I reckon Smarty (Three) and GiffGaff (O2) are both slugged one way or another. That's why I always recommend mainstream paper-shop SIMs or free from the mainstream provider for comparative testing. In the post in Three's case, and possibly others.

I was more surprised that a primary network provider (EE) was limiting the speeds of their standard plans, I can understand cheaper networks doing it, or with the case of Vodafone limiting the speed of their cheaper unlimited plans.

With my short and unscientific tests, Three and SMARTY offer the exact same speeds — it's possible SMARTY traffic may have a lower network priority, but there doesn't seem to be much contention over here, so I likely wouldn't see a difference between the two. One thing I did notice was when SMARTY's adult content filter was enabled, the speeds slowed down significantly from 55+ Mb/s to rarely ever passing 35 Mb/s. Could be coincidental as I didn't do any intricate or long-term testing, it was just an observation.

In reply to a post by thebigdolphin1:

I was more surprised that a primary network provider (EE) was limiting the speeds of their standard plans, I can understand cheaper networks doing it, or with the case of Vodafone limiting the speed of their cheaper unlimited plans.

EE used to on handset contract & SIM only contract they had a "60 Mbps" or "unlimited" and their PAYG services were limited to 30 Mbps. But these went about 2 years ago.

However third parties reselling capacity on the EE network can offer any service level they wish. EE supports quite a few MVNOs, including Plusnet and others. (BT mobile may also be one).

Three used to throttle people on the "One Plan" the original unlimited, so we know they have capability to.

Here's another weird one, just tried comparing Three's 3G vs 4G. Sure, speeds are quite a bit slower (25mb/5mb vs 70mb/30mb) but the latency seems to be substantially better... I thought 4G was supposed to offer a much greater improvement in latency over 3G?

3G: Minimum = 30ms, Maximum = 33ms, Average = 30ms
4G: Minimum = 41ms, Maximum = 60ms, Average = 51ms

I can only assume higher network contention, but the speeds I'm receiving don't seem to back that up (getting 65 down via 4G right now).

In reply to a post by thebigdolphin1:

Here's another weird one, just tried comparing Three's 3G vs 4G. Sure, speeds are quite a bit slower (25mb/5mb vs 70mb/30mb) but the latency seems to be substantially better... I thought 4G was supposed to offer a much greater improvement in latency over 3G?

4G can offer improved latency, and you might have seen that 8 years ago when 4G launched. Now of course 99% of users will be on 4G, and load has increased. Once 3G is turned off the radio capacity can be moved to the 4G signal.

I can only assume higher network contention, but the speeds I'm receiving don't seem to back that up (getting 65 down via 4G right now).

Always more users on 4G than 3G. You can't tell contention as the number of people in the coverage area will change randomly (human beings with phones move), unlike fixed line broadband systems.

Yeah you're probably right, I just assumed that because the 4G speed is fairly stable at its max attainable rate during the middle of a weekday, there'd be very little volatility and much lower usage. I suppose compared to 3G, it's still going to have a lot more devices connected to it.

3G seems to be very susceptible to terrible latency as soon as there's any network traffic though. As soon as I browse a webpage or download a file, it shoots up into the hundreds whereas 4G doesn't seem so bad (presumably just because there's more available bandwidth).

In reply to a post by thebigdolphin1:

Yeah you're probably right, I just assumed that because the 4G speed is fairly stable at its max attainable rate during the middle of a weekday, there'd be very little volatility and much lower usage. I suppose compared to 3G, it's still going to have a lot more devices connected to it.

Depends on the capacity deployed by the network (and what they are licensed to deploy) and the users using it at the time.

4G is the technology LTE (Long Term Evolution) and this is a packet switched technology.

3G seems to be very susceptible to terrible latency as soon as there's any network traffic though. As soon as I browse a webpage or download a file, it shoots up into the hundreds whereas 4G doesn't seem so bad (presumably just because there's more available bandwidth).

3G (UMTS) and 2G (GSM) are circuit switched technologies.

Think of circuit switched as a traditional phone call on a land line. When you make that connection nobody else can use the line until you finish.

Packet switched is where every user splits their conversation into packets, and sends them at the network, and the network gets them to the end where they are reassembled into order.

This is one of the major differences and why 4G keeps working but speeds drop in heavy load conditions, whereas 3G and earlier "saturate" with too many users.

This was frequently seen before 4G launched in the main London railway stations, lots of people with phones showing full signal but not able to get a 'session' on the mast. The "circuits" were all in use.


Your performance queries are interesting, but not really comparable except in speed. The networks have been for the last few years gradually reducing capacity on 3G as the radio spectrum capacity is capable of connecting more users at the same time on 4G. Most networks are planning for the shutdown of 3G in the next 2 to 5 years. (maybe sooner!)