Old speed test results

I have just run a speed test, and wanted to compare to my previous runs (especially with my previous ISP) however I can't find the old ones? Can anybody point in the right direction? I used to see the full list of old ones after completing a speed test, but this isn't offered.

Many thanks

https://www.thinkbroadband.com/my-profile

Assuming you are logged in and scroll down the page.

Thank you so much!

Accessing my old tests confirms it - previously the TBB figure and HTTP were very similar (even with this ISP but in the earlier days) however this time they are significantly different.

Avg TBB - 36.6Mpbs vs Avg HTTP - 65.1Mpbs

Hmmm.

Another aspect, Andrew.

With the earlier version, generally my graphs shot almost straight up, then turned horizontal; with very rarely any kinks or variations.

With the new version, the initial TBB climb is something like the front of an A4 streamlined steam loco.

Throughout the new version plots, both on the initial climbs and the subsequent horizontals, they are riddled with minor peaks and troughs.

The summary averages have remained reasonably consistent.

------------------

To add to that, when I go to download the small graph image to my PC, I am "told" that I am viewing someone-else's results - although logged in etc.

I deliberately did the offered follow-up test, but again the "someone-else" message came up.

Edited by deleted (Sat 13-May-17 07:31:27)

In reply to a post by eckiedoo:

Another aspect, Andrew.

With the earlier version, generally my graphs shot almost straight up, then turned horizontal; with very rarely any kinks or variations.

With the new version, the initial TBB climb is something like the front of an A4 streamlined steam loco.

Throughout the new version plots, both on the initial climbs and the subsequent horizontals, they are riddled with minor peaks and troughs.

The summary averages have remained reasonably consistent.

See this post. Andrew has explained it several times, both during the development phase when people not logged in on the main site ran a test, or logged in people chose to run what was then the labs version.

To add to that, when I go to download the small graph image to my PC, I am "told" that I am viewing someone-else's results - although logged in etc.

If I understand what you are saying, because I don't recognise the word "download" in this context, clicking a speedtest result link in a post on these forums, including ones own, I think has always done that. You don't get transferred to the Main Site page where you are logged in, you just view a linked graph from a database and your status on the site is not checked.

In particular, your username on the Main Site and on the forums does not have to be the same. Mine isn't. How is either the forums or the main site to know to know that and know I am logged in on the other?

It is perfectly possible for a random person to sign up to either with someone else's nick from the other if that someone else is using the same setup as me. It doesn't cause any problems unless said random person then starts posting News Article Comments. The chances of that are remote unless someone does it deliberately to cause confusion, which I'm sure would get sorted out. (However, as a precaution, by the time you read this I shall have moved to prevent it ). (Edit: It is no longer possible unless the second person knows the forum user's email contact address. Maybe this was always the case but I don't think so. I seem to remember the trick was once used).

The BQM is the same.

I deliberately did the offered follow-up test, but again the "someone-else" message came up.

Are you referring to some other post? What "offered follow-up test"?

Edited by RobertoS (Sat 13-May-17 09:12:38)

Less artificial smoothing and single thread is http rather than a custom TCP protocol so not as smooth. Now I could flatten things with maths but that would also mask issues some people have. Once I've seen one of your tests I can see if things fall outside reasonable

On the speed test and 'viewing someone elses' are the tests available in the My Profile section, if so proves the account linking worked at least, what I need is a link to a test that has done what you say.

If the old figures were flash version then that was a custom TCP protocol and with newer test (which has been running 3 years) the single thread is just one http download rather than the 6 of the httpx6 version

This means there may be a difference for some people

Thanks for your reply, explanations and the link.

I have just re-run the test -
When it finishes and does the initial, larger graph display, clicking the Blue "Results Page" button at the bottom-left, to get access to the small version of the graph, to download that under right-click "Save Image -

Above that small graph the following appears -

"You are viewing one of your previous results. Run a new speedtest now"
(On earlier runs over the days of the new test, this has been that I was viewing someone-else's results)

suggesting that the small graph contents are from a distinctly earlier test run on this occasion, not the immediate last one.

