The hugely scientific console reliability poll

[DuckThor Evil]

What? I'm claiming around 5% defect rate, which is actually rather bad, you're the one claiming the truly wild numbers. You've yet to offer ANY support for your argument yet you want me to support mine by breaking down numbers that do not need to? Simple math alone shows that claims of 20% are impossible.

You were the one who said that because MS had a loss of around 300M$ they possibly couldn't have that many repairs and at the same time posting mythical repair cost numbers. I pointed out that the loss of 300M$ was only during one quarter and you replied that it still wouldn't make sense based on "numbers" that apparently are a result of some sort of formula you were calculating in your champer... I would be very surprised if the numbers you were using would be facts instead of hard speculation, as we don't know the cost of repairing a unit and we don't know the percentage of losses that comes from the X360 business as the games and entertainment division includes more contributing factors than just X360. Also we don't how much it costs to make a X360 in the first place, so I'm very curious how you can simplify all this into simple math. Please post this simple math for us to see.

Hey maybe the 20% is not accurate, it was always a toss in figure anyways. Perhaps I should have said that around 20% or more of the X360 owners probably have had a broken unit at some point, as the working replacement unit naturally brings the failure rate number down.

My proove is that everywhere I go I see people discussing about the high failure rate of the console. MS said that 3-5% during the initial damage control stage, and I seem to recall that at some point they admitted the problem is higher than that. I also said that naturally over time the number will grow smaller, because more working units are entering into the market.

The main line of you people who are disregarding this poll for example is that "polls aren't accurate" yet nobody has yet to say anything about why this particular poll that we have here at Beyond3d is showing such a high number of failed units, except I, who said that we probably have lot more early adopters here, and because early units seem to be the most problematic. It probably adds inaccuracy to the poll. Until I hear more convincing stuff than just Polls aren't accurate, I'll certainly won't hold your support to your arguments very strong either.

I think I'm just going to end my part on this particular discussion by saying that I hope you are right.

 

[Cheezdoodles]

Of course for every failed XB360 there's a working XB360. But there are also instances of multiple failed XB360s too. And multiple working XB360's for some owners. As we can't do polls with numbers, only checkboxes, numbers were left out. The result is not a failure rate in %age of units. It is at best, crudely, a measure that when you go into a shop and buy a box, you've such-and-such a chance of it dying over the same time period as the current platform's lifespan.

Um, this is just flat out wrong, you have on perpus made the poll in a way that hides the reliability rate. And reliability rate would be the best measure of your chance of having a box thats going to die or not if you and buy one.

Not to mention, reliability rate would be the most interesting thing to measure, judging by your topic title.

Finally, I have to ask how do you actually interpret these results? It's all very well to highlight the faults and limits of polls, but unless you can present a convincing argument to discredit the entirety of data (such as identifying a poll base as biased in voting habits), the faults of polls are only going to cause a discrepency and not an outright change. .

I dont interpret them in any way, the sample size is to small and the lack of poll options (without even taking in mind that website polls are flawed to begin with, because the userbase your polling isn't an average of the population), isn't going to give any meaningful number.

As stated before, i reckon the failure rates on the X360 are high, but there is nothing that would statistically justify anybody from actually reading the poll results here, and draw conclusions based on it. Unless of course, you wanna draw conclusions only for the B3D forum goers, where the numbers are semi-okay to use (still not great because of lack of polling options), but then again the B3D forum is an to inaccurate sample to make predictions for the real world.

It's all very well to highlight the faults and limits of polls, but unless you can present a convincing argument to discredit the entirety of data (such as identifying a poll base as biased in voting habits), the faults of polls are only going to cause a discrepency and not an outright change. .

Completely wrong, just because the sample size is to small, and the consoles "in the poll" aren't an average of the total userbase in terms of how old and how much they are played, it can mean that the numbers are going to be totally different than your result. Just like my example earlier with the TV-show and the professional poll, totally different results.

 

[Shifty Geezer]

Um, this is just flat out wrong, you have on perpus made the poll in a way that hides the reliability rate. And reliability rate would be the best measure of your chance of having a box thats going to die or not if you and buy one.

Ignoring the fact I say 'at best crude' because the point of the poll wasn't to find a chance to fail rate (how often do I need to repeat this?!), you're saying that 107 people bought XB360s, 42 had those machines die, and yet that's not a measure of chance to have a box break, 42:107? If the reliability rate is far lower, at 5%, how come way more people have had more XB360's die than that? 100 people went into a store to buy a TV. They all left with the same TV. 18 months later 40 of those TVs were no longer working. Yet the failure rate is not 40%. I really don't see the logic in that.

Completely wrong, just because the sample size is to small, and the consoles "in the poll" aren't an average of the total userbase in terms of how old and how much they are played, it can mean that the numbers are going to be totally different than your result. Just like my example earlier with the TV-show and the professional poll, totally different results.

The TV show and poll examples aren't representative here, because they're measuring people's opinions, and not products. Different people have different opinions, and behaviour is likely related to opinions. Thus you'll find that given a question, there'll be a higher likelihood of someone expressing an opinion (voting in a poll) if they feel one way than those who feel the other way. In this case, we have a question about people's experiences with hardware. What reason is there to think that voting habits will vary with hardware bought and used? You keep saying the poll-base can't be accurate, but you haven't really given reasons why other than general complaints about internet polls and referencing other polls.

