Is bt speedtester a random number generator?

Why have my latest three bt speedtests shown downloads in single or double kbps figures, whilst thinkbroadband tests carried out immediately afterwards display over 3000kbps (still pathetic, I know, but bear with, bear with). Uploads equally disparate. Previously tests gave broadly similar results.

Perhaps one of you experts could kindly explain?

Maybe the server is suffering traffic problems, it has a habit of only being good for displaying the IP Profile rather than a place to go and test speed of your connection

It's all very well being only good for the IPP, but if you want to report a fault, you HAVE to use the BT tester using an ethernet cable.

From what it looks like on the BT care forum, the tester is completely broken, generally giving 0 upload speed. Many people find it impossible to do the TAP3 test.

Also, people all over the country have been reporting Infinity slow downs on a very regular basis. Usual cure isa reboot - advice is often "you have a broken HH3", "get BT to change your HH3". But, the speeds come back again once you restart with a new IP (well, in a lot of cases). But, they drift down again - for some people it's hours, others it can be a day or so.

The fact that it "self heals" as well does point to a serious network problem out there affecting a lot of the BT infrastructure.

The most irking thing is that the call centre don't have a clue about it. BT is not giving out any information and the CEO simply brushes it off as "it's only affecting a few people". Well, the ack that it IS affecting people should be enough to set a plan in action to locate and fix the problem. I'm not holding my breath ...

But, my connection is still good after 13 days now with no mortal intervention!

Go figure ...

Cheers,

Rob

How much do you think BT might pay to run the speed test part for them?

Their tester has never really coped that well, and one problem can be providers passing almost every caller through the system

Well Andrew ....

Why not ask them ... the speed testing part is rubbish at the moment. Mind you with the plethora of problems around, I wonder how hard it's being hit ...

But, the problem exists at the moment that in order to report a fault you have to do a BT speed test, the last result being stored somewhere on a server which the call centre I assume can see. Mind you, the last part of storing also fails quite a number of times.

I use your tbb speed test a couple of times a day to keep track.

I use it via my wireless g, and see around 13-18 megs down and 4.6 up. These are the values I get when I use ethernet and get 37.5 down and 8.6 up. So, if my wireless tests are in the right ballpark, then I know my wired test will be pretty much OK.


speedtest.net gives slightly higher downloads, but very much higher uploads - usually around 8.6 megs up, and around 18-20 down.

I know they are based on different engines, and also the way the average speed is sampled as well (speedtest is more a peak I believe from what I read somewhere else on these forums).

Strange though how much faster the apparent upload speed is - but, I have to say the tbb stats are much more consistent.

So, how much are you going to bid then????

Cheers,

Rob

Edited by deleted (Tue 03-Apr-12 22:52:50)

The speedtester lite almost certainly is a random number generator. I've been monitoring it on my line for months and it differs from BT's full fat tester by anything up to a factor of 10. One day I'll pull out all the numbers and do the statistical test, just to be sure! On the other hand, it's always lower than the real speed so if they switch to it they're making a stick to beat their own backs (as if they actually give a hoot...)

Thanks guys.

I too have concluded that, as others have posted here and previously, the basic bt speedtester is an example of rubbish programming.

Trouble is that in the past bt have refused to accept findings of any other speedtester. Also, it gives info which tbb can't - e.g. ip profile/con rate.

I find it gives nonsense results. The tester doesn't seem to give long enough to get an accurate reading.
I always get higher speeds using usenet.
I can confirm this as the speeds reported match the speeds reported on my menu bar in both cases.
So is speed tester is giving 3.0 I see 3.0 on the menu bar, in the case of usenet it's 4.6.

A new post on a similar thread on bt's own forums reveals the rivalry (?) between bt retail and bt wholesale - the latter being said to be responsible for the speedtest prog.
One might have thought that if bt retail are frustrated there would have been some mecvhanism to transmit their dissatisfactions to bt wholesale. Sadly, in mega-huge organisations it doesn't seem to work that way.

I have now received this amusing response to my own complaint re duff info:

I�m sorry that you are not happy with our BT Speedtester website. The reason we ask you to use this website is because we can see the results on our systems of the speedtest. I can confirm that the information on this website is correct.

i hate to be beaten but in this instance have decided that life is too short ...

In reply to a post by spranglebolt:

One might have thought that if bt retail are frustrated there would have been some mecvhanism to transmit their dissatisfactions to bt wholesale. Sadly, in mega-huge organisations it doesn't seem to work that way.

It is not all down to BT ... OFCOM have a lot to answer for.

I had a specific issue and a few years back if nothing was happening the Customer Service Adviser would have called the Clerical Assistant working for the local Manager of the team responsible for resolving it.

Now, the CSA has to gain approval from their manager/supervisor to escalate the issue to the local interface Point of Contact (withing BT Business), that PoC calls their opposite number in OR and also documents it. The OR POC then calls the OR faults to find out what has happened. Then a call will be made to teh Team Manager ...

So, it was 2 people and now anywhere from 4 to 6 because of OFCOM demands. And who suffers? The customer. Who pays? The customer. Who benefits? The extra two staff that have jobs/ Who profits? OFCOM maybe?

Sorry, it's me again, but

thinkbroadband @ 20:38 this evening dl = 3895kbps

bt @20:41 dl = 1.5mbps

speedtest @20.47 dl = 4.10mbps

Numerous other tests immediately surrounding this, all giving results approx 4mbps (except for uswitch, which disappeared up its own orifice). I accept that if the discrepancy was limited to this one occasion, it could be due to traffic congestion etc. But this was only the latest of six consecutive occasions over the past week. The bt support analyst wrote on 05/04 'I can confirm that the information on this website is correct.' It really infuriates me that an organisation can get away with this nonsense, but I guess I'll just have to go away and get a life.