Cumulative data on S65 engine failures.

[swartzentruber]

Actually what's even more interesting than how many of those are CA cars is that I believe NV and CO are both 91 octane premium states as well, so ~53% of all failures are in states with 91 octane. While I suspect more than 12% of the total M3s sold were sold in CA, I sincerely doubt that those 3 states make up 53% of cars sold.

Also, not that I was holding on to the theory anyway, but the fact that only 3 of the failures came from cold weather states makes my original cold start theory extremely unlikely. It does seem like premium octane rating is explanatory.

[catpat8000]

Originally Posted by SenorFunkyPants

Updated the first post with the latest data (added 3 fails).

Also added this:



Notes
Only includes USA cars and where a location state was provided.
Supercharged cars excluded.
California cars account for 35% of the failures against having 12% of the total USA Population.

Awesome work Mr. FunkyPants.

Have you thought about another graph normalizing by state population? Something like failures per 10 million people?

You'd see interesting things:

California has 3.2 failures per 10m people
NY has 1 per 10m people
Nevada has 7 failures per 10m people
Texas has 0.4 failures per 10m people

The data might not show anything but noise so far, mainly due to the low number of authenticated failures.

Pat

[catpat8000]

Originally Posted by catpat8000

Have you thought about another graph normalizing by state population? Something like failures per 10 million people?

You'd see interesting things:

California has 3.2 failures per 10m people
NY has 1 per 10m people
Nevada has 7 failures per 10m people
Texas has 0.4 failures per 10m people

The data might not show anything but noise so far, mainly due to the low number of authenticated failures.

Pat

Here are some other data items against which you might consider comparing as a proxy for number of M3s sold:

Car dealerships per state, and failures per 1000 dealers as of Jan 1, 2013:


CA 1364 F/D: 8.8
NV 0102 F/D: 20
NY 0877 F/D: 2.3
CT 0266 F/D: 3.7
FL 0853 F/D: 1.2
TX 1201 F/D: 0.8
MD 0303 F/D: 3.3
TN 0345 F/D: 2.9
NC 0579 F/D: 1.8
CO 0260 F/D: 3.8


Total annual new vehicle registrations by state as of Jan 1, 2013 and failures per million registrations:


CA 1.53M F/R: 7.8
NV 0.11M F/R: 18.2
NY 0.87M F/R: 2.3
CT 0.16M F/R: 6.3
FL 1.01M F/R: 1.0
TX 1.28M F/R: 0.78
MD 0.31M F/R: 3.2
TN 0.27M F/R: 3.7
NC 0.37M F/R: 2.7
CO 0.28M F/R: 3.6


Just some thoughts. It seems interesting to me that Texas is comparable to CA in terms of size, is generally much hotter than CA weather wise, and has so many fewer failures.

Pat

[US///M3]

Maybe because CA owners are more likely to run tunes, superchargers etc.

[catpat8000]

Originally Posted by US///M3

Maybe because CA owners are more likely to run tunes, superchargers etc.

Well first off, we have excluded supercharged failures from our data so there's that.

Second, your point seems unlikely to me. California has the strictest emission laws in the country, which tunes won't pass. Yes there are a lot of tuned cars but CA has 38 million people and 1.5 M annual vehicle sales. There are just a lot of cars here.

And anyway, there is no evidence whatsoever that CA owners are more likely to modify their engines with tunes. It's pure heresay.

Pat

[US///M3]

Originally Posted by catpat8000

Well first off, we have excluded supercharged failures from our data so there's that.

Second, your point seems unlikely to me. California has the strictest emission laws in the country, which tunes won't pass. Yes there are a lot of tuned cars but CA has 38 million people and 1.5 M annual vehicle sales. There are just a lot of cars here.

And anyway, there is no evidence whatsoever that CA owners are more likely to modify their engines with tunes. It's pure heresay.

Pat

Are you saying there arent a lot of cars in TX and FL? You combine TX and FL and you still have less than half of failure rates of CA.

[catpat8000]

Originally Posted by US///M3

Are you saying there arent a lot of cars in TX and FL? You combine TX and FL and you still have less than half of failure rates of CA.

edit: I thought SenorFunkyPants' data excluded modified cars. So mods are at best a diversion in this discussion. We are talking, I think, about stock failures.

[chris719]

I think we just have too small of a sample size here to be representative of the population.

[swamp2]

Excellent brainstorming and additional ways to look at the data catpat8000!

[regular guy]

Originally Posted by swamp2

Excellent brainstorming and additional ways to look at the data catpat8000!

+1

Kudos to SFP and catpat8000 for this work.

[US///M3]

Originally Posted by catpat8000

edit: I thought SenorFunkyPants' data excluded modified cars. So mods are at best a diversion in this discussion. We are talking, I think, about stock failures.

Most dont advetise having a tune when they end up with a blown engine for obvious reasons.

