Hello there folks.

[robd53]

Hi. I have recently been connected to BT Infinity here in Northwich UK. Here is an example of my d/load speed:

 

 

Is this typical for a testmy.net test result?

 

I have a connection that is rated at 76Mb by BT, and when tested using the BT wholesale (BTW) test site I usually get 45 - 53 Mb with a wireless link, depending on the time of day.

 

Thanks,

 

Rob.

[TriRan]

Welcome to testmy.net! Is it possible for you to test over a wired connection or is wifi your only option? I'd like to see if that is the cause or rule it out

[robd53]

Thanks for the welcome.

 

It isn't possible to test with a cable link at the moment due to the physical layout of the house and I'm reluctant to have cabling installed because of the age of the building.

 

I have tested download/upload speeds using Speedtest.net from servers in the US, Texas in particular, and I get test results comparable to the BTW performance test i.e. around 50Mbps d/load and 9Mbps u/load.

 

Unfortunately this leads me to doubt the validity of the testmy.net test results! Sorry.

 

Rob.

[CA3LE]

Thanks for the welcome.

 

It isn't possible to test with a cable link at the moment due to the physical layout of the house and I'm reluctant to have cabling installed because of the age of the building.

 

I have tested download/upload speeds using Speedtest.net from servers in the US, Texas in particular, and I get test results comparable to the BTW performance test i.e. around 50Mbps d/load and 9Mbps u/load.

 

Unfortunately this leads me to doubt the validity of the testmy.net test results! Sorry.

 

Rob.

 

You're testing your UK connection to servers all the way in the United States... switch to the Amsterdam server https://testmy.net/mirror -- International routes are often congested.  But the results you've been getting here are accurately representing your international speed.  

 

Believe whatever makes you feel better I guess.  But I can tell you this much... you're getting 13 Mbps to the middle of the United States... not 70 Mbps.  If you visit lots of sites in the United States, then that's the speeds you can expect to get.  Otherwise, try switching the server to one closer to the UK.

 

speedtestnet and the BT test your mention both use flash and both are opening more than one connection.  TestMy.net is single thread... because opening more than one thread can mask certain performance issues.  A properly running connection doesn't need to open multiple threads to attain full speed.  Now, if you hit a bottleneck that's only allowing 10 Mbps (for example) then you open more than one connection down that pipe... then you can still make the connection look good.  But it's NOT really pulling those speeds, it's combining multiple threads into that one speed.  Showing you your Max line speed but not really cluing you into any problems that may slow down a single thread download.  ....You do have a problem at that distance, otherwise you'd score better.

 

Doubt the validity of the speed test that's been around the longest... or start realizing that the other tests you've been taking are designed by the people who benefit by making your connection appear to be running right.  You're not here to make yourself feel good about how fast your connection is.  You're here to make sure it's running correctly.  So that's what I show you.  Speedtest.net which was started by a former ISP CEO and BT has a vested interest in making your connection look as fast as it can.  I don't.

 

Here's a couple things to read if you still doubt me.  First, read this about flash speed tests and tests designed by Ookla (like speedtest.net). Why do my results differ from speedtestnet/ookla -- not only do they multithread as I mentioned above, they admit in their own wiki to doctoring results by as much at 30% to offset a poor protocol choice. ---- Now, I'm no scientist but I'm pretty sure that for an experiment to be valid, the results can't be altered. -- also read TestMy.net Legitimacy if you have a second.

 

Cheers!

- D

[Pgoodwin1]

robd53....hi from SW Ohio in the US.

CA3LE speaketh the truth. Those other sites give results that are not the truth. The value in the tests here are that when there is a problem or bottleneck, it shows in the results. Start collecting test data to baseline your numbers here when your equipment is operating normally. If you do most of your communicating with sites in Europe, use the Amsterdam server he suggests. If it's here in the states, use the Dallas one. Either way, once you establish what your "normal" results are, you can use variation from that average to determine your connection quality. It's a more real world result, and valuable because of that fact.

The only time I got any useful results from the ISP's ookla type sites was when the ISP's equipment had actually failed so bad intermittently that my connection speeds on their site showed almost zero. On another occasion, when my ISP's equipment had a signal out of range hi, the quality of their signal was degraded to the point where TestMy's speed tests indicated a problem, while the ISP's flash-ookla system showed the meter pegged into the green.

[TriRan]

A flash based test won't detect issues with wifi such as link speed slowdown they pretty much show you what your connection is capable of under perfect conditions not what your actually getting however that is simply my opinion from my own personal experience since you can't bypass wifi it would be impossible for me to help diagnose your setup because diagnosing it will require process of elimination