We are seeing a growing number of retirement advisers shifting from using return assumptions towards using required return (a.k.a., "Goal-Oriented Target") numbers.

The first is the investment equivalent of fuzzy math while the latter limits itself to hard data. Significant potentials fiduciary issues appear to be driving this movement (see "Retirement Planning Using Return Assumptions vs. Return Requirements – A Fiduciary Perspective," FiduciaryNews.com, January 26, 2016).

This shift, however, still has a long way to go.

Many retirement advisers continue to rely on return assumptions to create retirement projection models. I asked several folks who fall in this category to explain to me what components go into making return assumptions and what are the sources of these components.

As veteran professionals might guess, there was no consensus in their answers.

While such traditional sources as Ibbotson and Morningstar (now one in the same) were cited, what individual advisers do with that base data varied.

Some adjust the historic data downward to reflect a more conservative philosophy. Others prefer to take a more aggressive stand based on expectations associated with specific asset sectors.

Finally, it's almost just as popular for advisers to simply make their own assumptions, perhaps giving a subtle nod to the guidance of history, but that's about all.

And that, in a nutshell, describes the fundamental problem with using return assumptions and any program or process that invokes them.

There's a lot of guesswork, a lot of opinion, and, basically, a lot of (usually) unintended hocus-pocus that can skew the expectations of the person they're supposed to help.

And when professionals encourage high expectations, they also increase their fiduciary liability.

It's almost better to simply use the historic data in a way that leaves it historic. In other words, when you use return assumptions, you're implying that you are assuming those returns will occur.

Even the much heralded Monte Carlo method – safely nestled in its probability-laden outcomes – assumes some veracity to those probabilities.

Not to get all Schrödinger's cat on you, but all the Monte Carlo method does is tell you your retirement assets have both will fail to grow sufficiently and they will grow sufficiently at the time of your retirement.

This non-sensical duality (referred to as the "superposition of state" in quantum physics) exists until you actually retire.

In the case of Schrödinger's cat, the feline hidden in the box is both alive and dead until you open the box. Once the box is opened the cat is either dead or alive, but most definitely not both.

And another thing: once you open the box to confirm if the cat is alive or dead, you cannot close the box and return the cat to this weird dead/alive duality. Nope, once you know for sure the cat is alive or dead, the cat stays alive or dead. You cannot undo opening the box.

It's the same with Monte Carlo projections. You won't know if you've succeeded until you retire (i.e.., open the box) and, once you've done that, you can't go back and have a "do-over."

At least when you stick to real unadulterated historical data, you can say something like "since we began collecting real data in 1929, the market has met or exceeded an average 8% annual return for 95% of the rolling 40-year periods" instead of saying "the odds of meeting or exceeding 8% annual returns for a 40 year period it 95%." Notice the difference between the two statements.

The first is a fact. The second sounds more like an implied promise.

Which version would you stake your career on?

Don't let me assume your answer.