I'd argue CR's testing methodology has improved in the past twenty years. Unfortunately, the in-depth reporting of said tests is all but gone, which is hugely disappointing. You get a couple of dumbed-down paragraphs and a chart.
Real-world testing: It would be difficult to standardize test results if their loads were randomly soiled, both with washers and dishwashers. Having said that, I'd love to see their test of how well cookware is cleaned go beyond baked on brownie batter. Starch rings from cooking pasta/rice and the fond created by sautéing at high temps would be a nice addition.
Reliability reporting: I think they only go back five years in order to err on the side of skewing to newer models rather than letting an unreliable newer model ride the coattails of more reliable machines from 15 years ago. The opposite is also true: Ten years ago, LG washers were not very reliable. They've improved quite a bit, so it's good their new machines aren't dragged down by the problems of older models.
What I'd like to see: Get rid of the red/clear/black circles that say "Excellent", "Very Good", etc. Give us the actual number for each component of the test. A washer may rate a Very Good for cleaning performance, but there's a 20-point range within that rating. My decision to purchase a particular model may hinge on whether a machine is only a point or two away from "Excellent" or a point or two away from "Good." They think we're not able to deal with numbers, I guess.