Although you say B3D users aren't representative of the average XB360 owner base, they have the same hardware as everyone else, so we're still measuring the same machines. Now if there's reason to think B3D owners do things to their machines which they shouldn't, which the masses won't do, then we can say they're not indicative. In which case, as that's your opinion on these results, what are B3Ders doing differently? Why are they driving their machines to break beyond what normal folk do?

As stated before, i reckon the failure rates on the X360 are high, but there is nothing that would statistically justify anybody from actually reading the poll results here, and draw conclusions based on it.

That wasn't what I meant by how you interpret the results here. Instead of using the word 'interpret', I'll say 'explain'. In your opinion XB360's hasn't anything like this failure rate. If so, how do you explain these results specifically? Are most of those XB360 votes lies, fanboys trying to sabotage the poll? Or are the users of these machines over-using them, or doing some other things wrong? I know you've given general points on how polls can be inaccurate, but I can't relate them to what's shown here. I can't myself believe a situation where 50% of votes for one option are fibs. Neither can I get my head around the idea that most people with a busted XB360 broke it through their own actions, yet haven't got the same failure rate due to misuse on other machines. What is it in your opinion about B3D forum goers that means their machines die more often than usual, and which is screing up the numbers here? And why isn't the same inflated failure rate shown on the other machines?

 

[ninzel]

I think there are some problems with taking polls this early in the consoles life cycles.
The longer the consoles are out the more chance there is of defects coming up. The amount of people who own the 360 shouldn't really effect the rate but the fact that it's been out for a full year longer could greatly skew the results right now.
At this point in time the 360 has been out for about 3 times longer than the Wii or PS3. After 6 years the difference will have dropped to 1.2 times longer.It will never be perfectly comparable but the longer you wait I think the better.
I'm not here to defend the 360 or Xbox brand as I've had big defect problems with them while my Sony/Nintendo experience has been trouble free but I still think it should be given more time.
If you are doing a poll that is specifically looking at initial reliability I think we have enough time and data,but title it that way. Beyond that problems could still come up for PS3 and Wii.
Edit: I think even when J.D. Power does quality survey's that call it Inital quality survey.Something like that.

 

[Cheezdoodles]

Ignoring the fact I say 'at best crude' because the point of the poll wasn't to find a chance to fail rate (how often do I need to repeat this?!), you're saying that 107 people bought XB360s, 42 had those machines die, and yet that's not a measure of chance to have a box break, 42:107? If the reliability rate is far lower, at 5%, how come way more people have had more XB360's die than that? 100 people went into a store to buy a TV. They all left with the same TV. 18 months later 40 of those TVs were no longer working. Yet the failure rate is not 40%. I really don't see the logic in that.

If every 1 console out of 5 is broken in the factory, then indeed, the reliability rate is 20%. However, this is assuming that the replacement consoles are working, the real reliability rate for the consumer would be lower. But never mind that, this is not the biggest issue at hand.

Although you say B3D users aren't representative of the average XB360 owner base, they have the same hardware as everyone else, so we're still measuring the same machines. Now if there's reason to think B3D owners do things to their machines which they shouldn't, which the masses won't do, then we can say they're not indicative. In which case, as that's your opinion on these results, what are B3Ders doing differently? Why are they driving their machines to break beyond what normal folk do?

B3D users would be considered on average, Hardcore gamers. The % of people here who went out and bought a launch unit is probably overwhelming compared to the real life average, do you think launch consoles break more often than consoles made today?

There are a lot of other abnormalities that prove this, look at how many Xbox 360 B3D owners that have live. How many have XBL gold in real life? 20% ? How many people here have XBL gold? 80%? Or more judging by the amount of Gamertag posts in the XBL gamertag thread.

Further, going by the amount of launch\ early consoles, its a pretty good assumption to say that B3D people, have used their console more than the average X360 owner, correct?

There are many indications that the B3D userbase, is not at all a good userbase for measuring console averages. Amount of launch\early generation consoles alone in this userbase would be sufficient proof of that.

In your opinion XB360's hasn't anything like this failure rate.If so, how do you explain these results specifically? Are most of those XB360 votes lies, fanboys trying to sabotage the poll? Or are the users of these machines over-using them, or doing some other things wrong? I know you've given general points on how polls can be inaccurate, but I can't relate them to what's shown here. I can't myself believe a situation where 50% of votes for one option are fibs. Neither can I get my head around the idea that most people with a busted XB360 broke it through their own actions, yet haven't got the same failure rate due to misuse on other machines. What is it in your opinion about B3D forum goers that means their machines die more often than usual, and which is screing up the numbers here? And why isn't the same inflated failure rate shown on the other machines?

You keep saying the poll-base can't be accurate, but you haven't really given reasons why other than general complaints about internet polls and referencing other polls.

If you cannot relate them to this, its either because i didn't explain it good enough, or because your not willing to understand, or because you simply do not understand.

Let me try again to tell you exactly why this poll is worthless for making any statistical assumptions in regards to the consoles failure rate as a whole:

1. For a proper sample size out of 10 000 000 consoles, you would need far more than 100 consoles, to have any number thats has a statistical value. A bare minimum of sample size required to get a result that can be said as 95% accurate (95% accuracy is the lowest statistical accuracy at where numbers can be accepted at a scientific level) would in this case (roughly calculated) be around 16.000 (assuming the consoles for the sample are the average of all consoles in terms of how old and how used they are) . Consequently, the number of participants in any sample is directly related to the standard deviation of the sampling distribution. The more participants, the narrower the distribution, and the greater the likelihood that any differences will be discovered.

