Meteor is by far the fastest Mobile Broadband on Speedtest.net

dublin geek

I just thought I would share some research that I did when choosing Meteor for Mobile Broadband. I haven't seen this comparison done before on Boards - but I think it is useful.

Before anyone starts lecturing me on the deficiencies of mobile "midband" - I am well aware of these. I already have Eircom 7.6M DSL at home and an Eircom fibre connection at work, so for me Mobile Broadband is purely for use out and about, when away from fixed line Broadband & Eircom Wi-Fi hotspots.

There are a lot of posts on Boards about individual experiences on Mobile Broadband - but individual experiences mean nothing because 3G is so hit-and-miss. In order to determine which is best, what is required is a comparison using a high number of samples, in order to average out the local variations. Speedtest must have this information, but they don't appear to publish it on their site.

However, when you do a speedtest, in addition to showing your individual result, it also shows a bar indicating your "ISP average". It's just a graphic - not a number, but by comparing the length of this bar with your individual measurement, you can get a reasonable estimate of your ISP. You Tube has reasonably recent videos of speedtests done on Vodafone, O2 and 3, so I was able to get the averages for each of these ISPs without having to buy 3 dongles. I decided to try a Meteor dongle as I couldn't find a speedtest video of it. I now have a full set of ISP averages for all the mobile networks. I didn't get a video or try an Eircom Mobile dongle, but I expect the performace is the same as for Meteor, since Eircom own Meteor.

Based on this methodology, the ranking for the average download speeds is as follows,

1. Meteor - 3.3Mbps
2. O2 - 2.2Mbps
3. Vodafone - 1.7Mbps
4. 3 - 1.5Mbps

As you can see - Meteor is the fastest, by a significant margin. It may also be reasonable to assume, since I have never seen the full 7Mbps, that to maintain an average of 3.3Mbps, Meteor must rarely drop below 1Mbps. This is certainly my experience so far.

Another thing I liked about Meteor is the option for a day-pass - this suits my occasional usage pattern.

I think this ranking method is worthwhile, but I would welcome other views. Clearly, it must come with a health warning that averages don't necessarily mean that the perfomance in a specific location will mirror the average ranging.

For reference, the Speedtests on YouTube are as follows,
Vodafone - http://www.youtube.com/watch?v=SzIJxbxBp_8
3 - http://www.youtube.com/watch?v=AQopVYwjjv8
O2 - http://www.youtube.com/watch?v=1vX_JU_IFxE

jor el

Since Meteor has very little coverage, and very few customers, they're obviously not feeling the pinch of contention just yet. That will change, guaranteed.

PogMoThoin

Ssssshhhhhh, don't tell any of Your neighbours Your Meteor dongle works or they'll all sign up. This is what happens. 3 had great speeds, the word got out and within 6 weeks there was huge problems.

Firetrap

And you joined just to tell us this Contention is the killer of all mobile internet.

Kensington

Member's 1st post.
Plugs Meteor broadband as high speed.
Plugs Eircom 7.6MB broadband at home and Eircom fibre line into business.
Plugs Eircom WiFi hotspots for Eircom broadband customers.

So tell us, what department do you work for in the Eircom group?? :rolleyes:

As mentioned before - since Meteor's 3G coverage is so sparse they're not going to have any great amount of customers on it. You're not going to get much load on a 3G network which feck all subscribers on it! Give it a few months - as customers increase, speeds will decrease...

Sponge Bob

I will predict 3 things

1. eircom will pick up another 10 or 20 3g bb customers in Dublin or Cork ( where they actually run 3g services ) .

2. the service will be as sh1te as it is anywhere else at 8-12pm

3. The shill will disappear ...because they gave him/her/it a job writing scripts in Bangalore and the food does not agree with him/her/it .

This must be the most pathetic social engineering effort that eircom ever came up with, obviously a junior marketing person.

Leave boards to your very excellent and competent engineers in future . They know how to handle us

Sponge Bob

Forgot to mention .

http://irelandoffline.org/2009/10/international-broadband-quality-survey-blasts-dublin/

IrelandOffline today received their copy of the “Oxford University Said Business School Global Broadband Quality Study” which is issued annually.
Commenting on the results of the study Eamonn Wallace of IrelandOffline said
“We are not quite sure whether this is Salvador in Brazil or Salvador in El Salvador but this city is now ranked as the 86th City in the world
for the quality of its Broadband Infrastructure and is one place ahead of Dublin in 87th place for the Quality of its Broadband” (1)
Cities ranked ahead of Dublin include Ankara, Ljubljana, Bratislava, Talinn, Budapest, Sofia, Vilnius, Riga, Prague, Athens and many Chinese cities. (2)(3)

Most of these cities are former Soviet states and not noted for their infrastructural achievements. Athens has always been many places below Dublin for
the quality of its telecommunications infrastructure.
Wallace also added:
“The only notable European capital cities that score worse than Dublin are Warsaw and Rome and not by very much may I add.
I also note Soweto in South Africa is currently in 102nd place with a score of 25 and it scores nearer the Dublin score of 28 than Dublin scores against Lisbon with a score of 38.
We recommend that Minister Ryan should visit a Conference in Soweto next year rather than a Conference in Seoul ( score 65 ) as he did last year.

