Since the Discourse has the power of a black hole for me, here are a couple interesting ideas I have heard about the process of playtesting.

Firstly, the concept of a "playtest budget" is a fruitful one. As a designer and/or a publisher, figure out how much playtesting a given design will need. That can by anything from no playtesting to five years of testing depending on the designer & game.

Secondly, there are a number of portions of games that can be dealt with outside of a playtesting format. If you want to know how your complex dice probabilities work? You might be able to figure that out via stats, brute force, or https://anydice.com/ 

Third, the most powerful and useful playtesting in my experience relates to player engagement, emotional reactions, decision-making, emergent properties, and points of confusion.

Fourth, playtesting can be an exhausting process in terms of time, emotional energy, and actual cash. There is a point of diminishing returns after which it is not worth playtesting more. Sometimes that point is day 1!

Fifth, playtests can be a powerful tool to force a designer to look more closely at the points of friction and unnecessary complication. If you find yourself overburdened with explaining something or reflexively skipping it at the table "this time", nuke it from orbit.

Sixth, playtests are also often a stealth form of marketing. Playtesters often become invested in the game when they help shape it, so they can double as game demos.

Seventh, playtests are often most useful when you intentionally identify specific components that you are trying to get more information about. "You are in a car chase because I need to see how those rules work in practice. ". It doesn't have to be an "authentic" game session.

Eight and final? The excellent @snarkbat proposed a new term we should use for playtesting without the designer present. "Unguided Playtesting" instead of the more problematic "Blind Playtesting". Let's normalize it.