You would have thought after all this time the TBB and VM could have arranged for a better connection given that accurate results are in everybody's best interest.

I am assuming, of course, that accurate VM stats are in TBB's best interest.

Re speedtest.net. I guess if the OOKLA's algorithm shows everybody off at their best, that is at least a level playing field for all ISPs.

Ookla can suffer from peering issues too, and often it is harder or impossible to get feedback, as least with us Seb and John can investigate the network, and has shown in the past that by alterating routing things can be improved on a temporary basis.

Virgin is more than welcome to put some private peering in place, the odd thing is that it does not affect all Virgin users equally, i.e. we do see people getting good speeds.

In reply to a post by MrSaffron:

Ookla can suffer from peering issues too, and often it is harder or impossible to get feedback, as least with us Seb and John can investigate the network, and has shown in the past that by alterating routing things can be improved on a temporary basis.

Virgin is more than welcome to put some private peering in place, the odd thing is that it does not affect all Virgin users equally, i.e. we do see people getting good speeds.

Iv also noticed thar flash based speed tests arnt as accurate has java based tests.

In reply to a post by Bobby_Valentino:

Iv also noticed thar flash based speed tests arnt as accurate has java based tests.

It might depend who programs the flash test, I have no problems with this flash one for example:
http://www.thinkbroadband.com/speedtest/broadband.html

If people are getting different speeds on the TBB Flash and Java testers that shows an issue either with the link from TBB to their ISP, or from their ISP's border routers to the customer.

In reply to a post by MrSaffron:

Also the algorithm used by ookla has a tendency to show a good result even when a link is congested

Yes that's interesting. I had a chance to play in the US last month (family over there). They subscribe to Time Warner Cable's highest speed service - 50mbps download, 5mbps upload.

Using a wired connection to their ISP required all in one box (same idea as VM) they achieve 50mbps download and 5mbps upload from the local speed tester (ookla powered) at Time Warner, and using one about 500 miles away from a big competitor Comcast. Almost all the independent testers were pretty awful around 10mbps max!! (I didn't try thinkbroadband, as generally the UK connectivity is slow, I'd anticipate 5mbps on a good day).

Wireless (N with WPA2) was always poor, only 30mbps download, and kept stuttering - inSSIDer showed lots of local networks but all low enough they shouldn't affect it. I think its a really poor wifi unit. The stuttering showed up browsing on phones (Android and/or iPhone) and iPads, and on desktop PC showed in "outlook as lost connection" messages. Annoying.

But that is the max they can get, their entire housing estate (neighborhood) can't get DSL, too far from the CO (exchange) and so Time Warner have a captive market!

Many are not aware of the lack of choice for most US residential customers and often the same in Europe particularly with the new fibre solutions.

Agreed; I work with people across europe. Mostly similar as US, where distance means DSL would be 128kbps at best :-/ Some however have multiple providers over cable-tv coax; something VM should offer. (e.g. more "pro services" with fixed IP etc).