At least his phone might work in Soweto.”
The survey shows Ireland frozen in broadband mediocrity. With the lowest Telecommunications Investment in the developed world (as a % of GDP) we are therefore guaranteed to remain in mediocrity.
The low Broadband Quality outside our major cities is also noteworthy.(3) 60% of our population currently live outside our major cities and while Ireland scores slightly ahead of Ukraine we are behind Greece, Spain, Portugal, Poland, Slovakia and Hungary right now.
Furthermore, most of those countries have large scale fibre installation plans or installation projects in train where we have no concrete plan.
All of these future network plans are glibly entrusted to his “liberalised market” by Minister Ryan and his dysfunctional department.

Sources Referenced Above..

1.http://www.sbs.ox.ac.uk/NR/rdonlyres/0353B17F-CAF7-4169-A294-9DAB696398A2/0/BroadbandQualityStudy2009Appendix.pdf Page 11
2.http://www.sbs.ox.ac.uk/NR/rdonlyres/99AAE7AC-C951-42CB-8F62-D43F51CEDE87/0/BroadbandQualityStudy2009PressPresentationfinal.pdf page 19
3.http://www.sbs.ox.ac.uk/NR/rdonlyres/99AAE7AC-C951-42CB-8F62-D43F51CEDE87/0/BroadbandQualityStudy2009PressPresentationfinal.pdf page 20
4.http://www.sbs.ox.ac.uk/news/media/Press+Releases/Cisco+BQS+2009.htm Link to Said Business School Press Release and Documents including 2008 Data

dublin geek

Kensington wrote: »

Member's 1st post.
Plugs Meteor broadband as high speed.
Plugs Eircom 7.6MB broadband at home and Eircom fibre line into business.
Plugs Eircom WiFi hotspots for Eircom broadband customers.

So tell us, what department do you work for in the Eircom group?? :rolleyes:

As mentioned before - since Meteor's 3G coverage is so sparse they're not going to have any great amount of customers on it. You're not going to get much load on a 3G network which feck all subscribers on it! Give it a few months - as customers increase, speeds will decrease...

Thanks for the warm welcome to Boards, guys!

Having scoured the Internet looking for an objective assessment of Irish Mobile Broadband average performance - I failed to find any. I simply came up with a methodology for comparison using existing measurements, which can be verified independently by anyone who chooses.

Nobody seems to have published this before in one place. I thought it might be appreciated by others wanting to make a similar decision - but clearly some people want to maintain their existing prejudices and don't want new evidence.

If anyone wants to challenge my calculations or methodology, please feel free - but I have no interest in continuing a discussion based on prejudices or speculation as to what might happen in the future. I just presented the facts as they are now. I also gave a huge health warning, because I agree with most of the posts on Boards that mobile "Midband" is hit-and-miss and isn't a substitute for fixed Broadband. Perhaps Meteor's performance will deteriorate with more users - I didn't say that it wouldnt. However, all the networks will get more users, and hence the ranking may stay the same?

By the way, I have nothing to do with Eircom or Meteor, and had no part in the decision to choose Eircom at work. Just stating the context, that I have fixed Broadband alternatives and don't depend on Mobile, lest anyone read this post thinking of choosing Midband as a permanent solution. Clearly it's a crime to use Eircom or Meteor if you are going to post on Boards. I must remember this in future.

Sponge Bob

dublin geek wrote: »

Thanks for the warm welcome to Boards, guys!

Having scoured the Internet looking for an objective assessment of Irish Mobile Broadband average performance - I failed to find any. I simply came up with a methodology for comparison using existing measurements, which can be verified independently by anyone who chooses.

Nobody seems to have published this before in one place. I thought it might be appreciated by others wanting to make a similar decision - but clearly some people want to maintain their existing prejudices and don't want new evidence.

I agree with most of the posts on Boards that mobile "Midband" is hit-and-miss and isn't a substitute for fixed Broadband.

That is all true save a few qualifiers

1. performance is by sector by cell, the difference between good and bad speeds is a few users and their position relative to a particular mast . Great speeds require a near empty sector apart from yourself.

2. If you tell everyone that eircom is good then everyone will abandon 3 and go to eircom which will make eircom crap. Consequently the best way to guarantee your good eircom speeds is NOT TO TELL ANYONE ABOUT IT !!!

3. Average performance is crap, there are around 60 midband users for every sector in the country ....averaged. Some more and some less.

That would be based on around 6000-7000 sectors for around 400,000 midband users . Eircom were around 10 users per sector average at end july 2009 .

Now shhhhhh! Willya :cool:

Kensington

dublin geek wrote: »

Clearly it's a crime to use Eircom or Meteor if you are going to post on Boards. I must remember this in future.

No, I just find it very, very, very odd that you specifically state that it's Eircom products you use at home and at work. Most people would just simply state that they have high speed broadband at home and a fibre line in work. But you go to very specific pains to list all of the Eircom group internet products aswell as your nice little plug of Meteor broadband.

stepbar

Here's my specific dilemma:-

I have Digiweb Metro broadband and pay €39.95 for the privilage. I also have a data package on my mobile (with Meteor) which costs €10 a month.