2. Poll not accounting for multiple consoles\multiple replacements: Some people own more than 1 console, others own only one, but has had it replaced several time, others again has had 1 brake down and then received a proper working console,

3. The Beyond 3D userbase is not an average representation of the console userbase as a whole. There is much more people who bought their consoles at launch\ near launch than in the average console population. Launch consoles are probably going to crap out on you more often than a console bought in december 2006. There is also indications that far more people than average have Xbox Live Gold, do B3Der's play roughly the same as the average joe who bought a console in order to play some Gears of War and Fifa\Madden? There is a number of unkowns. Maybe B3Der's play to long each sitting, which for some reason hurts the consoles lifetime? (Im not saying that the latter part its likely, it just examples of unkowns here)

4. The same argument you said about TV shows and voting on opinions can be turned around and used in this place. People with broken consoles are more likely to be on the forum than people with working consoles. Just like people with broken X360's are more likely to make a post about how their X360 is broken, rather than a person going to a forum saying his X360 is working fine.

5. Because the sample size is not big enough, and because the B3D's consoles arent a representative of the average console, these numbers would by anybody who has had statistics at a university. largely be written of as abnormalities. If i where to poll 100 people left and right on different forums, i would sooner or later encounter a group of people with 100% working consoles, the results from that group would also be obviously wrong.

To end this post, i would like to tell the following example:

There are two hospitals: in the first one, 120 babies are born every day, in the other, only 12. On average, the ratio of baby boys to baby girls born every day in each hospital is 50/50. However, one day, in one of those hospitals twice as many baby girls were born as baby boys. In which hospital was it more likely to happen?

It is much more likely to happen in the small hospital. The reason for this is that the probability of a random deviation of a particular size (from the population mean), decreases with the increase in the sample size.

The larger the sample the less likeliness for abnormalities, this sample is far to small to make any conclusions going towards the all the consoles out there, the only thing you can really make as an conclusion would be the failure rate @ B3D, which isn't worth much speaking on a global basis.

 

[Shifty Geezer]

Let me try again to tell you exactly why this poll is worthless for making any statistical assumptions in regards to the consoles failure rate as a whole:

1. For a proper sample size out of 10 000 000 consoles

Of course. However this is the best we can do, and unless there's good cause to think the average has that much variety, the small poll size should be indicative, if not accurate.

2. Poll not accounting for multiple consoles\multiple replacements:

Yes, but there's no scope for that in polls on this board, so instead of trying to pin down a proper scientific number, the poll was set up just to see what the current state of affairs was.

3. The Beyond 3D userbase is not an average representation of the console userbase as a whole...Maybe B3Der's play to long each sitting, which for some reason hurts the consoles lifetime? (Im not saying that the latter part its likely, it just examples of unkowns here)

Though I can agree with that, I wouldn't say that in itself would cause a higher than average fail rate for XB360s without a logical reason why, and so wouldn't discount the poll out of hand. The machines are switched on and off and games are played. It's the same activities for hardcore gamers as total lightweights. Thus what are B3Ders likely to be doing that'll drive up the fail rate? Buying early machines is of course one example, but that would tell use that the high fail is confined to the early machines if not the latest machines - the cause of the high numbers is still faults in the box, and way above MS's figures for launch unit failures. I've said before that numbers here aren't going to tell us future fail rates, but the number do tell us what the current failure rate has been in terms of customers with broken hardware. And unless there's reason to think B3Ders mistreat their machines, such as on for 24/7 (and one would expect that hardcore gamers actually treat their machines with better respect, like better aeration around the units, because they understand and value the tech more), the number given has to be mostly as a result of units failing within proper functioning.

4. The same argument you said about TV shows and voting on opinions can be turned around and used in this place. People with broken consoles are more likely to be on the forum than people with working consoles. Just like people with broken X360's are more likely to make a post about how their X360 is broken, rather than a person going to a forum saying his X360 is working fine.

Yes with the making a post, but no with making a vote, IMO. People do tend to talk more about down sides than up. For every forum post about a busted box, there'll be however many working units that people don't post about because that's the norm. However a poll is different. It's open to all opinions, and invites people to vote working units as well as non working. Now if people are still twice as likely to vote 'my machine's busted' than vote 'my machine works fine', that behaviour should be uniform across all polls. Which means the comparison with the PS3 and Wii and last gen is valid. If the XB360 failure rate isn't as high as this poll shows, neither is the XB or PS2 failure rate, which means XB360 is still comparatively extraordinarily high.

5...If i where to poll 100 people left and right on different forums, i would sooner or later encounter a group of people with 100% working consoles, the results from that group would also be obviously wrong.

But the chances of that happening are relative to the failure rate. If the failure rate is 4 out of 5, eventually you could run a poll that finds 100 working machines, but it'll take a long, long time in all likelihood! Now we're not trying to get accurate results - we've already determined that's impossible. The best we can do is get feedback from users, and see what info that gives. The info, if the poll is valid, is not accurate but is within a certain percentage of error. Now I'm no statistician, but from what I know, random sampling can get accurate results from small samples when dealing with large populations. In this case isn't the statistical accuracy probably (90% chance of being) something like +/- 10%? Or a 1 in 10 chance that this survey has abnormal results, and a 9 in 10 chance of this survey being within 10% of the real results if you surveyed everyone (ignoring population differences)?

