this guy says hes on his 3rd 18 engine

[bootlegger]

Why would I ever be defensive? I honestly don't care about the issue I just hate bad statistics.

I wouldn’t call 62+ people with the same exact issue bad statistical input. It would be bad statistics to ignore the data we do have.

Sponsored

 

[kidamer]

So you posses the Ford warranty complaint records? Do show us all the numbers.

oh you must , oh wait your using a forum poll with a whopping 100 total votes, so theres that.

 

[Stormtroopin5.0]

Bootlegger, no point in trying to explain statistical analysis to folks who don't value trend insight that can be attained. The only gripe I have with using the poll data is as you has mentioned previously, those of whom who voted no but now have the noise are not captured in the dataset therefore misrepresenting the actual sample size. I much like you voted no initially but now at a hair over 500mi have both noises.

 

[thehunterooo]

Bootlegger, no point in trying to explain statistical analysis to folks who don't value trend insight that can be attained. The only gripe I have with using the poll data is as you has mentioned previously, those of whom who voted no but now have the noise are not captured in the dataset therefore misrepresenting the actual sample size. I much like you voted no initially but now at a hair over 500mi have both noises.

You mean people abusing statistics? Explain to me how only having the poll in a thread with the related problem is a good sample? You can defend it troll. A bunch of paranoid people is not a trend let alone a bunch of drama YouTubers crying for views in order to gain subs and make money.

 

[thehunterooo]

I wouldn’t call 62+ people with the same exact issue bad statistical input. It would be bad statistics to ignore the data we do have.

Where did I say it was not an issue? You are implying that from a biased sample that a certain amount of people have the problem.

A person did it with the evap core and it was silly then. Omg wow 50% of the people who came into the thread have the same issue that means 50% of everyone has the issue! Yep that is a trend. Stop with the troll stats.

Put the poll in the general section and ask all 2018 owners to vote on it. Should be similar numbers right?

 

[bootlegger]

oh you must , oh wait your using a forum poll with a whopping 100 total votes, so theres that.

124. At least use the correct numbers when trashing the data.

 

[bootlegger]

Where did I say it was not an issue? You are implying that from a biased sample that a certain amount of people have the problem.

A person did it with the evap core and it was silly then. Omg wow 50% of the people who came into the thread have the same issue that means 50% of everyone has the issue! Yep that is a trend. Stop with the troll stats.

Put the poll in the general section and ask all 2018 owners to vote on it. Should be similar numbers right?

I am implying that out of 124 people answering a poll, 63 have the same issue. I never said the percentage was exact to the entire population of Mustangs sold, only that the results had some statistical significance. Even if there were 1000 2018 GT owners on the forum (there aren't), the 63 would be rather significant.
I would start a poll in the General section, but I really don't feel like dealing with the militant fanboys who bash anyone interested in discussing the topic. I have already seen how that turns out in other threads. If you want to do it, you have my support. More data points is better. If you are calling me a troll for being interested in an unknown issue and showing concern for the large number of members reporting this noise, I think you are way off base.

 

[SlaughterOfTheSoul]

Probability is about perspective but we can get some absolutes. Yes, sampling here will have bias if you try to represent it as a proportion of the whole. However, being an engineer that uses stats regularly, that's not what caught my attention. Please excuse these numbers, for they are only to illustrate an example for the conclusion.

So Ford sells 7k units a month. There's been 10 months since the 18s have really been getting out there. Assuming GTs are 5% of production, that's 3500 cars in total so far. How many of them actually visit forums? That's hard to say, it's not 100%, but it's a dipole composition between enthusiasts and people who are having problems. Unfortunately, we're seeing people represented in this thread who were enthusiasts before they bought their 18. That perspective significantly reduces a general forum bias. It doesn't eliminate sub-section bias, but it's still a clean point of data.

But we don't have enough samples for a good survey? Well, we can do a best case estimate with this information. At this point, lets say there are 64/3500 cars with the noise, or ~2%. And 50% of those have the noise, then that's 1%. Not so bad, right?

Not so fast.

The chances of those enthusiasts being previously present and not showing issues with previous platforms yet simultaneously demonstrating the exact same issue across multiple samples changes things drastically. This is where the perspective aspect comes into play. Yes there is bias because of the section, but there is also another POSTERIOR factor due to the type of demographic being represented.

So at this point, you can take a sample of all enthusiast members and compare them to the total number of 18 GT owners on the site, as we've previously determined them to be relatively unbiased samples (at least not biased against Ford).

