Difference in spped tests

speed tests at dsl zone and broadnandmax are over 6.7MB, thinkbroadband down as 1.3MB. can anyone shed light on the reason why such a difference?

speed tests at dsl zone and broadnandmax are over 6.7MB, thinkbroadband down as 1.3MB. can anyone shed light on the reason why such a difference?

They work in different ways, and the one here often gives stupid results.

The best test, by far, is to actually download a file (using a download manager, such as Flashget), with the download split into multiple parts (say, 10), and see how fast it goes.

it all depends on the server traffic also, if its in high demand it will cause u to get a slow result.

Stupid - as in that was what was measured at the time? Or stupid as it totally disagrees with cross checks?

e.g. watch DUMeter and see how long the download takes, some also use two streams which gets around some simple traffic limiting or computer setup issues

Another issue of course is what ports the various shapers use.

Stupid - as in that was what was measured at the time? Or stupid as it totally disagrees with cross checks?

Actually works fine for me (except on IE7 - some problem with Java which I can't be bothered to fix), but I've seen too many 0.1Mbps results (and have had some on other connections) to believe there isn't a fundamental problem with certain connections.

I'm sure the problem isn't the speedtester itself, but how it interacts with the connections available from some ISPs.

The 0.1Mbps results cannot possibly be attributed to traffic shaping (what kind of ISP would limit to 100kbps - unless you've gone over the usage limit?).
I'm sure the problems are related to a few ISPs, and the small initial check carried out (by the TBB speedtest) to determine how large the actual speed check data should be.
If anything interrupts this small test, or the initial check comes back too slow, the size of the real test will be much smaller, and liable to the same "interruptions".