If I've understood and remembered right, what we've got is a fairly accurate result of the local population. That is for XB360 owning B3Ders, 40% have had a console die on them, and that figure is within 10% of the real figure for B3Ders. If that population can be proven as different from the rest of the millions of XB360 owners, then it's not representative. If B3Ders aren't hacking their machines about any more than Joe Public, then it becomes far more representative. Even ditching the 40% as not being sensible to one's own expectations of failure giving economic impacts of releasing such faulty hardware, it's very hard to argue down the poll results as low as 5%. The sensible conclusion has to be 'we don't know what the percentage failure rate actually is, but at the moment it's far higher than normal for CE goods, and at the moment these boxes die more often than any other console made to date.'

 

[Cheezdoodles]

Of course. However this is the best we can do, and unless there's good cause to think the average has that much variety, the small poll size should be indicative, if not accurate.

This is an assumption based on broken logic. The sample size is to small to be an indicative of anything on a global level, as i have already explained to you.

Yes, but there's no scope for that in polls on this board, so instead of trying to pin down a proper scientific number, the poll was set up just to see what the current state of affairs was.

But since the poll doesn't have the proper options, the current state of affairs would be wrong anyway.

Though I can agree with that, I wouldn't say that in itself would cause a higher than average fail rate for XB360s without a logical reason why, and so wouldn't discount the poll out of hand.....

...Buying early machines is of course one example, but that would tell use that the high fail is confined to the early machines if not the latest machines - the cause of the high numbers is still faults in the box, and way above MS's figures for launch unit failures.

Your not making any sense. I clearly pointed out that the B3D userbase is in no means an representative average of the global x360 userbase. The fact that the amount of launch numbers \ early consoles are much higher than in proportion to the global userbase would alone cause a higher average fail rate. It being higher than MS's PR figures or not is irrelevant, its still wrong.

which means XB360 is still comparatively extraordinarily high.

.... im not arguing that the failure rate is not extraordinarily high, so stop writing your posts around that. What im telling you is that the poll results is worthless to use for trying to tell anything outside of the B3D userbase.

Now I'm no statistician, but from what I know, random sampling can get accurate results from small samples when dealing with large populations. In this case isn't the statistical accuracy probably (90% chance of being) something like +/- 10%? Or a 1 in 10 chance that this survey has abnormal results, and a 9 in 10 chance of this survey being within 10% of the real results if you surveyed everyone (ignoring population differences)?

Random sampling can get accurate results, but this is not a random sample. In order for these random samples to be an accurate indication of what we are discussing here, the sample would need to be carefully selected to be an representative of the average consoles. Which its not, as i have explained to you several times already.

In this case, the statistical accuracy is way smaller than that. You would need, assuming that your samples where taken from a proper average, 1300ish consoles to get something has has a 50% chance of being statistically accurate (these numbers are just rough numbers, i dont have SBSS on my laptop). Your sample size of 100 is to low for me to measure.

Where have you gotten this 10% figure from anyway?

If I've understood and remembered right, what we've got is a fairly accurate result of the local population. That is for XB360 owning B3Ders, 40% have had a console die on them, and that figure is within 10% of the real figure for B3Ders. If that population can be proven as different from the rest of the millions of XB360 owners, then it's not representative.

I have already proven it to be different, just judging by the amount of launch consoles and other things in my earlier post. Not only that, your using broken logic here.

The B3D population can simply be an abnormality, the sample size is TO SMALL, and not random for it to be an representative of anything outside of B3D.

it's very hard to argue down the poll results as low as 5%. The sensible conclusion has to be 'we don't know what the percentage failure rate actually is, but at the moment it's far higher than normal for CE goods, and at the moment these boxes die more often than any other console made to date.'

I have never said anything about 5%, i already said in my first post here that i believe the reliability for the X360 to be significantly worse than with its competitors. Still, the sensible conclusion still is based on nothing from this poll, as its not statistically reliable for anything.

 

[warb]

If I've understood and remembered right, what we've got is a fairly accurate result of the local population. That is for XB360 owning B3Ders, 40% have had a console die on them, and that figure is within 10% of the real figure for B3Ders. If that population can be proven as different from the rest of the millions of XB360 owners, then it's not representative. If B3Ders aren't hacking their machines about any more than Joe Public, then it becomes far more representative. Even ditching the 40% as not being sensible to one's own expectations of failure giving economic impacts of releasing such faulty hardware, it's very hard to argue down the poll results as low as 5%. The sensible conclusion has to be 'we don't know what the percentage failure rate actually is, but at the moment it's far higher than normal for CE goods, and at the moment these boxes die more often than any other console made to date.'

Didn't MS admit that the failure rate of consoles manufactured before 01/01/2006 was unusually high? Since this is a console discussion board, it might be safe to assume that quite a few of those dead 360s were manufactured in 2005. Another poll might clear it up.

How many Xbox360 consoles were manufactured in 2005, 400000? I wouldn't be surprised if that unusually high failure rate was up around 50%. If we can believe that post 2005 360 deaths are at a much more acceptable 3%, there would be 488000 broken consoles total; less than 5% of 10,000,000.