I'm now finding that I am increasing needing to use broadband on the move, however the mobile it not sufficent for the work I need to do. I find the data package useful for checking gmail, news etc but thats about all.

On average I use approx 10gb a month (Metro) and TBH I just check email, Boards, Youtube and a few other things.

I'm thinking of just buying a new laptop and getting rid of both packages I have. I would then obtain a midband solution and work from there. The saving would be c€30eur a month which would be used to pay for a laptop over a year.

Now, in the past I've wireless modems from Irish Broadband and Clearwire. To put it mildly both solutions were sh1te. The next solution I got is the one I have now - Digiweb Metro and I've had no problem with it so far. However, I'm paying €50 a month for broadband that doesn't really suit me at this point (I could live but life would be handier with a laptop and reliable broadband "on the move"*).

I suppose what I'm asking is, am I setting myself up for trouble (based on my previous experience with IBB/Clearwire)? The last thing I want it having to hand back the midband solution and having to get Digiweb back in.

Any comments / observations welcome.

*On the move means anywhere down the N4/M4 between Dublin and Sligo.

Sponge Bob

Meteor midband will work in Mullingar and Longford and Sligo ...and of course Dublin . Rural coverage is cack BUT there is EDGE in some spots .

Eircom only really give 5gb a month and crucify you thereafter so you will need 2 packages to get 10gb . Read the pricing carefully .

Amalgam

I changed over from 3 this evening after a week of 'Error 619' every couple of hours. The E180 is 5 cent short of 50 euro in Xtra-Vision, cheaper if bought direct from Meteor online. A Huawei E180. Works fine with Ubuntu and Mint too.

The last straw with 3 was being told in my account I had 4 days and 4GB of data allowance left.. yet being cut off.

Here's the current Speedtest.



I dont have to reconnect every ten minutes, south Dublin, in a low coverage area.

Amalgam

During the day, lunchtime, same location as above..



Still, relatively speaking, more responsive than 3.. which just goes to show how crap things are overall. No need to reconnect every so often.

Amalgam

Another odd quirk of using Meteor, PG2 (PeerGuardian) blocks 'FiberConnexion (Media Defender)' every 5 minutes or so, 3 attemps, on the clock. I've just started using the dongle and don't torrent using Midband.

Pyro Boy

Amalgam wrote: »

Another odd quirk of using Meteor, PG2 (PeerGuardian) blocks 'FiberConnexion (Media Defender)' every 5 minutes or so, 3 attemps, on the clock. I've just started using the dongle and don't torrent using Midband.

OT but don't use PeerGuardian it doesn't work if anything it could make the situation worse for torrents.

foggy_lad

any time i have been using meteor broadband around dublin it is totally useless even in the city centre! yesterday evening between 2pm and 10pm i was travelling between naas and dublin and there was no broadband, the 3g network was there but not able to stream a youtube video.

this seems to be very much oversubscribed in the dublin area now probably due to people switching after o2 started blocking access to certain sites?

tara73

I agree with kensington, the post from dublin geek sounds completely like an advertisement.

I'm using this meteor to go, subscribed 6 month ago and for the last 5 month it was fine, speed was fast and never problems with the connection.

since the last month or almost two it changed, it's completely crap and kind of unusable at all.
most of the times it takes ages to open pages, literally I can go and make a cup of tea:rolleyes:....additionally the connection cuts off several times a day...I live in Dublin City Center, so no rural place!!!!!
I tried to talk to the customer service, their first sentence when I explained the issue was, that meteor has NO problems with their network. when I tried to discuss things further he hang up!!! this is unbelievable. I also had to run after my deposit in a ridicilous way.
I'll cancel this thing with meteor, it's not worth the hassle...

Sponge Bob

tara73 wrote: »

I agree with kensington, the post from dublin geek sounds completely like an advertisement.

I'm using this meteor to go, subscribed 6 month ago and for the last 5 month it was fine, speed was fast and never problems with the connection.

since the last month or almost two it changed, it's completely crap and kind of unusable at all.

Well you have explained the situation quite nicely Tara .

It ONLY took eircom 5 months to max out your city centre cell and then their network became as bad as any other 3g providers . The service launched in March and was maxed by August.

It took three Ireland about a week to max out their grossly inadequate Dublin network two years before then.

The 3G operators in Dublin will not have any new spectrum until 2014 when MMDS is turned off but they could rollout microcells and picocells in certain areas to improve the coverage .

tara73

Hi Sponge Bob / All,

I tried to figure out how this "broadband to go" (which I learned already is no broadband at all) is working....as I understand now, the more users there are, the slower it gets??
but how is that possible? it works obviously with a signal, but as far as I know the normal wireless broadband works with a signal as well and is not affected by the amount of users...Is that right??
And why does this happen to all "broadband to go" offers after a couple of month??
Sorry to ask this questions, but I'm not a computer / broadband professional but really interested to understand this whole strange thing.....
thanks for answers!
tara

jor el

tara73 wrote: »

but how is that possible? it works obviously with a signal, but as far as I know the normal wireless broadband works with a signal as well and is not affected by the amount of users...Is that right??

This image should help to explain how it works. A cell sector has a limited bandwidth available, and it's divided among the users, with priority given to those closest to the mast.


