When should opening study start?

There are a lot of questions being asked on opening theory on the forums here, and some of the responses coming from the stronger players are along the lines of "Beginners should not be studying openings" or "At your level you shouldn't be studying opening theory".

So at what level are you supposed to be studying opening theory?  Maybe I'm confused as to what a "beginner" is?  Is it someone rated 1000 on Chess.com or 1000 OTB?  Or both?  Or maybe these players mean that you shouldn't be studying trees of opening theory without understanding what the plans are behind the moves?

It's not that I doubt the playing strength of someone 700 rating points higher than me, but sometimes I think that the strong players give a response for the very same reason cyclists shave their legs: Because everyone does/says that.

I would really like to hear some thoughts on this

When you start losing because of the opening would be my best guess. 

As soon as you realize that your assesment of the position in the opening (which you may not even realilze that you are in a well known opening) you are currently playing is wrong.  That there are thousands of games in data bases that have gotten to the same position and the 2 or 3 correct continuations are well established chess theory. 

It is fun to give flight to your imagination as long as you don't mind losing time after time after time.  This is known as the school of hard knocks.  Eventually your mistakes in previous games will lead you to better moves that are more in keeping with established theory.  But you can save yourself alot of pain and losses by reading some books on openings so that you can learn from other chess players' mistakes in the past.  

Opening study should ideally be started when your child is 4-5 years old.

Thanks Shadowknight911 and ciljettu!   I think you have answered my question

I more or less agree with shadowknight BUT one caveat. I disagree with the old classic idea of studing the endgames and working backwards. That theory came out of (and works great) russian chess schools where players were working and expected to work on becoming very strong players. For the vasts majority of players a complex rook or king pawn endgame rarely occurs in a game so the time spent on studying them is really better utilized on something else. Middlegame plans yes and basic opening traps, lots and lots of tactics. 

I have felt this way for a long time and was felt validated when IM Silman came out with a book expressing the same view. In his endgame book he shows which endgames are important based on rating (studying 1 level ahead is good but more than that is unwarranted) 

if your rated 1000 know how to mate with queen and king and basic pawn opposition but other than that work on tactics. Lucena and Philidor Rook endgames are interesting but worthless if more of your games are won or lost because a side drops a rook...