Far fetched?

 

[PatrickL]

The poll can't give accurate results for an obvious reason: you don't know how many people on a board give really honest answer. How many of them push their agenda thru polls and other topics ?

You have better chance to get an accurate result asking the same question to random people in the streets around your home I guess

 

[Shifty Geezer]

I guess we'll just have to agree to disagree, as you take the view that according to the principles of statistics, the results can be ignored without a rational explanation why the could be incorrect. As an example, consider a bag with 100 marbles in a bag of the same size and material, and a small inclined plank. We take out 3 marbles and place them each in turn at the top of the plank, and they all roll down. According to statistics that's no evidence that the rest of the marbles will roll down the hill, but using other reasoning, we can see that the rest of the marbles are the same, and under the same conditions should act the same way.

Applying that reasoning to the XB360 situation, we know it's the same box undergoing the same general uses, so the outcome should be the same. If the outcome is to differ, there must be a reason why! You've presented a couple of areas where they could be wrong :
1) That there's false voting. For that to be true enough for these results to be totally and utterly wrong, are you saying that aeriic, Clockwork, mrcorbo, NuvNacST3, rusty, TheChefO and [Maven] are all lying, at least? And why then aren't that many people lying about PS3 and Wii failures? Surely the proportion of truthful votes is the same, so the relative failure rates are the same?
2) That B3Der's are doing things that shorten the life of their consoles. Like what? Why would B3Ders console live a shorter life?

XB360 has been on sale for 18 months. It's the same box in the shops for whoever goes in and picks one up. The odds of B3Ders picking up a duffer are the same as Joe Public. Of 109 XB360's bought, 42 have died. That does not give a %age failure of XB360s of 40%, because it's not statistically valud. But on the converse, you can't just ignore that info by saying it's not statically valid. Either these machines have been prematurely killed by their users, or the results are lies, or it's true that machine deaths are common enough for an inaccurate poll to get such rates. That is, if the fatality rate were only 1%, we wouldn't be seeing these reported figures (unless they're lies or people abusing the systems). This then feeds back info into the questions of XB360's failure rate, where some people (not yourself) have said it's 3-5%, and some say it's more. If the failure rate is only 3-5%, we shouldn't be getting these results. Without this survey, we'd have no numbers at all to see if the failure rate is 5%, and instead be reliant on people saying 'my XB360 died, so it must have a high failure rate'.

i already said in my first post here that i believe the reliability for the X360 to be significantly worse than with its competitors. Still, the sensible conclusion still is based on nothing from this poll, as its not statistically reliable for anything

Have you seen official returns numbers then? If not, what is reason for believing XB360 reliability is significantly worse than its competitors?

 

[Rainbow Man]

This is an assumption based on broken logic. The sample size is to small to be an indicative of anything on a global level, as i have already explained to you.

Methinks thou doth protest too much.

If we make a poll asking wether a particular unit's broken down and ask the same question for 3 different brands of X..and one of those brands come out with a failure rate that is 10x higher than any of the other two you can't seriously argue the poll is at fault.

If it WAS the neach of the three brands would have an abnormally high failure rate but obviously that is not the case here. Just one of them is.

Now we could aruge until the cows come home wether the failure rate shown in this poll is accurate or not (given the limitations of the board software's poll options and sample size iit probably isn't) but it surely gives an indication of proportional failure rates compared to the other console brands.


However living in apparant cpomplete denial with regards to the results isn't going to be helpful to reach any sort of conclusion.
Peace.

 

[PatrickL]

If the return rate is so high, I find strange that no class action started yet ? With the short warranty US had , a lot of people should have been left with just dead consoles no ?

 

[DuckThor Evil]

How many Xbox360 consoles were manufactured in 2005, 400000?

Way more than that. There were already more consoles than that available at day 1.Around 1 million was sold during 2005 so many more were manufactured.

BTW I just listened the latest 1up podcast and they said that probably 75% of their boxes had broken... I personally think I have seen enough evidence with this poll being part of that, that it's ok for me to draw a conclusion of them breaking left and right, despite how many babies are born in a small Norwegian hosbital...

 

[Cheezdoodles]

I guess we'll just have to agree to disagree, as you take the view that according to the principles of statistics, the results can be ignored without a rational explanation why the could be incorrect.

WHAT???

I have in detail given you exact information as to why the results can and should be ignored in terms of making any indicative of the reliability wordwide. YOUR the one who is sitting here, basically disgreeing because you do not understand the basics of statistics.

I cannot believe how your argumenting here. Your basically ignoring everything i said, and standing by your point that this poll should be an indicative of something on a global basis aswell, simply because you do not understand statistics.

This is a poll, this is statistics. In order for you to make ANY assumptions that have any value, you need to understand the basics. Yet you disagree with this, because its your poll, and you dont understand it, and therefore all my explanations are not rational.

As an example, consider a bag with 100 marbles in a bag of the same size and material, and a small inclined plank. We take out 3 marbles and place them each in turn at the top of the plank, and they all roll down. According to statistics that's no evidence that the rest of the marbles will roll down the hill, but using other reasoning, we can see that the rest of the marbles are the same, and under the same conditions should act the same way.