But for example I can positively identify it as the immediate last test run as when it was running, the "meter needle" shot up to a very high reading, which on the larger graph is stated to be -

"6.8s 175.69 Mbps"

which even allowing for the lessened smoothing, seems extremely high on a 40/10 contract.

This was the test?
http://tbb.st/1494680101762718455

The spike does not happen all the time as you had a 'normal' one earlier http://tbb.st/1494655397347177855

Looking at your tests the system is seeing you as logged in, and do the tests appear on the speed test list under your My Profile page at https://www.thinkbroadband.com/my-profile

On the needle spiking, if the browser tells us that was the speed not much we can do, I could code something that looked at the theoretical speed for you and limit the figures displayed but that would be bad for many other reasons.

Afternoon, Andrew.

Yes, that is the one.

I was taken aback for such a high speed to appear; but being under 240 Mbps, ie 6 * 40 Mbps, in itself may be correct. Not well enough acquainted with such details of Broadband!

Agreed that it has not appeared every time with the new test - this morning was the first time I was aware of a similar spike.

And given that there is limited averaging now, that probably explains it; and in that circumstance, I would not expect you to make any changes. Just trying to understand it.

Are you using either Kaspersky or (IIRC) Avast? This used to be a big problem with Kaspersky but was cleared, then popped up on the other for a while.

Believed to be related to buffering at the user end while the initial traffic is vetted, then released from the buffer with a burst then a steady stream.

Evening Bob

Not that I am aware of, using Kaspersky or Avast.

It is simply looking at the new TBB Speed Test Results; and trying to fathom out the differences relative to the old Test.

----------------

There are no obvious differences in using the PC on the Internet, so puzzling as to why the differences.

Fair enough about there being less averaging, that explains the minor peaks and troughs.

But that very high peak; and the statement -
"You are viewing one of your previous results. Run a new speedtest now"
when clearly it is the most recent etc.

---------- Just done another new TBB Test, getting that statement; but now there are no extreme peaks, athe highest being 65.02 Mbps.

Interestingly though, the overall or headline average has risen from generally around 37 Mbps to 41 Mbps.

I was meaning to cause the high initial peak. That's what they did.

What IS system do you have then? I said I wasn't sure the second one was Avast.

Not sure by what you are referring to as "system".

I am on an EE VDSL 40/10 Contract for almost 3 years now, selected for price, as I do not watch videos etc.

Bright Box 2
Fast Path
G993.2 (VDSL2)
Noise Margin - Up 22.0, Down 18.0 (Clean on the QLT)
Line Attenuation - Up 0.0, Down 13.4

Data Rate Up 9999 Kbps, Down 40000 Kbps

Uptime 1,342 Hours


-------------

About 300 Metres D-side length from the FTTC although ironically the FTTC is only about 10 Metres from my front door and I can see inside it when opened, from my lounge window.

The PCP is about 40 M away, both from the FTTC and the front door.

The 250 M original D-side length is absolutely proven, although the literally-precise routing is not clear, apart from an unusual junction box about 125 M away on a "near-neighbour's" house - where noises from a poor connection were cured, about 30 years back.

IS (Internet Security) system, as in Kaspersky/Avast and many others. Which was the subject we were discussing that could be the problem. I was simply asking what you use, in case I was mis-remembering Avast, as I had said I might be when I first mentioned it.

To be honest, I don't know.

Looking down through Task Master, possibles are -

Advanced SystemCare
AVG
IObit Malware Fighter
McAfee Security Scanner
MS Malware Protection

Windows Processor
Antimalwasre Service Executable
Local Security Authority
-------

I appreciate that some or all of these may not be relevant; and also there may be degrees of conflict similar to running two or more anti-virus programs simultaneously - but I have not had any such conflicts flagged up, so on the "sleeping dogs" principle ...

Thanks for your interest etc.