I've posted about this misunderstanding before. I don't mean to jump on you in any way, Gearhead, but this is one of those myths that seems to have taken on quite a life of its own. I believe it stems from the desire to avoid overload conditions. Specifically, if one boosts too many filters to too great a degree, one can cause the next stage of amplification to overload if there is no independent gain adjustment following the tone shaping.
Overload conditions aside, because most players use fixed (or parametrically adjustable) filters to achieve tone shaping, the best rule of thumb is to apply the smallest amount of gain or cut that achieves the response that's desired. This will result in the smoothest spectrum.
For example, suppose you need a small relative boost in the mid-frequency region (i.e., 1-2 kHz). You are far better off applying a small boost in that region than you would be by cutting the surrounding regions. Doing the latter would leave you with a bit of what is referred to as a "rippled spectrum."