This is something completely different and a totally irrelevant example. your just not willing to understand or read into my comments. First of all, you cannot use any meaningful statistics to calculate something when you only measuring one outcome. However, i can tell you right now, that if you were to measure where these marbles traveled, there would be abnormalities, even tho they would be of the same size and material.

I'm sorry if you do not understand this, but at least try to listen and try to understand what other people who know what they are talking about are saying. Instead of blindly defending your poll.

All i am doing, is simply explaining why this poll result is not enough to give any indication of the situation globally.

Applying that reasoning to the XB360 situation, we know it's the same box undergoing the same general uses, so the outcome should be the same. If the outcome is to differ, there must be a reason why! You've presented a couple of areas where they could be wrong :
1) That there's false voting...... And why then aren't that many people lying about PS3 and Wii failures?
2) That B3Der's are doing things that shorten the life of their consoles. Like what? Why would B3Ders console live a shorter life?

You have selective memory. I gave several examples of numbers of unkown, for example the average B3Der could be playing more than the average joe. However the main argument here is that there is significantly more launch\early generation consoles in the B3D userbase, than it is world wide.

Surely the proportion of truthful votes is the same, so the relative failure rates are the same.....

This is also flawed logic, you are assuming that there is the same amount of Wii, PS and Xbox fans all around. The Console boards for B3D is largely PS positive, compared to a global view. Again, it was an example as to why the polling results could be wrong, not an answer.

XB360 has been on sale for 18 months. It's the same box in the shops for whoever goes in and picks one up. The odds of B3Ders picking up a duffer are the same as Joe Public.

Wrong, because your statistics are based on a number that largely consists of a population that has bought launch\early generation consoles. Which i have pointed out 3 times now, but you brush this of, why i dont know.

But on the converse, you can't just ignore that info by saying it's not statically valid. Either these machines have been prematurely killed by their users, or the results are lies, or it's true that machine deaths are common enough for an inaccurate poll to get such rates.

It could be a number of things, shifty, i have already pointed out something called an ABNORMALITY . Tiny small samples like this are much more likely to have abnormalities because the probability of a random deviation of a particular size , decreases with the increase in the sample size.

Have you seen official returns numbers then? If not, what is reason for believing XB360 reliability is significantly worse than its competitors?

The reasoning why i belive the numbers are significantly worst, because of media reports. There has been way more reports that the X360's are breaking, thats why i believe the failure rate is higher than normal.

Now please, if you do not understand statistics, either read up on it, or do not try to make any more arguments. What your doing right now, is simply brushing off everything i have said, because you do not understand it. If you want to have a proper debate, try to understand things, or just don't.

It would be like me arguing with somebody else that the (excuse my english) sum of all angels in a triangle is not 180 degrees, because i do not understand math. My arguments would be something along the lines of yours, that surely you could make a triangle with 59.8, 60 and 60,1 degrees in the angles.


Thats basically what your doing. Your arguing because you do not understand.

 

[Cheezdoodles]

I personally think I have seen enough evidence with this poll being part of that, that it's ok for me to draw a conclusion of them breaking left and right, despite how many babies are born in a small Norwegian hosbital...

FFS. im not arguing that the X360 failure rate isn't high.

Im just arguing that you cannot use this poll or any other web-poll around this to give you an result that would be of interest statistically.

You not being able to understand the analogy and simply math behind the hospital example, is not my fault.

Okay, so 1up reported 75% of their launch consoles to be broken. You think thats a statistically correct number? You think that worldwide, 75% of all X360's are broken? Its whats called an abnormality. If 75% was the case, Microsoft could just close up the X360 production, and instead, buy every gaming division Sony owns including the PS franchise, for the money they would have lost on repairs.

 

[NovemberMike]

As an example, consider a bag with 100 marbles in a bag of the same size and material, and a small inclined plank. We take out 3 marbles and place them each in turn at the top of the plank, and they all roll down. According to statistics that's no evidence that the rest of the marbles will roll down the hill, but using other reasoning, we can see that the rest of the marbles are the same, and under the same conditions should act the same way.

Say I flip twenty coins and they all land heads (I have done so before). I can see that the rest of the coins are the same, so they should all land heads too, right?

And since the Cell is the same for every one, and the three I selected all have 8 SPEs working all Cell's will have 8 SPEs working?

My god, you just proved Statistics wrong! Quick, get to the Nobel people before somebody steals your work!

 

[Cheezdoodles]

Here is an example shitfy, on the logic you are using right now:


Since according to you, this poll is an indicative of something wordwide, obviously NPD and all other console sales tracking outlets out there, are plain wrong! Playstation is selling much better than reported, it has 200% more sales than the Wii!!! A poll of 100ish consoles owners at B3D, is for some reason an indicative of world sales.

OMG! Lets go to www.ps3forum.com and do a poll there! People are going to have 99% PS3's and some multiconsole owners! And we could probably get like 50000 people to vote in that poll (its a popular forum). That would be big sample, and surely prove that the PS3 is breaking every sale record in the book!!

See why this logic doesn't make sense?

What you see in the poll, is obviously an abnormality compared to the population average. Because world wide, the Wii is outselling the PS3 by 2:1 (if not more).