20 users will bring a cell sector to a halt, and in an urban area, that limit could be reached very easily.

Fixed wireless does suffer from contention too, but it will depend on the technology used and the bandwidth available.

More info on www.radioway.info

rebel.ranter

Actually I think the majority of issues being experienced on 3G networks are not related to cell breathing at all. It is down to capacity issues, capacity issues which occur no matter how close you are to the site, which is different to cell breathing. Don't believe everything you read.

jor el

Cell breathing will effect those at the extremities. Capacity will effect everyone. On a 3.6Mbps sector, 20 users would get dial up speeds, due to capacity.

Sponge Bob

Capacity issues are part of it , some 3g cells ( not meteor but 'somebody' ) had bog standard 2mbit leased lines for backhaul even as late as 2008 . That would be 3 x 3.6mbit sectors scrunched down a single 2mbit pipe .

However the pattern on the ground where the service disappears completely and in a disorderly manner is consistent with cell breathing , were it capacity then everybody would get 'their share' of the 2mbits ....like 20k each for example

Instead they get nowt

rebel.ranter

I still think the effects of cell breathing are overstated. Cell breathing is more of an interference phenomenon. See here:

Cell Breathing

If all the codes are used up then there is no possible code to be allocated to a user. With cell breathing the user's ability to connect is not related to the fact that a code cannot be allocated, it is due to noise, the noise floor is too high for a distant user to connect.
Books are great for the theory on this stuff & there are many out there that have done extensive reading on this subject & could probably recite lots of stuff but as with many things the theory does not always match up with reality.
You are correct to point out also that backhaul is also another limiting factor. In fact I believe that in the long term this will be more of an issue than on-site capacity or cell breathing. There is plenty of scope to upgrade node B capacity at the minute with Tx capacity upgrades, additional channel elements, 2nd & 3rd carriers, etc.
At the moment operators in Ireland are using microwave with multiple E1s or leased lines as backhaul. Really there needs to be fibre to all these sites to match the capabilites of a 3G node B.
Running fibre is an expensive business, this will be the main limitation (the costs that is). I think this is an issue for the future of all broadband & midband services fixed or moble. Who pays for it?

Sponge Bob

The fibre issue is an immediate one ...given the imminence of LTE .

Thankfully these capacity issues are not just an Irish thing. Our peers in the Phillipines have the same problems and do not have a widespread copper network to fall back upon unlike we do.

http://www.telegeography.com/cu/article.php?article_id=30668&email=html

Filipino telecoms operator Globe Telecom’s wireless broadband network has been ‘temporarily overwhelmed’ by a massive surge in subscribers following a price war among local operators. The Manila Times writes that Globe slashed the price of its ‘Globe Broadband Tattoo’ offer from PHP1,895 (USD40.24) to PHP895 in response to rival offers from Smart’s ‘Smart Bro Prepaid’ kit (now PHP995) and Sun Cellular’s ‘Easy Broadband Plan (PHP799), resulting in significant subscriber uptake that crippled the service.

At least they were honest about it, unlike 3 were in 2008 .

Flamed Diving

dublin geek wrote: »

I just thought I would share some research that I did when choosing Meteor for Mobile Broadband. I haven't seen this comparison done before on Boards - but I think it is useful.

Before anyone starts lecturing me on the deficiencies of mobile "midband" - I am well aware of these. I already have Eircom 7.6M DSL at home and an Eircom fibre connection at work, so for me Mobile Broadband is purely for use out and about, when away from fixed line Broadband & Eircom Wi-Fi hotspots.

There are a lot of posts on Boards about individual experiences on Mobile Broadband - but individual experiences mean nothing because 3G is so hit-and-miss. In order to determine which is best, what is required is a comparison using a high number of samples, in order to average out the local variations. Speedtest must have this information, but they don't appear to publish it on their site.

However, when you do a speedtest, in addition to showing your individual result, it also shows a bar indicating your "ISP average". It's just a graphic - not a number, but by comparing the length of this bar with your individual measurement, you can get a reasonable estimate of your ISP. You Tube has reasonably recent videos of speedtests done on Vodafone, O2 and 3, so I was able to get the averages for each of these ISPs without having to buy 3 dongles. I decided to try a Meteor dongle as I couldn't find a speedtest video of it. I now have a full set of ISP averages for all the mobile networks. I didn't get a video or try an Eircom Mobile dongle, but I expect the performace is the same as for Meteor, since Eircom own Meteor.

Based on this methodology, the ranking for the average download speeds is as follows,

1. Meteor - 3.3Mbps
2. O2 - 2.2Mbps
3. Vodafone - 1.7Mbps
4. 3 - 1.5Mbps

As you can see - Meteor is the fastest, by a significant margin. It may also be reasonable to assume, since I have never seen the full 7Mbps, that to maintain an average of 3.3Mbps, Meteor must rarely drop below 1Mbps. This is certainly my experience so far.

Another thing I liked about Meteor is the option for a day-pass - this suits my occasional usage pattern.

I think this ranking method is worthwhile, but I would welcome other views. Clearly, it must come with a health warning that averages don't necessarily mean that the perfomance in a specific location will mirror the average ranging.

