General Discussion
  >> General Broadband Chatter


Register (or login) on our website and you will not see this ad.


  Print Thread
Standard User MHC
(legend) Thu 16-Feb-12 09:46:39
Print Post

speedtest.net


[link to this post]
 
It is better than it was and the quirky results have almost disappeared.

However, I do question the "Connection Grade" where it takes a 30 day rolling average for the country or world and then compares my results to that.

My results are around 37.5 to 38.0 Mbps - 37.81 equates to a 38716 sync. It grades te connection as A- 80% which indicates that 20% of users in the UK get average speeds greater than that when using speedtest.net.

How true or accurate is their assessment - even allowing for VMs product offering it does seem quite a way off.





~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~


M H C


taurus excreta cerebrum vincit
Standard User Deadbeat
(knowledge is power) Thu 16-Feb-12 09:54:38
Print Post

Re: speedtest.net


[re: MHC] [link to this post]
 
Speedtest.net often returns results which are way above my sync so I take the whole thing with a pinch of salt.
Standard User CurlyWhirly
(eat-sleep-adslguide) Thu 16-Feb-12 23:36:04
Print Post

Re: speedtest.net


[re: Deadbeat] [link to this post]
 
In reply to a post by Deadbeat:
Speedtest.net often returns results which are way above my sync so I take the whole thing with a pinch of salt.

Same here, not a reliable speed test IMHO.


Register (or login) on our website and you will not see this ad.

Administrator MrSaffron
(staff) Fri 17-Feb-12 09:34:33
Print Post

Re: speedtest.net


[re: MHC] [link to this post]
 
That will include leased lines, unis and large businesses

Andrew Ferguson, andrew@thinkbroadband.com
www.thinkbroadband.com - formerly known as ADSLguide.org.uk
The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
Standard User MHC
(legend) Fri 17-Feb-12 09:49:47
Print Post

Re: speedtest.net


[re: MrSaffron] [link to this post]
 
But those are probably just a small proportion of the unique results they have.

They also refer to the "Household Download Index" and say "this index compares and ranks consumer download speeds around the globe". I know it is not easy to remove all corporate based lines or tests but it suggests they may be doing so.





~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~


M H C


taurus excreta cerebrum vincit
Administrator MrSaffron
(staff) Fri 17-Feb-12 09:51:46
Print Post

Re: speedtest.net


[re: MHC] [link to this post]
 
They may be, but last time I drilled down into the results there was a lot of 100 Meg+ lines at a time when it was not sold to consumers - other than millionaires

Andrew Ferguson, andrew@thinkbroadband.com
www.thinkbroadband.com - formerly known as ADSLguide.org.uk
The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
Standard User MHC
(legend) Fri 17-Feb-12 10:18:03
Print Post

Re: speedtest.net


[re: MrSaffron] [link to this post]
 
How many is a lot? It would need to be in the hundreds of thousands. Most of the hi speed lines will probably be symmetrical and not a 100/2 ! and the tests would affect the upload results too - however the UK average upload is just 2.13 on their data.

They claim over 17M unique IP addresses for 86M tests - but just 4M being used in the current 30 day rolling average. Those figures suggest that 800k tests have given results of 40Mbps or better and a large number of tests will be run when a problem is perceived - which will pull averages down.

How much confidence do you have in their claims? For me it is currently low ...

I could download the raw data but it is a 1.5GB file !





~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~


M H C


taurus excreta cerebrum vincit
Administrator MrSaffron
(staff) Fri 17-Feb-12 11:39:22
Print Post

Re: speedtest.net


[re: MHC] [link to this post]
 
Well we have had people with leased lines run automated tests against our servers in the past, so without spending a few days crunching the numbers hard to talk with confidence.

Takes ages for us to filter out all this stuff from our results, hence why we don't do league tables very often.

Andrew Ferguson, andrew@thinkbroadband.com
www.thinkbroadband.com - formerly known as ADSLguide.org.uk
The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
Standard User michaelh
(fountain of knowledge) Fri 17-Feb-12 18:39:49
Print Post

Re: speedtest.net


[re: CurlyWhirly] [link to this post]
 
In reply to a post by CurlyWhirly:
In reply to a post by Deadbeat:
Speedtest.net often returns results which are way above my sync so I take the whole thing with a pinch of salt.

Same here, not a reliable speed test IMHO.


I find the results fairly consistent with other test sites

Standard User MHC
(legend) Fri 17-Feb-12 22:34:04
Print Post

Re: speedtest.net


[re: michaelh] [link to this post]
 
I am now finding that with a 40Mb FTTC connection the results are in line with other testers - maybe just a little high but consistent. I do still see some real way off results on the fixed speed (512k, 1MB & 2Mb) and up to 8Mb Max services.





~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~


M H C


taurus excreta cerebrum vincit
  Print Thread

Jump to