Search This Blog

Tuesday, 3 August 2010

Ofomc and samknows claim average broadband speed increase

Really? Might that be because in the 3 years ish this "experiment" has been running, our requests for the special routers to be placed in rural areas rather than purely urban areas have managed to achieve the grand total of 5 being sent out. One I know blew up as it was plugged into my house when we were hit by lightning, the other 4 are on varying degrees of rural connections which on occasion fail entirely to actually connect to the outside world.

This blog post can be read at

If you, Ofcom and Samknows, are going to claim the average speed of British connections is increasing, then please define by location (urban vs rural) exactly where these routers are. Because I hate to be a cynic but aren't the majority of your routers in places being cherry picked by the telcos for ADSL 2+ etc???

I could say the average speed on all British roads is 60+mph if all I measure are the parts of the motorways with least traffic and most lanes. Aren't you skewing the statistics in a similar way? After all, who is paying for the end reports???

Put an absolute minimum of 50% of your routers in rural areas, as defined by accepted statistics and/or definitions, and then re-present your data please. Otherwise, I hate to say it, but these claims mean nothing.

"Torture numbers and they will tell you anything"


Somerset said...

Pointless doing a survey if the sample size and distribution is not accurate.

What's 'rural'?

GuyJ said...

I rather like ofomc as an acronym and I wonder what the letters might stand for?

GuyJ said...

A statistically significant sample of the Final Third (quarter, fifth, tenth?) would be good...

The three Theoretical Exercises currently being open source supplied by some of the Industry Day attendees to BDUK and in between each other cover areas that are each and all challenging - high hanging fruit

1000 logging routers to capture SamKnows info across each of the 3 localities - do this for a month, for 90 days and onwards.

Instead of theory, invest in capturing actual rural customer broadband performance data and use that to inform future practical improvements to service delivery.

Our friends at DABS will do 1000 of WRT54GL routers for
Delivery (764.000kg) £273.54
Total (inc VAT) £45,945.79

Decent value for Taxpayers Money - less than £50k pays for a dedtailed and valid analysis of the high hanging fruit, in the very garden of Eden - now that would be a huge value add for Rory Stewart's Conference - 10 routers per Parish perhaps?

Cybersavvy UK said...

I've just spent an entire day with stats ppl. I had no idea what a statistically small sample was until today.

Now I do!!!!

And for those who don't know, ofomc is a Nokia N97 shortcut. If it upsets you, then consider how it feels being a volunteer without a decent smart phone in a smart world. Think about it.

Cybersavvy UK said...

@somerset define rural. OK. We have spent nigh on 15 years making sure we are 'in the zone'.

I think it's your turn.....define rural.........

Anonymous said...
This comment has been removed by a blog administrator.
Cyberdoyle said...

I got one of the 5 rural routers. The stats have never been collected because the customer uses one of the ISPs that aren't on Samknows list. We have never been given any stats even though they said we would.
If you reflash these routers with ddwrt they make good mini meshboxes, guess that is what we will do to ours. Will be more useful that stats that mean nothing. Putting the routers in a 10 km circle round ANY exchange would yeild far more valuable data with more accurate average for UK broadband, not cherrypicking the larger ISPS and more tech savvy people.
In this case 'rural' means anyone further than a few Km from an exchange. Final third first means the same. The people who never will get access to decent broadband without a fibre feed.