For reference, the Speedtests on YouTube are as follows,
Vodafone - http://www.youtube.com/watch?v=SzIJxbxBp_8
3 - http://www.youtube.com/watch?v=AQopVYwjjv8
O2 - http://www.youtube.com/watch?v=1vX_JU_IFxE

To anyone with even a leaving certificate understanding of descritive statistics, will know that reporting averages is meaningless without measures of distribution.

Standard deviation? Variance?

tara73

whatever the issue is, I strongly believe it would be solvable very easyly for this companies who offer the service.
And every child can think that far: the more customers = the more money for the company = duty to the company to buy more capacities or whatever to fix the problem and ensure the service for the customers.
But it's all a well planned rip off, it's 100% that people from meteor and all this crap are protected from the government or even sit in the government.
People fly to the planet mars, but it's not possible to deliver a working internet connection?????

rebel.ranter

Sponge Bob wrote: »

The fibre issue is an immediate one ...given the imminence of LTE .

I wouldn't be holding my breath! The operators are only getting €20 per month from a suscriber at best for a HSDPA service, LTE involves significant investment both in backhaul & Access network equipment. They would need to get more than €20 per month for LTE to be worth while rolling out.

There would need to be a compelling reason for the operators to fast track the rollout of LTE (i.e. guaranteed income).

jor el

tara73 wrote: »

whatever the issue is, I strongly believe it would be solvable very easyly for this companies who offer the service.

Not really. There is a technological limitation with 3G/HSPA that simply cannot be overcome. The number of connections in any sector has a physical limit.

tara73 wrote: »

And every child can think that far: the more customers = the more money for the company = duty to the company to buy more capacities or whatever to fix the problem and ensure the service for the customers.

Capacity can be bought, but that will only go so far. The "whatever" is far more complicated.

tara73 wrote: »

But it's all a well planned rip off,

It's well planned miss-selling of a service that is not supposed to be, and was never designed to be, a fixed line or fixed wireless broadband substitute, or alternative.

tara73 wrote: »

it's 100% that people from meteor and all this crap are protected from the government or even sit in the government.

You've lost me there. Are you saying that Meteor form part of our government?

tara73 wrote: »

People fly to the planet mars, but it's not possible to deliver a working internet connection?????

Again, you've lost me. Even if people had gone to Mars, which they haven't, what's that got to do with the price of spuds?

jay93

where did you get your statistics from ?? for starters vodafone broadband is barely even 0.86 mbps in the day 3.06mbps at night .. also 3 brodband speeds are not 1.5mbps ther 3.10mbs as i have done speed test with them not sure of o2/meteor speeds but i would expect meteor to be the fastest but thats only because they have way less people connecting to the internet than all other 3 providers .. so just give it time and meteor will be the slowest i can guarentee that if their gprs network can barley even handle calls i can just imagine how bad their broadband service will be ..

dublin geek

jay93 wrote: »

where did you get your statistics from ??

If you look at my original post at the start of the thread - I gave the sources for the Vodafone, O2 & 3 Speedtests on YouTube as follows,

Vodafone - http://www.youtube.c/watch?v=SzIJxbxBp_8
3 - http://www.youtube.com/watch?v=AQopVYwjjv8
O2 - http://www.youtube.com/watch?v=1vX_JU_IFxE

I didn't do these tests - just found them on YouTube. I did the Meteor test myself. Comparing the average bars (rough measurement on screen), this gave the following ranking.

1. Meteor - 3.3Mbps
2. O2 - 2.2Mbps
3. Vodafone - 1.7Mbps
4. 3 - 1.5Mbps

The Meteor averages on Speedtest.net are holding steady at 3.3Mbps since my first post, despite the predictions of some people on Boards that they would deteriorate.

If anyone has more recent data for average speeds for the other networks, please post. Meteor were so far in the lead, I would be very surprised if any of the other networks have overtaken them. I emphasised that the summary data refers to average speeds. As everyone agrees on Boards, all midband products vary widely depending on location and time of day. I simply wanted to have a method for identifying the network with the best average speed. Nobody should assume from these results that the average ranking automatically applies to their specific location. I also emphasised that mobile midband is no substitute for fixed line Broadband - I only use mobile when out and about, which is why the average matters most to me.

Flamed Diving

Dublin Geek - you are not providing the full picture. Without providing either the variance or the standard deviation of these results, then what you are asserting is baseless. I am going to assume that you have these figures, and are deliberately concealing them because they would spoil your conclusion that Meteor is the best provider.

Feel free to prove me wrong, kid.





*For those who don't get what I mean, consider two providers: Alpha and Beta. Five (Mon-Fri) random speed (mbps) samples were taken for each provider:

Alpha = {0.9, 1.2, 1, 1.2, 1.1}
Beta = {0.5, 1.5, 2, 0.3, 1.1}

Mean for Alpha = 1.08
Mean for Beta = 1.08

So, looking at the mean only one would assume that there is no difference between both providers, right? What about the variance?

Variance for Alpha: 0.017
Variance for Beta: 0.492

As you can see, the degree to which the service of Beta varies (from the average) from day-to-day is 29 times greater than Alpha.

