Fanbois tend to dislike car reliability surveys. With their Weltanschauung (can’t have enough fancy foreign words) threatened by any suggestion that their car of choice is less than perfect in any way, they will argue that most of the problems reported for their car are trivial, and thus should be ignored. But this implies that most of the repairs reported for other cars are serious. Why not only count serious problems? If this were done, which cars would prove truly unworthy of devotion?

I’ve been conducting TrueDelta’s car reliability survey for nearly a decade. We’re about to update our stats for the 35th time. At first we had stats that included all problems, weighting them equally. Then, a year ago, we started posting stats that included only powertrain and chassis problems. Though only about a third of the total for nearly new cars, their share does increase to two-thirds for five-year-old cars and three-quarters for ten-year-old cars.


But I can see that this isn’t going far enough for people who want to count only “serious problems.” Most powertrain and chassis repairs aren’t all that serious: oxygen sensors, valve cover gaskets, wheel bearings, CV joints, and the like. Starter, alternator, and battery failures most often strand a car, but these aren’t difficult or expensive to fix. What about inarguably serious problems?

Unfortunately, it’s first necessary to define “serious problem.” Sure, you could simply ask owners to only report problems they considered serious, but definitions of “serious” vary widely. Some people will consider a rattle serious if they expect their brand new car to not have one. Other people don’t consider a transmission failure serious if they are treated well. (Hassle-free warranties and complimentary loaners go a long way.) Time also plays a role. Something that happened a year or even a month ago seems far less serious than something that’s happening right now.

For the sake of discussion let’s posit that a serious problem is one that requires an expensive repair to the engine or transmission. Which cars commonly require these?


First up, engine rebuilds and replacements. Among the over seven hundred model / model year combinations covered by our current reliability stats (mostly 2007 and newer cars but including some back to the 2000 model year), the analysis uncovered only three cases of somewhat common somewhat major engine problems. About one in ten owners of the 2009-2011 Audi A4 2.0T and of the 2010 Chevrolet Equinox / GMC Terrain with the four-cylinder have had engines rebuilt to fix high oil consumption. (Some 2011 Q5s with the 2.0T engine have also required rebuilds, for the same reason.) The only other case where we have a decent sample size: about one in fifteen 2012 Jeep Wranglers has required a cylinder head replacement, to eliminate a ticking noise. Notably, in none of these cases are engines actually failing. And in none of them are entire engines being replaced. Put another way, we have zero cases of common engine failures.

With a larger sample size of Mazda RX-8s we might have one case. Two of 17 owners of 2005 RX-8s have reported a rotary engine failure in the last two years. The infamous IMS bearing in early-aughts Porsches? Covered that earlier.

What about turbochargers? Many people remain wary of boosted engines. Should they be? Here, one case stands out, the 2008 BMW 535i. Six out of 31 owners reported a turbocharger replacement. A couple of other cars, the 2007 Mazda CX-7 and the 2004 Subaru Forester XT, might be in the same ballpark, but the sample sizes are smaller so these could be flukes. One thing that probably isn’t a fluke: each case is the first year a turbocharger was offered in the model. The CX-7 is a well-known horror story. The Forester case has a known cause. Subaru included a “banjo” joint with a small oil filter in the line that lubricates the turbocharger’s bearing. Over time this small filter can clog, starving the bearing of oil. Some smart owners have preemptively removed the filter. Finally, roughly one in twelve turbos has failed in the 2012 and 2013 Passat TDI (diesel), even though these cars have only 15 to 25 thousand miles on them. The biggest finding, though: in the great majority of models with turbochargers, failures have been rare over the past two years, perhaps two to three percent.

Many people have raised concerns with carbon build-up in direct-injected (DI) gasoline engines, and we have quite a few reports of this problem. But, when you look at the frequency for specific models and years, it’s rarely high. In fact, it’s only high for two engines, both of which happen to be in 2007-2010 VWs and Audis: the 2.0T in the A3, CC, and Passat, the 3.2-liter V6 in the A6 and Q5, and the 3.6-liter VR6 in the Q7. The frequency in these cases varies from one car in ten to one in five, so pretty high, but far from “all of them.”

Missing from both the turbocharger and DI lists, at least so far: Ford’s EcoBoost engines.

My gut told me that transmissions are more likely to fail than engines. But even here failures simply aren’t common. One well-known case sticks out: Acuras and Hondas with V6 engines from 1999 to 2003 comprise about half the set. I’ve also become well aware of wave plate failures in the 2007-2009 GM large crossovers (Buick Enclave, GMC Acadia, Saturn Outlook). These have been running about one vehicle in ten. The same problem also affects the 2008-2009 Saturn VUE V6, and possibly also the Saturn Aura and Chevrolet Malibu of that era. Automatic transmission failures also might be relatively common in the circa-2003 Chrysler minivans, circa-2003 Golf and Jetta, 2006 Ford Explorer (first model year), 2011 Elantra Touring (possibly earlier years as well, limited by sample sizes), perhaps one in ten first-year Nissans with the 4-cylinder and CVT (2008 Sentra, 2009 Rogue, 2009-2010 cube), and the 2013-2014 Nissan Pathfinder and Infiniti QX60 (but not the JX). Owners of the last commonly report a “judder” at low speeds, not outright failures.

Considering that we do have good sample sizes for over 700 model-model year combinations, once again not a long list.


Anyone who spends much time talking cars on the net seems to hear engine and transmission horror stories all the time. How can we not have more cases in our stats? Partly it’s a matter of our sample. We don’t cover many models over eight years old, and serious failures no doubt become more common once cars have well over 100,000 miles on them. But there’s also the power of the Internet. The web has made it possible for each of us to come into contact with tens of thousands of other people. This didn’t happen 20 years ago. If we also include the people these people come into contact with, well, we’re no longer dealing with six degrees of separation. You want an engine failure in a specific car model? Hell, I can find you ten. By three o’clock this afternoon.

But these will be ten out of thousands of cars. To get a sense of how common “common problems” actually are, it’s necessary to gather data through a properly structured process. Do this, and you’ll find that unarguably serious car problems are rarely common among cars with fewer than 100,000, even 120,000 miles. And if you happen to own one of the exceptions, such as the Mazda RX-8 in my garage? Feel free to love it anyway.

Michael operates, which gathers and shares information from car owners on reliability, real-world fuel economy, and why (not) to buy the one they did. The more car owners participate, the better this information becomes.