54mb. Although it appears your tests did take more facts into consideration and eliminate a lot of the variables, it's still flawed. Tests need to be performed in "working" environments (I'm guessing Northern Ohio isn't the cell phone capital of the world), tests should be performmed at the same time, the same phones should be used, and the tests should be performed in multiple cities at locations with full coverage for all carriers. AT&T may be capable of the fastest internet, but do they actually provide it? How many users are spread out on each cell tower on AT&T vs Verizon. Either way, all of these tests will forever be flawed and no results really usable. AT&T was great for me when I lived in Texas and was horrible for me in San Diego. Verizon was untouchable in California and is good in Arizona, but not as good.

23. SellPhones82

Posts: 569; Member since: Dec 11, 2008

Here are some real tests that can not be faked by using Wifi. http://www.dslreports.com/mspeed?domains=1 These are from a mobile speed test site which lists your domain so if you use Wifi it will not show up under the providers domain name. These are much much lower than the numbers above and are a true real world test. Also the number of samples are all more than 2000 for the big 4 with Sprint having the most samples at 3899 and T-Mo the least at 2171.

24. vzw fanman

Posts: 1977; Member since: Dec 11, 2008

it won't let you do those tests with blackberries though... i use to do it with my voyager and it worked great.

26. SellPhones82

Posts: 569; Member since: Dec 11, 2008

Good info fanman had no idea you can't do a speed test with a BB. Looks like it has something to do with the browser blocking certain info and you will receive inaccurate info due to compression and data filtering that bes/bis servers perform. Makes me wonder how many people on this Wired test were on BB's making the results even more worthless...

29. rwolf1984

Posts: 536; Member since: Jun 06, 2009

I provided results of tests done at one location with 3 different carriers, during peak hours during the week and 4 different but comparable smartphones. I compared VZW blackberry to AT&T blackberry. AT&T 3G smartphone to VZW 3G smartphone. For that test I did not use the Bold because the processor is not comparable to the curve and the phone was much newer than 8330. The phones performed the same tasks 3 diffrent times. This is as real world as it gets. . .for the test site. I did not claim any network was faster nationwide. Many variables determine network speed. You cannot guarantee full coverage at sites. Which brings me to another point I dont think any carrier is capable of providing a true coverage map or accurately estimating speeds for ideal use nationwide.

40. behold--me

Posts: 679; Member since: Jun 22, 2009

dude. read ur article n see if it makes sense. DUH!!!! uh thats just ohio. not the entire us like this test. DUH

19. jskrenes

Posts: 209; Member since: Dec 11, 2008

Technically the numbers aren't unscientific, (any sample size over 30 is sufficient) but the study could have been done a lot better. I would have pulled 100 urban users, 100 suburban users, and 100 rural users from every carrier. Give them all or make sure they are all Blackberry Curve users (since the Curve is common across all carriers, and make sure you pick models that have comparable processor speeds), and cripple wifi (not a problem for VZW). Have them use their phones during different times of day to account for traffic, and you come a lot closer to accounting for a lot of those variables.

32. Kiltlifter

Posts: 742; Member since: Dec 11, 2008

Yeah... to get a truly accurate test you would have to pot a tower that has Sprint, VZW, T-Mo, and AT&T all form the same location, use roughly the same phone, like the curve 3g then test each one.... sadly this is almost impossible to test.... hills, network :"holes" etc.... but the test can be at least semi-accurate that vzw does invest a ton into their network which may be why there numbers were bigger... but their could be 1000s of variables scewing the data.

35. vzw fanman

Posts: 1977; Member since: Dec 11, 2008

it doesn't matter if all the carriers have a cell site in the same place... if 100 people are using verizon and 8,000 use at&t, at&t's network will perform slower and less reliable. and some people's phones constantly switches which cell site they connect too even if you just walk a couple of feet.

41. behold--me

Posts: 679; Member since: Jun 22, 2009

see that is something for t-mo to brag about. vz i think is the fastest def. but att last? wow someone in their company is reallly pissed right now lol

42. phoneguy

Posts: 213; Member since: Jan 08, 2009

I call bullshit on this article. There is no way in a frozen hell that t moblie of all carriers came before Sprint. VZW and Sprint use almost I dentical technologies for their EVDO service. Tmobile has one 3g tower in the world. CRAP!

43. TheCellDepotVZW

Posts: 26; Member since: Dec 26, 2008

its good to know that with Verizon phones (except the select few) no Wi-Fi is even avaliable and with 1000 people doing that test it is still the fastest for speeds and does not actually need Wi-Fi to have a fast network... Although it would be nice every once in a while!

45. j.naueattrep

Posts: 24; Member since: Jul 19, 2009

Guarantee you they weren't all downloading the same fileset from the same websites. The bandwidths, and website server speeds factor in too. Retest with the same files from the same website because VZW and Sprint CAP OUT max speed 1.5Mbts while At&t and T-mobile cap out at 3.6Mbts.
This copy is for your personal, non-commercial use only. You can order presentation-ready copies for distribution to your colleagues, clients or customers at https://www.parsintl.com/phonearena or use the Reprints & Permissions tool that appears at the bottom of each web page. Visit https://www.parsintl.com/ for samples and additional information.