[catpat8000]

Originally Posted by catpat8000

It seems interesting to me that Texas is comparable to CA in terms of size, is generally much hotter than CA weather wise, and has so many fewer failures.

Pat

Here's another test: a quick search on autotrader.com for 2008-2013 M3s for sale within a radius of 50 miles of Beverly Hills shows about 90 cars. The same search over Houston yields about 30 cars. But the LA area has at least double the pop of Houston, maybe higher, so this back-of-the-envelop test suggests no more than a 50% higher M3 density in LA than Houston. Doing the same test in the Bay Area suggests the per capita M3 ownership rate is lower up north and is likely comparable to Houston.

So it doesn't seem like a 10x difference in M3 ownership between CA and TX. It actually seems pretty close, likely not more than 50% higher in CA. So you wouldn't expect to see an order of magnitude fewer failures in TX unless:
- the data is still too noisy
- something else is different in CA

Pat

[regular guy]

Originally Posted by catpat8000

So it doesn't seem like a 10x difference in M3 ownership between CA and TX. It actually seems pretty close, likely not more than 50% higher in CA. So you wouldn't expect to see an order of magnitude fewer failures in TX unless:
- the data is still too noisy
- something else is different in CA

Pat

Age and maturity of owner?

[US///M3]

Originally Posted by catpat8000

Here's another test: a quick search on autotrader.com for 2008-2013 M3s for sale within a radius of 50 miles of Beverly Hills shows about 90 cars. The same search over Houston yields about 30 cars. But the LA area has at least double the pop of Houston, maybe higher, so this back-of-the-envelop test suggests no more than a 50% higher M3 density in LA than Houston. Doing the same test in the Bay Area suggests the per capita M3 ownership rate is lower up north and is likely comparable to Houston.

So it doesn't seem like a 10x difference in M3 ownership between CA and TX. It actually seems pretty close, likely not more than 50% higher in CA. So you wouldn't expect to see an order of magnitude fewer failures in TX unless:
- the data is still too noisy
- something else is different in CA

Pat

Or maybe CA owners join car forums at a higher rate than TX and FL owners combined.

[SenorFunkyPants]

Originally Posted by catpat8000

Here are some other data items against which you might consider comparing as a proxy for number of M3s sold:

Car dealerships per state, and failures per 1000 dealers as of Jan 1, 2013:


CA 1364 F/D: 8.8
NV 0102 F/D: 20
NY 0877 F/D: 2.3
CT 0266 F/D: 3.7
FL 0853 F/D: 1.2
TX 1201 F/D: 0.8
MD 0303 F/D: 3.3
TN 0345 F/D: 2.9
NC 0579 F/D: 1.8
CO 0260 F/D: 3.8


Total annual new vehicle registrations by state as of Jan 1, 2013 and failures per million registrations:


CA 1.53M F/R: 7.8
NV 0.11M F/R: 18.2
NY 0.87M F/R: 2.3
CT 0.16M F/R: 6.3
FL 1.01M F/R: 1.0
TX 1.28M F/R: 0.78
MD 0.31M F/R: 3.2
TN 0.27M F/R: 3.7
NC 0.37M F/R: 2.7
CO 0.28M F/R: 3.6


Just some thoughts. It seems interesting to me that Texas is comparable to CA in terms of size, is generally much hotter than CA weather wise, and has so many fewer failures.

Pat

Great ideas..which made me think about the number of BMW dealership per state and how that might fit in with your figures above?
Importers are usually very strict about the placement of franchised dealers and prospective BMW dealership owners can't just set up anywhere they like but in general the bigger the market for their (BMW) cars the more dealerships there will be per population.

[SenorFunkyPants]

Originally Posted by catpat8000

edit: I thought SenorFunkyPants' data excluded modified cars. So mods are at best a diversion in this discussion. We are talking, I think, about stock failures.

The table above is only USA cars whose owners gave their location.
Excludes S/C cars but includes all other simple mods.
The problem is that often owners with engine failure wish to keep quiet about any mods/tunes in case it affect their claim. So its impossible to know how many who said no tunes did in fact have a tune (its quite obvious in some cases but its not up to me to call them a fibber).

Its interesting to download the Excel file then you can sort the data anyway you like...just use the downward pointing arrows to exclude whatever you like.

Originally Posted by catpat8000

So it doesn't seem like a 10x difference in M3 ownership between CA and TX. It actually seems pretty close, likely not more than 50% higher in CA. So you wouldn't expect to see an order of magnitude fewer failures in TX unless:
- the data is still too noisy
- something else is different in CA

regular guy; Age and maturity of owner?

US///M3; Or maybe CA owners join car forums at a higher rate than TX and FL owners combined.

Often if it walks like a duck and quacks like a duck it is actually just a duck.

[regular guy]

Originally Posted by SenorFunkyPants

regular guy; Age and maturity of owner?