So, which provider would you pick?

dublin geek

Flamed Diving wrote: »

Dublin Geek - you are not providing the full picture. Without providing either the variance or the standard deviation of these results, then what you are asserting is baseless.

Flamed Diving - I don't have Variance or SD stats for any of the mobile networks. I took my data from Speedtest.net, and they don't appear to publish anything other than averages. They only barely publish averages in the form of a graphical bar, which you have to measure on screen.

I agree that it would be very nice to have variances and standard deviation data also. However, the examples you give have been contrived to prove a point - you have no information to suggest that any of the networks follow these profiles.

Most people on Boards and elsewhere are forming their opinions of the different networks based on marketing nonsense, or anecdotal evidence from a small sample of experiences. Speedtest.net averages thousands of samples. Averages aren't perfect, but far more scientific that what most people are using to choose a network.

The examples you contrive show two groups of data with the same average. In this case, clearly the samples with the narrower variance would be the better network. However, Meteor are so far in the lead on average speed, that the variance profiles would have to be very significantly different in order to shift the ranking of networks based on variances rather than averages. It is statistically possible, but unlikely, given that they are all using similar technologies.

Flamed Diving

It is statistically possible, but unlikely, given that they are all using similar technologies.

Hmm, Dublingeek, you don't seem to know what you are talking about. In fact, you seem to be trying your best to squeeze the square into the round hole. Unless you have data to back this assertion up, I declare it void. Allow me to help you some more.

Sample one of wages (Euro) = {5, 2, 6, 1000, 4}
Sample two of wages (Euro) = {8, 9, 8, 10, 11}

Sample one average wage = €203.4
Sample two average wage = €9.2

Hey, on average, group one is better off!

You see, this is the biggest problem in the mind of the "average" person. All they look at is the mean, they probably don't even know what an SD/VAR is. But to anyone with even an undergraduate (hell, even LC!) understanding of descriptive statistics should know that any data which conceals the full statistics, conceals the full story and should be treated with caution. Now, back to my "contrived" dataset.

Median of sample one = €5
Median of sample two = €9

There, that is a much better representation. Tell me, Dublingeek. Do you even have the median? Or are you basing your entire dogmatic belief on what is a very shaky sum?

Your statistical analysis is no better than what you read in the Sun or Daily Mail. Stop wasting our time.

dublin geek

Flamed Diving wrote: »

Sample one of wages (Euro) = {5, 2, 6, 1000, 4}
Sample two of wages (Euro) = {8, 9, 8, 10, 11}

Sample one average wage = €203.4
Sample two average wage = €9.2

Flamed Diving - this is more severely contrived data to prove your point. While totally hypothetical wages can have a unlimited variance eg 2-1,000, the variance on 3G is limited by the upper limit of the technology. Meteor seem to be using generally 7Mpbs technology, although I did come across greater than 7bps recently in Galway - maybe they are using 14Mbps in some places. Either way, this puts a absolute limit on the extent to which a spurious high reading could be distorting the average in the way suggested in your contrived data sets. If you have data from Mobile Midband performance to prove your point, please publish it. Otherwise, there is no evidence to support your theory that widely diffierent variances are going to change the performance ranking of the 4 Irish mobile operators.

Flamed Diving

*GROAN*

Dublingeek, my "contrived" (this word makes you seem so silly) data was so to illustrate a simple point. Without measures of dispersion, or alternative measures of central tendency then your statistics say F-ALL. They are completely misleading, and the way you throw them around as if they are gospel leads me to be highly suspicious of your motives. I could very easily demonstrate what I did earlier within the limits of midband mbps range and come up with a similar result. In fact, any statistics textbook will tell you the dangers of using only one measure, but sure that doesn't suit your agenda now does it, Dublingeek? Too bad. I have very simply illustrated with my "contrived" (LOL) datasets that a ranking of averages is very likely to be deceptive and should be taken with a HUGE pillar of salt. This goes for ANY ranking of averages. ANY.

Here, go away and educate yourself instead of spreading nonsense.

http://www.dummies.com/store/product/Statistics-For-Dummies.productCd-0764554239.html


To anyone reading, do you think that:

widely diffierent variances are going to change the performance ranking of the 4 Irish mobile operators?

Would you sign up a "widely varying" service, even if its average speed was faster? Do you think that you would still only pay attention to the performance ranking set out by Dublingeek, where only average speed matters? Do you think that this "performance ranking" is somewhat missing the full picture of what defines good performance?

Flamed Diving

I should probably point out at this stage that I don't care which provider is the best. In fact, it may be the case that Meteor would be strengthened further by the inclusion of the dispersion data. However, this is not the point. Barstool statistics being presented as fact must be one of the most irritating things in modern society. Can you imagine Dublingeek presenting his findings at an academic seminar? No measures of dispersion you say? It wouldn't change the measure of performance?

GEEEEEEEEEEEET OUUUUUUUUUUUUUUUUUUUUUUUTTTTTTTTT!!!!!!!!!

dublin geek

Flamed Diving wrote: »

*GROAN*

Dublingeek, my "contrived" (this word makes you seem so silly) data was so to illustrate a simple point.

Flamed Diving - please cut out the emotional attacks on my intelligence, and give us some hard evidence that provides a better means of ranking mobile midband products.

