Original URL: https://www.theregister.com/2014/08/23/the_register_monopoly_pubcrawl_mobile_map_vodafone_victor_in_london/

The Register Monopoly Pubcrawl Mobile Map: VODAFONE VICTOR in LONDON

Unless you just want the best 4G: Then EE

By Simon Rockman

Posted in Networks, 23rd August 2014 11:00 GMT

Special Report Get a load of our fabtabulous interactive mobile Monopoly map - now complete with 3G/4G data results - right here

If you think “data” is what you do with your girlfriend then move on, probably to another website: there is nothing to see here. But if you think mobile data in London is interesting you need to read this.

Jasper and Simon play monopoly

Jasper and Simon wonder if network planners have a better map

In the last week of May and the first week of June, mobile survey coverage company GWS with a little bit of help from Vultures Jasper Hamill and yours truly trudged around twenty eight locations in London, taking in Old Kent Road and Whitechapel Road through to Park Lane and Mayfair which - as any Brit who’s suffered a family Christmas knows - are the squares on the London version of the Monopoly board.

The precedent for this was a similar test I worked on twenty years ago at What Mobile magazine where we tested analogue and 2G voice. Things have moved on and we wanted to see how the networks compared. We looked at 2G and 3G voice, and now here are the results for 3G and 4G data.

Stress and nervous tension are now serious social problems in all parts of the galaxy and it is in order that this situation should not be in any way exacerbated that the following facts will now be revealed without further ado: in broad outline the best 3G was from O2 and the best 4G was from EE. But Vodafone was second best at both 3G and 4G, and that in combination with being best at voice - in our view - makes Voda the overall best mobile network in London at the moment. But it should be noted that we originally set out on this quest to examine 4G data coverage specifically, and if that's your priority then EE is the winner.

Our tests, using state of the art SwissQual Freerider II backpacks, gave a tremendous level of insight into what went on with the networks. In particular by being locked to 3G and 4G respectively we could measure not just the four brands but the eight networks. The intention was not to focus on user experience, but rather the performance of each network technology (3G, 4G) deployed by the wireless operators. Hence for each of the four operators a test device was solely locked on 3G data network, while the other was solely locked onto 4G data network.

It’s important to understand how relevant 3G is. Only about 10% of the UK mobile base has a 4G enabled subscription – although more may have 4G capable phones. National coverage of 4G is rolling out apace but it’s still early days. So with only a few of the people using 4G some of the time, 3G is for most of us where it is at.

Of course this is The Register, the most tech-savvy readership on the web, so 4G matters more here than the UK numbers might suggest. Also we were looking at London so 4G signal is going to be better than in most places.

Three had the poorest 4G network coverage overall. At Trafalgar Square the Three 4G network is practically non-existent giving 8% reliability. Ironically when What Mobile did its test twenty years ago we uncovered a major Vodafone hole in the same spot. At Coventry Street and Northumberland Avenue Three only had 57% reliability and at another six locations reliability less than 90%.

O2’s 4G network struggles in Mayfair providing 63% (whereas its 3G network had nearly 100%). At Vine Street, Bow Street and Trafalgar Square reliability for O2 4G was less than 90%.

For the rankings we measured the networks on upload speed, download speed and task success rate to measure reliability.

For 4G, EE stood out top in all categories. Vodafone came joint top on reliability second on speed. O2 was third on speed and reliability and it was a straight fourth place in every category for Three 4G.

On 3G things were very different. Vodafone and O2 shared the top spot for reliability while O2 came top for speed and the others tied for second position. This put O2 top with Vodafone a whisker behind, Three third and EE fourth.

EE’s guilty secret

Looking at the numbers in detail it seems clear that – in the sites we looked at, at least – EE is sacrificing its 3G service to promote its 4G one. No other coverage survey has unmasked this as they usually can’t tell which technology is being measured.

To explain at how we arrived at these results we looked at a variety of criteria. For 3G these included the signal strength measured as the Received Signal Code Power, where - across the whole test - EE and Vodafone had the weakest signals.

The monopoly box

We looked at which 3G mode the data transfer was using: DC-HSPA, HSPA+, HSPA or UMTS. These are the different generations of 3G. When the networks launched ten years ago or so they were all UMTS. Then the faster version HSPA (High Speed Packet Access) came along and that evolved into HSPA+ and DC-HSPA (Dual Carrier). All four networks use a mix of all four modes, so one might think the network with the most DC-HSPA and least UMTS would have the best 3G performance: but it’s not that straightforward. While EE had the most DC-HSPA mode (60%), it also had the most UMTS mode at 22% versus O2 with only 1%! Vodafone attained the fastest download speeds, attributed to 43% DC-HSPA, 40% HSPA+.

We also looked at the modulation and at the number of code channels assigned where on average Vodafone had more than 10 channels while the others had less than 8.

For all the scenarios we looked at reliability and all were pretty good with Three at more than 97 percent and O2 and Vodafone at over 98 percent. While EE’s 90 per cent looks poor it must be remembered that we’d locked the phone to 3G and if it had been allowed to drop down to 2G with EDGE the reliability would have been better, albeit at a much lower speed.

Of course the testing reflects usage as well as the capability of the network. Vodafone’s 3G network had highest download throughputs and slowest upload throughput. Over on the 4G side of things EE, O2 and Vodafone pretty level pegged for signal strength, Three had worst signal strength with an average of -104dBm, only 33% of the time attaining a signal of greater than -98dBm. With the other operators -98dBm was observed nearly 70% of the time, but as you might expect the reliability wasn’t as good as with the old technology where packet for packet Three was only 87 percent reliable, O2 took second place at almost 95 percent while EE and Vodafone tied top for reliability with over 97 percent.

Where EE, predictably perhaps, kicked things out of the ballpark was speed - this is thanks to its “double speed” 4G using more spectrum. EE’s 4G speeds were over four times that of its 3G speeds and 30 to 40 percent faster than its 4G rivals. A download network stress test concluded that EE was able to demonstrate an average of 22 Mbps over the whole network and we saw an amazing spike of 106 Mbps during a download in Bow Street. EE also clocked in the fastest complete download with 61.5Mbps in Whitehall.

For all the networks the 3G signal was at 2100MHz with Vodafone having some additional 900MHz spectrum. On 4G Three was using 1800MHz, EE was split between 1800MHz and 2600MHz while O2 and Vodafone were on 800MHz. When we break things down using the Monopoly colour groups, looking at 4G reliability Three came top in Brown, Electric, Green and Orange and was sole winner in Jail. EE won in posh Purple and shared first place in Pink, Electric, Red, Water Works and the stations. O2 was jointly first everywhere else bar Orange, Red, Purple and the Jail (Pentonville Prison in real life). Vodafone shared top spot everywhere bar Purple, green, water works and Jail. Ultimately the use of the colour groups is only a bit of fun and pretty much the only part of the testing which wasn’t scientific.

You can interrogate our interactive map, compiled using MapHook, right here.

Testing is always controversial and never perfect. But we made no less than 22,915 file transfers during this operation. Measuring multiple parameters on each. The result of this work is a highly professional analysis from both and engineering and statistical perspective, and to show our confidence in this we’ll be taking readers’ questions.

On 29 August, we’ll be holding a live chat with Dr Paul Carter, the boss of GWS, to explain the decisions that were made in testing and to look at some of the interesting lessons learned, such as when mobile networks have better upload than download capabilities and who is sacrificing 3G to promote 4G.

You can email questions ahead of the session to livechatquestions@theregister.co.uk or log on next Friday (29 August) to post them live. ®