I don't think a system that can promise you 3 standard deviations of improvement has been invented. Look at twin studies.
I agree that 3 standard deviations of improvement of a random person is a lot to ask for. However, I can easily see that someone with built in potential to be 3 sd above average could be brought down to near average by the wrong system.
My expectation of my children's potential is very dependent on how heritable intelligence is, and I admittedly do not know much about that.
If it's worth saying, but not worth its own post (even in Discussion), then it goes here.