As I have stated, Averages are not perfect, but the best we have...if you have better data, you are welcome to provide it. I disagree that Averages are of no use. Most people, including scientists use Averages to assess a wide variety of data. Variances and Standard deviations don't change the validity of averages, but they provide additional perspectives - nice if you can get them, but very often you have to make decisions without a full statistical analysis.

I could very easily demonstrate what I did earlier within the limits of midband mbps range and come up with a similar result.

Not possible - the contrived data set you used included one reading which is more than 200 times the average of the others! No data set that you obtain from any of the midband networks will show this level of variance from the mean. The average for Meteor is 3.3Mbps - the absolute technical max of the technology is only 14.4Mbps. Hence the maximum possible variance is only a little only 4 times the mean. In practice, most people are probably using slower speed modems, and I have rarely seen more than 5Mbps from Meteor, which suggests that the performance varies over a much narrower range, closer to the mean. We don't know how this compares with the variances on the other networks, but we know that the variances are all limited by similar upper boundaries of the technology.

Flamed Diving - perhaps you can tell us what superior scientific method you used to choose or recommend a mobile midband provider, based on existing evidence?

(1) Compare the marketing hype from each provider?
(2) Pick the provider that you think is the least "silly"
(3) Compare the anecdotal evidence on Boards?
(4) Compare the anecdotal evidence from friends?
(5) Choose the network your existing phone is from?
(6) Other method?

Flamed Diving

dublin geek wrote: »

Flamed Diving - please cut out the emotional attacks on my intelligence, and give us some hard evidence that provides a better means of ranking mobile midband products.

As I have stated, Averages are not perfect, but the best we have...if you have better data, you are welcome to provide it. I disagree that Averages are of no use. Most people, including scientists use Averages to assess a wide variety of data. Variances and Standard deviations don't change the validity of averages, but they provide additional perspectives - nice if you can get them, but very often you have to make decisions without a full statistical analysis.

What. Utter. Nonsense. Speaking on behalf of scientists while displaying a depth of ignorance about the very basics of statistics. You would be thrown out of a lab for the nonsense you have been spouting above. I can't believe that you would even have the cheek to attempt it. Good god, do I really have to say what I'm about to say. If you have the data to calculate an average (wait for it) you have the data to calculate the SD and VAR! And guess what, it does make a huge difference to results, and it does call into question the "ranking" that a simple mean result provides. Bloody hell! I have already explained why this is! Get a clue!

dublin geek wrote: »

Not possible - the contrived data set you used included one reading which is more than 200 times the average of the others! No data set that you obtain from any of the midband networks will show this level of variance from the mean. The average for Meteor is 3.3Mbps - the absolute technical max of the technology is only 14.4Mbps. Hence the maximum possible variance is only a little only 4 times the mean. In practice, most people are probably using slower speed modems, and I have rarely seen more than 5Mbps from Meteor, which suggests that the performance varies over a much narrower range, closer to the mean. We don't know how this compares with the variances on the other networks, but we know that the variances are all limited by similar upper boundaries of the technology.

IT WAS AN ILLUSTRATION!!!!!!!!!!!!!!!!!! Do you even know what that is? good lord, you are insufferable! :rolleyes::rolleyes::rolleyes:

dublin geek wrote: »

Flamed Diving - perhaps you can tell us what superior scientific method you used to choose or recommend a mobile midband provider, based on existing evidence?

If I was to perform a test, I would collect data. Actual data and perform a statistical analysis. A rudimentary version using Excel would suffice for such a banal topic. Standard deviation, max/min, IQR, median, that sort of stuff. Intermediate level statistics. I wouldn't post a Speedtest.net average and then tell everyone that Meteor is definitely the fastest and best, based on this result. However, since the data is unavailable (is it? i have my doubts about that also), I would just pick the best priced.

dublin geek wrote: »

(1) Compare the marketing hype from each provider?

What, like posting simple averages and concealing the important stats? That is certainly the kind of stunt I see marketing types pull all the time. But you wouldn't know anything about that.

dublin geek wrote: »

(2) Pick the provider that you think is the least "silly"

No, pick the least silly analysis.

dublin geek wrote: »

(3) Compare the anecdotal evidence on Boards?

Nope, anecdotal evidence is not terribly reliable, just like averages.

dublin geek wrote: »

(4) Compare the anecdotal evidence from friends?

Nope, anecdotal evidence is not terribly reliable, just like averages.

dublin geek wrote: »

(5) Choose the network your existing phone is from?

Nope, pick the best price.

dublin geek wrote: »

(6) Other method?

Generally pick the best price.

Flamed Diving

Has anyone else noticed the way that Dublingeek has never posted about anything but how great Meteor are?

Hmmmmmm...

dublin geek

Flamed Diving wrote: »

If you have the data to calculate an average (wait for it) you have the data to calculate the SD and VAR!

You are clearly not reading my earlier posts - I obtained the averages from Speedtest.net. I never claimed to have the raw data to calculate the averages - hence I don't have the raw data to calculate SD & VAR. Speedtest.net should have this data, but they don't appear to publish it.

However, since the data is unavailable (is it? i have my doubts about that also), I would just pick the best priced.

This is completely unscientific nonsense.

