Depends what goes wrong.

Lots of little niggly faults costing little to repair will give poor index for reliability and a low cost to repair.

Few faults but big ones will give a low index but high cost to repair.

Of course, you are comparing pigs and apples when one car costs twice as much as the other so parts costs are more expensive as well..

The index itself takes into account the average cost of repair - WD say

As a guideline, the average RI number on the 250 models we compare is 100.The Reliability Index figure is calculated as a combination of:

• the number of times a car fails,
• the cost of repairing it,
• the average amount of time it spends off the road due to repairs
• the average age and mileage of the vehicles we have on our books.

The one I'd rather they left out is the time off road. I can't imagine it's a very reliable figure anyway.

The rest of it amounts to burn costs (frequency x severity of claim) adjusted for age of vehicle (probably using a standard distribution curve for claims fall by age).

You can infer I think that of two cars with the same RI, that if once has double the average claim cost then it breaks down roughly half as often (at the same age) subject to the level of confidence in the data.

In my opinion the biggest limitation of the usefulness of the Index is that it is historic - it's generally true that there are more reliable and less reliable manufacturers, and the newer cars will often run to type - but that isn't always the case. And epidemic faults on a particular variant will potentially swamp the trend for both the model and the manufacturer.

Something else to think about - if I understand it correctly, then it's not actually a Reliability Index, it's an Unreliability Index. Because, even for an 'unreliable' model, most examples will not break down in a given year, tyhis can give a false impression of the relatively likelihood of having a trouble-free car. Comparing two models, A & B - if 5% of As break down, and 10% of Bs, other things being equal B will have double the reliability index of A and be 100% more likely to fail

But looking the other way, 95% of As won't break down, and 90% of B's won't. So A is only 5.5% (95/90-1) more likely to be trouble free than B.

So I don't get too upset when I see that the Outlander has a RI of 182. Especially when I see that 40% of problems are engine related, and at an average age of 4.4 years very few of the cars in the cohort will have the same engine as mine.

Edited by Manatee on 15/08/2014 at 17:04