FWIW, in at least some versions of Adobe software, the "warnings" can be turned on or off, and they might happen to be turned off at the moment you need them, so I don't consider this a foolproof method. In other software, clipping warnings might not even be available, so I usually rely on the presence of a tall, narrow spike at either end of the histogram as the way to tell if I have an appreciable number of pixels that are either blown, or are a featureless black. I consider this to be the most important, first-thing-to-look-for when evaluating a histogram.
WRT to not seeing identical R, G, and B histograms, the only subject / settings that will give you such a histogram is a perfectly gray photo. This recommendation is nonsense. It is definitely NOT something to strive for in color photography.
WRT MIke's suggestion for a flatter histogram, this is essentially the modern version of the old advice that for good negatives or prints, one should have "something in every zone" (ie, when using the decades-old "zone system"). However, this advice must be taken with a large grain of salt, because there are plenty of dramatic, award winning images with large blocks of uniform tone/color. The histograms for these will be very jagged / spikey. Having a nearly flat histogram maximizes the mathematical information content of the image, and can look very, very good. Some of the classic "Look" and "Life" magazine B&W covers, famous Ansel Adams prints, etc. were in this category. However, a flat histogram can also be produced by an image containing nothing but noise / grain.
A reasonably flat histogram is neither necessary nor sufficient, nor even a predictor of a good image. It's more important to use the histogram to help one quickly check that no area in the image that has visually important content isn't squished up against either the LHS or the RHS of the histogram. On the other hand, long before histograms became available, the same could be done with a hand-held spot meter, LOL.
HTH,
Tom M
