David Daiches, Scotch Whisky: Its Past and Present, London, 1969, p130:
But later experience of comparing different ages and proofs leads me to believe that additional age over twelve years does not add all that much in quality, and (within limits of course) a twelve-year-old at a higher proof tastes better than an older whisky at a lower proof. But the twelve-year-old is decidedly better than anything younger
I wonder if, at the time Daiches wrote this, there was less demand for quality casks? (99% of casks went off to be blended, and the matured whisky he was trying at the time had come out of the post-World War II years). In other words, lots of whisky was going into first-fill casks, and maturing at such a rate that the whiskies were peaking at 12 years old. Being first fill casks, any more than 12-15 years, and perhaps some of them were getting too woody and past it?
Today, however, global demand for whisky dictates that casks have to be re-used far more than they may have been in the past. Much of the stock now being released as single malt is now maturing in second and third-fill casks, which means the rate of maturation is slower, and some of these whiskies may actually be peaking at older ages, i.e. 15 to 25?
Just speculation, of course, but the theory holds up!