In fact shifty, since you are refusing to listen to my statistic explanation, and you truly believe that this is an proper indication of failure rates globally. I have the following challenge for you:

If this is to be a an indicative of something globally, (and you refuse that there is any thing like people being dishonest, not voting, etc) , this poll should also be an indicative of world sales. Now, please, explain to me how every tracking outlet in the world is mistaken, and that the PS3 is actually outselling the Wii 2:1. Because according to you, the sample size, and all statistics and all my arguments are wrong.

Therefore this should be a proper indicative of how many PS3\X360\Wii have been sold on a world basis. (Now i can explain this as an abnormality, and that the B3D userbase, is largely interested in high-end electronics, and graphics, and is therefore a sample thats not statistically large enough, nor is it a sample representing the average console buyer, but thats statistics..)

 

[Shifty Geezer]

Since according to you, this poll is an indicative of something wordwide, obviously NPD and all other console sales tracking outlets out there, are plain wrong! Playstation is selling much better than reported, it has 200% more sales than the Wii!!! A poll of 100ish consoles owners at B3D, is for some reason an indicative of world sales.

No, it isn't. We know that because the population of B3Ders are more hardcore gamers, they're more likely to buy the different consoles. The buying habits probably differ from the mainstream. There's a logical reason to think B3D isn't indicative in that respect. Before the consles ever released, we could look at this population and predict that there'd be more PS3's purchased and less Wii's purchased because of the hi-tech nature of B3Ders. Thus it's unsurprising that the sales figures don't tally with the rest of the world. What I don't see is a reason to expect B3Ders consoles to die more often. Before the machines ever released, I wouldn't have cause to think that a lot more machines would die in the hands of B3Ders than the general public - that's what I mean by a rational explanation. There's a rational reason to expect buying habits to differ from the rest of the world. What is the reason to think reliability will differ? The only given explanation in that regard so far is 'they're likely early adopters' which only means that high failure rate is due to early hardware being dodgy. The hardware still failed, and there's reason to wonder why and what may have changed in future hardware to correct that problem. Thus that bit of information is accurate and useful. It doesn't tell us that current and future machines are highly likely to die. But it tells us that the machines bought so far have had a problem. There's no reason to think that it's the owners who have broken the machines themselves.

And as for liars, of course I accept they exist, but not in significant numbers. I don't see this forum being polarized that much towards PS that you'll get more than, say, 10% of votes being made up. And I'd expect some made up votes from the other side too. I can't see reason to think that the real figures are much differnt from those recorded here. There probably is about 40 in 110 XB360's that died. It's not going to be 10 out of 80 with 30 made up votes just to make the XB360 look bad, unless I totally misread this forum and expect more balance and maturity from it's inhabitants than there is.

In fact shifty, since you are refusing to listen to my statistic explanation, and you truly believe that this is an proper indication of failure rates globally.

No I don't! It was never a statistical argument. It's only really statistical because the poll puts numbers on. Maybe you're missing the irony in my title and think I'm sincere in calling this a statistical poll? The numbers in this poll aren't any use other than giving a general impression of the current state of hardware ownership, and from considering that impression, we can question larger issues. There is a high failure rate of XB360s recorded. We can now question why and if that cause has been remedied, or we can determine that the results are absolutely bogus and my attempt to find the current state of hardware among B3Ders failed.

The reasoning why i belive the numbers are significantly worst, because of media reports. There has been way more reports that the X360's are breaking, thats why i believe the failure rate is higher than normal.

But media reports are as likely to be swung be topic momentum as forum posts. The media only reports what it sees and hears and is just as likely to be swung by itself. I remember media coverage of stupid things like Flesh Eating Bugs that gained momentum from nothing more than a few headlines encouraging other newspapers and programmes to investigate the same topic. The media is just as guilty of picking up forum complaints and passing them on, no? We've certainly had lots of wrong media coverage before on various topics.

The fact is we have no scientific data to form opinions on console reliability. None. Nada, Zip. Not even my hugely scientific poll. All we have had is a lot of forum chatter, and media sites posting articles. The same limits to this poll's reliability affect those reports. People are more likely to post threads or write articles about things going wrong than everything working just fine. So do XB360s break down more often than other consoles? Or no more often? Or rarely, and there's just been some momentum that's snowballed? If we're to talk about XB360's breaking more often than not, and if it's a mobo design or whatever, we need some sort of data that shows how much more problematic the hardware is. There's no point saying 'The X-clamp is the culprit' if the number of XB360's that fail is no different to the average for CE goods caused by random component failure. So I ran a poll to get some numbers. It shows XB360 has a bad reliability. It doesn't show scientific numbers! It never even tried to. It made no attempt to say 'an XB360 bought today is 12.36 times more likely to die within a year than a PS3 bought today.' Nothing of the sort! All it shows is that yes, there are issues with XB360. It shows PS2's over 5 years have had their share of problem but haven't broken down as much, and GC was good for reliability, and so far, this early on, if you get a PS3 or Wii with a fault you're very unlucky. The information is useful even if not statistically correct. You don't need data to be statistically correct to be able to get useful information from it.