In summation, I disagree that we have an insignificant portion of failures, as the composition of 18 GT owners on the site can be reasonably assessed and compared against the number of enthusiast(people who would not otherwise complain) failures. However, we WOULD need the number of 18 GT owners on the site, but even if 10% sign up, that's only 35. See where I am going with this?

 

[kidamer]

Probability is about perspective but we can get some absolutes. Yes, sampling here will have bias if you try to represent it as a proportion of the whole. However, being an engineer that uses stats regularly, that's not what caught my attention. Please excuse these numbers, for they are only to illustrate an example for the conclusion.

So Ford sells 7k units a month. There's been 10 months since the 18s have really been getting out there. Assuming GTs are 5% of production, that's 3500 cars in total so far. How many of them actually visit forums? That's hard to say, it's not 100%, but it's a dipole composition between enthusiasts and people who are having problems. Unfortunately, we're seeing people represented in this thread who were enthusiasts before they bought their 18. That perspective significantly reduces a general forum bias. It doesn't eliminate sub-section bias, but it's still a clean point of data.

But we don't have enough samples for a good survey? Well, we can do a best case estimate with this information. At this point, lets say there are 64/3500 cars with the noise, or ~2%. And 50% of those have the noise, then that's 1%. Not so bad, right?

Not so fast.

The chances of those enthusiasts being previously present and not showing issues with previous platforms yet simultaneously demonstrating the exact same issue across multiple samples changes things drastically. This is where the perspective aspect comes into play. Yes there is bias because of the section, but there is also another POSTERIOR factor due to the type of demographic being represented.

So at this point, you can take a sample of all enthusiast members and compare them to the total number of 18 GT owners on the site, as we've previously determined them to be relatively unbiased samples (at least not biased against Ford).

In summation, I disagree that we have an insignificant portion of failures, as the composition of 18 GT owners on the site can be reasonably assessed and compared against the number of enthusiast(people who would not otherwise complain) failures. However, we WOULD need the number of 18 GT owners on the site, but even if 10% sign up, that's only 35. See where I am going with this?

yes I see where you going the sample size is to small

 

[bootlegger]

yes I see where you going the sample size is to small

*Too

And you did completely miss the point.

 

[Bald Menace]

something I have been thinking about .. the block is cast aluminum with a spray in liner so block and bores are going to expand outward as they warm up. pistons are hypereutectic which is known for it's low rate of expansion which allows tighter piston to cylinder wall clearances. so these parts expand at different rates therefore changing cold and warm clearances drastically. A cast iron liner doesn't experience this issue because iron has a lower rate of growth when heated compared to aluminum. for example, when we install aluminum cylinder heads on a cast iron Buick Block we have to factor in about 7-10 thousandths growth in the cylinder head for lifter preload or lash adjustment. perhaps this is an issue in the gen 3 coyote rattle issues

 

[GT Pony]

So Ford sells 7k units a month. There's been 10 months since the 18s have really been getting out there. Assuming GTs are 5% of production, that's 3500 cars in total so far.

However, we WOULD need the number of 18 GT owners on the site, but even if 10% sign up, that's only 35.

10% of 3500 is 350, not 35.

 

[Zathras]

FWIW, for MY2015, about 40% of Mustangs were the GT model (60% were ecoboost or V6). So if that still holds, as an estimate, and there have been 50k '2018 mustangs sold to date (based on the July production numbers just released) that's very roughly 20,000 '18 Mustang GTs in owner's hands right now...

 

[ElAviator72]

If the engine is dying due to some sort of misuse/abuse, and the owner repeats the same thing over and over....Einstein said the definition of insanity is repeating the same thing and expecting a different result

How are these engines faring in the F-150? Does the F-150 get the DI/linerless version of the Coyote?

 

[ElAviator72]

something I have been thinking about .. the block is cast aluminum with a spray in liner so block and bores are going to expand outward as they warm up. pistons are hypereutectic which is known for it's low rate of expansion which allows tighter piston to cylinder wall clearances. so these parts expand at different rates therefore changing cold and warm clearances drastically. A cast iron liner doesn't experience this issue because iron has a lower rate of growth when heated compared to aluminum. for example, when we install aluminum cylinder heads on a cast iron Buick Block we have to factor in about 7-10 thousandths growth in the cylinder head for lifter preload or lash adjustment. perhaps this is an issue in the gen 3 coyote rattle issues

Aluminum engine blocks without sleeves have been a disaster just about everywhere...ask GM who tried it on the Vega and BMW who had the imfamous nicasil lining in their V8's in the early 1990's.

Sponsored