If you reach endgames with material parity in less than 10% of your games (ie a beginner or lower intermediate player -- sometimes much less than 10%), is it still worth studying endgames?

Yes, but only very basic ones like how to mate with king+ queen vs. king and king + rook vs king. Actually a good reference for this kind of thing is Silman's endgame course.

http://www.amazon.ca/Silmans-Complete-Endgame-Course-Jeremy/dp/1890085103

Endgames are incredibly important and will spring your play from intermediate to advanced in almost no time at all.  The point isn't the endgame technique... it's knowing what type of exchanges favour your middlegame!

Only if you seek to improve.

"Intermediate to advanced"... so what about for beginners, as in the original post.

I say no -- for real beginners, nothing beyond basic overpower mates will be of much immediate use.

 Yes, because you will learn how the pieces work together and cooperate. Chess is almost best learned in reverse. As the beginner progresses s/he will start to see how simple theoretical positions (K vs Pawn/opposition) relate to more complex theoretical positions (Lucena, R & 2 connected pawns vs R, etc) and then how the more complicated thoeretical positions relate to technical endgames and then how the technical endgames relate to the late middlegame and so on. It is much harder to go from the complex to the simple.

You're right, I misread the original post...

It still applies to some degree though, In the (admittedly limited) coaching I've done I've made it a point of demonstrating basic mates followed by king/pawn and then king/rook endgames... only then do I set the pieces up at play a game of chess.

However I agree... for anyone under about 1000online rating (800 ELO), mating is all you need to know for now... focus on not hanging pieces first.

One nice strategy is to become an endgame buff and just trade off in the opening and middlegame :)

 That's still my focus and I am just over 2000 on ICC! :-)

the BEST way to learn chess properly is to start with endgame techniques - fewer pieces so 'simpler' in a way. Certainly this is how all the soviet masters/grandmasters were trained from a young age.

Most of the things I was going to say were said, but I vote for studying endgames for several reasons. First, even though I'm a beginner I do win some games if only on my chess computer. So I need the skills and insights of the endgames for those situations. Plus I think it helps to see where you're going during the previous part of the game.

Learn how to stop dropping pieces randomly first before studying the endgame. For example, http://www.chess.com/livechess/game.html?id=21618196 20. Rd5? loses the game at once (you where losing but still...)

You should still know how to mate with only a rook and how to win when you are a knight up or so...

Endgames are more or less 100% tactics. Being good at endgames helps the rest of your game as well.

The answer is simply YES.  Just starting with the easy stuff like mating with a rook or two bishops.

Mating with two bishops? I'd say thats more an intermediate endgame.

Yes--knowing mating patterns is key--knowing how to take an advantage such as an extra piece and be able to convert it to a win (like forcing a pawn to Queen), as well as understanding how to either Queen a pawn or prevent your opponent from queening are all key features in one's chess understanding

Yeah, mating with 2 bishops isn't so easy.

And endings are 100% tactics? That's a new one on me :Z

I think that endgames are important, but one can advance much better if only solves a few midlegame problems every day...try it and you'll see...

Maybe the question should be rephrased into something more like:  "with only X number of hours per week, how much time should a beginner devote to endgames?"  The problem in this post is that one group will be arguing for learning more and more, while the other group argues it's not necessary for that level, back and forth ad nauseam.

If X = 4, then I'd say spend all your time playing and studying tactics.  Skip the K+2B v K (which I've never had in the thousands of games I've played in my life!), skip the K+B+N v K (rather difficult and also rare), but learn K+R v K, K+P v K, and of course K+Q v K.

On the other hand, if X = 10 or more, learn them all.