In SoCal, the BMW ///M3 scene is filled with rich kids whose maturity isn't very high. If you go to a car meet, it's not uncommon to see/hear them revving the engines at full throttle, in neutral, bouncing off the RPM limiter for many seconds at a time. That's just one example I've seen repeated at least a few times a year.

What I'm trying to say is that I assume the young kids who drive M3's don't know how to take care of them and are more prone to abusing them. That's par for the course when you haven't earned your own money and don't have an association to the cost and repair bills. So I was asking what the age demographics differ specifically between SoCal and TX.

[///mawana]

Originally Posted by catpat8000

Here's another test: a quick search on autotrader.com for 2008-2013 M3s for sale within a radius of 50 miles of Beverly Hills shows about 90 cars. The same search over Houston yields about 30 cars. But the LA area has at least double the pop of Houston, maybe higher, so this back-of-the-envelop test suggests no more than a 50% higher M3 density in LA than Houston. Doing the same test in the Bay Area suggests the per capita M3 ownership rate is lower up north and is likely comparable to Houston.

So it doesn't seem like a 10x difference in M3 ownership between CA and TX. It actually seems pretty close, likely not more than 50% higher in CA. So you wouldn't expect to see an order of magnitude fewer failures in TX unless:
- the data is still too noisy
- something else is different in CA

Pat

Analysing the data in this manner would necessarily suggest SFP's octane angle does have merit. TX generally has easier access to 93 than CA with 91 largely.

[regular guy]

Originally Posted by ///mawana

Analysing the data in this manner would necessarily suggest SFP's octane angle does have merit. TX generally has easier access to 93 than CA with 91 largely.

There was always an assumption that there would need to be a connection between low octane and signs of detonation -- as the detonation would be the cause of the engine failures. But so far, none of the bearings or very few sets of pistons examined from NA engines experiencing failures showed any signs of detonation.

So I think an alternate hypothesis needs to be presented to connect low octane to the failures that excludes detonation. So far, no other hypothesis has been presented. I've got to believe that the brightest and smartest engine guys on the forum could be able to come up with something -- yet there's still a void connecting low octane without detonation to the engine failures. So for me, the connection is still very remote and unlikely.

[signes]

Originally Posted by swartzentruber

Actually what's even more interesting than how many of those are CA cars is that I believe NV and CO are both 91 octane premium states as well, so ~53% of all failures are in states with 91 octane. While I suspect more than 12% of the total M3s sold were sold in CA, I sincerely doubt that those 3 states make up 53% of cars sold.

Also, not that I was holding on to the theory anyway, but the fact that only 3 of the failures came from cold weather states makes my original cold start theory extremely unlikely. It does seem like premium octane rating is explanatory.

Not sure where the car was in NV, but certainly for Colorado you need to take elevation into account. A mile up the air is a lot thinner. Less air going into the engine effectively lowers the compression ratio meaning less chance of detonation. I saw an estimate once that the 91 we have here at 5200' of elevation is comparable to 93-94 at sea level.

This is an interesting article explaining this in a nice, concise manner.

[swartzentruber]

Originally Posted by signes

Not sure where the car was in NV, but certainly for Colorado you need to take elevation into account. A mile up the air is a lot thinner. Less air going into the engine effectively lowers the compression ratio meaning less chance of detonation. I saw an estimate once that the 91 we have here at 5200' of elevation is comparable to 93-94 at sea level.

I am fully aware that due to elevation, there is less of a need for the higher octane. But still, 91 octane is 91 octane. I'm not going to discount that fact that it represents another failure in a state with 91 octane just because in general there is less of a need for higher octane with elevation. I can think of a few possibilities: Maybe:
a) A colorado car doesn't need 93, but 91 is still too low (i.e. needs 92)
b) I don't really know CO, but I assume they sell 91 statewide, and I assume not the whole state is highly elevated. Perhaps the owner did most of their driving in lower elevation.
c) As RG suggests ealier, perhaps the issue isn't detonation, but something else somehow correlated with the lower octane gas

[swartzentruber]

Originally Posted by regular guy

So I think an alternate hypothesis needs to be presented to connect low octane to the failures that excludes detonation. So far, no other hypothesis has been presented. I've got to believe that the brightest and smartest engine guys on the forum could be able to come up with something -- yet there's still a void connecting low octane without detonation to the engine failures. So for me, the connection is still very remote and unlikely.

Could the engines be actually running a bit cooler with lower octane in such a way that the "heavy" 10W-60 oil is flowing less well, leading to increase bearing wear? That's all I got for now, but I'll keep thinking.

Also, I'm still not entirely convinced that enough sets of non S/C'd M3 bearings in failed engines have been examined to conclude that detonation is rule out. I know the ones that you and kawasaki have looked at didn't show what appears to be signs of detonation, but there are still many failed engines whose bearings were never examined. Since S/C'd engines are showing signs of failure due to detonation, it doesn't seem like it should be that much of a stretch that cars could be failing in lower octane states for the same reason.