Price has nothing to do with performance, and if it did, a lower price would probably be more likely to suggest a lower performance.

In practice, price seems to have more to do with the power of the Brand, so big companies like Vodafone, 3 and O2 can probably command a higher price, even if their performance is poorer.

Average speed has to be a better proxy for performance than Price.

Flamed Diving

dublin geek wrote: »

You are clearly not reading my earlier posts - I obtained the averages from Speedtest.net. I never claimed to have the raw data to calculate the averages - hence I don't have the raw data to calculate SD & VAR. Speedtest.net should have this data, but they don't appear to publish it.

This is completely unscientific nonsense.

Price has nothing to do with performance, and if it did, a lower price would probably be more likely to suggest a lower performance.

In practice, price seems to have more to do with the power of the Brand, so big companies like Vodafone, 3 and O2 can probably command a higher price, even if their performance is poorer.

Average speed has to be a better proxy for performance than Price.

Ha. Unscientific nonsense. Do you know what is truely unscientific my Meteor-loving friend? It is when you see the end product of a 3rd parties results and then go around treating them as fact and publishing them in public. For all you know, those results were, ahem, "contrived" :pac:. This is getting boring, geeky. You are boring me because you cannot say anything useful or interesting. You cannot defend your position and all you can rely on is silly rhetoric or attempting to fumble away with the childrens illustrations I gave to help you understand. You obviously work for Meteor. Either that or your only interest in life is Midband services. :rolleyes:

My work here is done... for now.

Flamed Diving

Speedtest.net should have this data, but they don't appear to publish it.

http://www.speedtest.net/datainquiry.php

It took me about two seconds. Not free though.

dublin geek

Flamed Diving wrote: »

You cannot defend your position and all you can rely on is silly rhetoric ...

I have been well prepared to defend this attack on my analysis. Your attack has descended into personal abuse rather than any attempt to provide a better methodology. If someone has a better analysis, I will be quite happy to back down. Regarding me promoting Meteor midband only - nonsense - I have at all times pointed out the weakness of all midband products.

Midband should not be relied upon as a substitute for real Broadband. However, I still it find it extremely useful to have a midband solution when I am away from home/work, for both laptop and Smartphone purposes. I wanted to make sure that I bought the best product for this purpose, and I found the average speed methodology the best means of arriving at a decision. It worked for me, and 90% of the time I am happy with the outcome. There are times - typically weekday evenings in some locations, where Meteor midband lets me down - just like all Midband solutions.

Flamed Diving

dublin geek wrote: »

I have been well prepared to defend this attack on my analysis. Your attack has descended into personal abuse rather than any attempt to provide a better methodology. If someone has a better analysis, I will be quite happy to back down. Regarding me promoting Meteor midband only - nonsense - I have at all times pointed out the weakness of all midband products.

Midband should not be relied upon as a substitute for real Broadband. However, I still it find it extremely useful to have a midband solution when I am away from home/work, for both laptop and Smartphone purposes. I wanted to make sure that I bought the best product for this purpose, and I found the average speed methodology the best means of arriving at a decision. It worked for me, and 90% of the time I am happy with the outcome. There are times - typically weekday evenings in some locations, where Meteor midband lets me down - just like all Midband solutions.

If I made a personal attack, feel free to report the post because I did no such thing. Secondly, congrats on your first post not telling people how great Meteor is. There are dozens of other forums that may interest you.

Go explore.

jay93

hmmm maybe he just wanted to point out that meteor are the fastest at the moment but will slow soon as more people sign up ..im with 3 since yesterday and have to admit they are the best i have used for so many people signed upto it (dont work for 3 as im only 16 haha so dont start )) :D:D:D

Flamed Diving

jay93 wrote: »

hmmm maybe he just wanted to point out that meteor are the fastest at the moment but will slow soon as more people sign up ..im with 3 since yesterday and have to admit they are the best i have used for so many people signed upto it (dont work for 3 as im only 16 haha so dont start )) :D:D:D

I doubt it. All his posts have something to do with Meteor. Why on earth would you sign up to boards and only talk (positively) about one company?

Well, there really is only one reason.

pablomakaveli

I've had both 3 and now Meteor broadband. The speeds i got with 3 when i first got the modem (around 1Mbps)was good but it got progressively worse. I changed over to Meteor and i've gotten double the highest speeds i got with 3 (around 2.5Mbps). I know this will probably change over time with more people signing on but i reckon if you're in the coverage area Meteor is the better choice.

jay93

yeah suppose so !! :p:p you cant really say which brodband provider is the best speeds vary in alot of places so just depends where you are really where i am must not have alot of 3 customers i was with vodafone and the speeds where no good tryed o2 ther ok signal is ****ty so 3 is the best option for me where as other areas the other providers are better in differant places :rolleyes::rolleyes:

gar318

hi just to let you know stay well clear of bill pay internet from meteor they are <snip> im currently signed up with them for 17 euro a month they rang today and told me i owe them 237e for last month haha what a joke i download nothing only browse online they can sing for there money there stick is already in the post on its way back to meteor oh and they also want another 152e for me cancelling my direct debit as they couldnt get access to the 237e they wanted f**kers:mad::mad::mad::mad::mad::mad::mad: