How do you determine the overall failure rate of a "bundled" product
based on the failure rates of the individual components.  For example,
a bundled product consists of a computer monitor, keyboard and mouse. 
Those components have individual failure rates of .7%, .1% and .25%
for a 1 year period?  How do we determine the overall failure rate
of the bundle?

Mscash --

The most-conservative assumption would make them additive.

Of course there's the possibility that two components will fail from
similar causes (dirt, humidity, heat) or fail simultaneously, creating
a co-variance that reduces the OVERALL failure rate.  But without
experience to predict that, a most-conservative assumption would be
simply to add the 3 failure rates.

Best regards,

Omnivorous-GA