eg. Mr. Mungo's Sweet Factory produces two new sweet products with a view to selling to the 10 million strong UK children's market. They take these sweets to a local school with a population of 200 kids and let them all taste the sweets individually without knowing what other kids have said. The first kid likes Sweet A but not sweet B. The second kid likes sweet A but not B. The third likes A but not B. The fourth likewise. And the fifth. How many kids need to be tested on these sweets to determine if they're a viable product or not? From your earlier numbers, 16,000 are needed to get an accurate measure of the population, no? So if all 200 kids in this school like A and not B, would you then go test another 15,800 children to get a statistically accurate figure to form an opinion of? Or would you ignore proper statistical protocol and think to yourself, even after 20 kids in a row have said 'yes' to A and 'no' to B, 'This sweet B wasn't a good idea'? I'd do the latter. My reasoning is that children are pretty much the same with the same tastes, and if you get such a strong negative reaction straight off, you can extrapolate response, unless there's reason to think these kids are non representative. If these children are different, you can test elsewhere. If you find out all these kids were eating toothpaste before you came along with your sweet samples, you can ignore the results as unrepresentative. But without due cause to think these kids tastes are different from the rest of the UK's children, a tiny sample is enough to base an opinion, even if statistically that's not the right thing to do. Statistics aren't the be all and end all of useful information!

In this poll, without statistically accurate figures or nice percentages that we can feed into a spreadsheet, we can see that XB360 has a noticeably higher error rate than other consoles, and so can start asking if it's the X-clamp, or the mobo layout, or whatever. Without any figures at all, all we have are 'some guy on website said his XB360 died. I know a mate who's XB360 died too. Must be a huge failure rate.' 'Nah, I know three people with XB360's since launch all working fine. There's no problem.' That hearsay is no use as a basis for smart debate!

I for one am happy to use non-statistically correct statistics to help form opinions. When buying products I check reviews which often only have a few for thousands of users. I'll probably choose the shop that got 8/10 mean average in three reviews rather than the one that got 4/10 mean from 2 reviews if looking at the stores myself shows no major discrepency. If I was on Who Wants To Be A Millionaire and used my 'ask the audience' option and 85% voted for B, I'd pick B. If I'm after an anti-ageing lotion and 85% (of 127 polled) of women said they noticed a difference in 10 days with Neuroregenimide while 70% (of 113 polled) of women noticed a difference in 10 days with Fuesli, I'll likely try the Neoregenimide first. And if I see an open public poll on a platform neutral site that show lots of people having busted XB360s and nowhere near as many busted XBs and PS2s, I'm happy to accept there was something about that XB360's that caused them to fail and start wondering what it was. Truly statistically accurate data is hard to come by, so I'm happy to use fuzzy logic and rough figures.

(Edit : If nothing else, this thread has at least brought some more orignal analogies to the forum than damned cars!)

 

[deepbrown]

Whenever these statistical polls are performed by members who are not the norm of consumers...there will of course be disparities. And of course I don't expect the percentage of failure to be as it is on this poll. But it certainly shows a trend.

Being a member of other forums...most staunch xbox360 "supporters" have admitted that they are not still using their first Xbox360.

It's a problem, and I don't think consumers should stand for it.

If this failure rate was occuring with the ps3 (might it in a years time ) this would be all over the news and the sales of the ps3 would slow to a hault...I'd almost feel there'd be public outrage. But this problem for the Xbox360 has not met the consumer. It's one that's in my mind, and I know it's in the mind of informed 360 owners...they generally don't like to leave their 360 running too long - folding@home wouldn't be an option for many, simply because they would be scared the machine would fail.

These numbers are indicative of a wider trend...and i don't feel the reported 3%-5% is accurate at all...and personally I don't think this has had enough media coverage...i feel the media and fans have been very leniant.

(I feel the same about the ps2 problem...though mine is still working since launch)

 

[Cheezdoodles]

N What I don't see is a reason to expect B3Ders consoles to die more often......

So you believe that a sample that has a higher consistency of a lanch\early production consoles, have no baring on the reliability rate at hand?

The numbers in this poll aren't any use other than giving a general impression of the current state of hardware ownership, and from considering that impression, we can question larger issues. There is a high failure rate of XB360s recorded. We can now question why and if that cause has been remedied, or we can determine that the results are absolutely bogus and my attempt to find the current state of hardware among B3Ders failed.

Actually, you cannot do this any more than you could before. Because the data collected are worthless for any sort of real meaningful analysis, it wouldn't matter if you did this poll or not.

So I ran a poll to get some numbers. It shows XB360 has a bad reliability. It doesn't show scientific numbers! It never even tried to.

YOU tried to argue that they were a proper indicative. How can something thats not scientifically accurate, lead you to a true conclusion?


I specifically said that the numbers are worthless in any global view, YOU tried to argue they are not. NOW your suddenly saying that they aren't.

All it shows is that yes, there are issues with XB360.

Yes they show that there are issues wit the X360, i never said there wasn't any issues with it. However, all i said was that this poll isn't worth a damn for for talking about a the reliability globally.

I for one am happy to use non-statistically correct statistics to help form opinions.

^^ but that is "just" as bad as just listening to an average forum poster saying that he has 100 friends and all of the consoles died. Since the numbers aren't correct in the first place, they are therefore worthless.




And now please, take the time to respond the the prior post as well. Because your switching opinions here, first you try to argue with me, that i haven't provided any real answers as to why these numbers are flawed. And now suddenly, you have switched your opinion completely.

Suddenly, this is no longer important. I hate when people try to argue with you about something they obviously dont understand, and suddenly switch their whole argument around.

All i argued was that the numbers are in no way an indicative or really anything on a global basis. YOU argued that they were