Why do they say study the endgame first?

If you don't know how to turn a winning position into an actual win - basic checkmates, King and pawn endings, dull stuff like that - then you may well end up playing tactically and strategically superior chess but still have no idea how to close out the game. This is why I keep seeing kids who've played their way to a position where they're a Rook and two minor pieces up (or, at any rate, blundered away less of their material than the opponent) and are now wandering around the board continually checking the other guy's King and hoping they're somehow going to stumble on a checkmate. Maybe it's fun to play that way but if you've any ambitions to become a better player, you need to start at the end and work back. What shall it profit a man to be a Queen up if he doesn't know how to mate with King and Queen?

Exactly! Here is a game: https://www.chess.com/live/game/2027555431

So first I made a mistake and lost my knight in K+R vs K+N after 96.Ka5. Normally I would immediately resign in such position but I got this feeling that my opponent was not a great endgame player. And he has proved it! In K+R vs K he couldn't deliver a checkmate and I've claimed a draw on 50-move rule.

The point is not to get the exact same position in the book.  The point is to understand the endgame concept and employ it in all relevant cases.  Also, having the knowledge of how to win certain endgame positions factors in to the evaluation of trading pieces in the middle game.

Let's say you open up a beginner's endgame book, and it explains that with a White King on d6, Black King on d8, and White pawn on d5, no matter who is to move, White wins, but move everything back 1 square from White's perspective, and it's only a win for White if Black is to move.

Now, this does not mean that this rule only applies on the d-file.  This applies to any file except those featuring a Rook Pawn, so 6 of the 8 files the same rule applies, except in the case of a Knight pawn, you have to be careful as only 1 move wins if it is White to move, not 2 moves.  Going the other direction leads to stalemate.  The King should go to the outside, not the inside, in this case.

Same thing goes for R+P vs R.  Same rules apply for all pawns except Rook Pawns.  So seeing an example with a d-pawn doesn't mean you need that exact position to apply the knowledge.  You could have the same basic position with an f-pawn instead of a d-pawn and the same rules apply!

What does it profit you to play 148 moves against a player who is so poor that he cannot win with K and R vs K??

To become stronger you should try to play against stronger opposition.

It never happens to you to play against someone with rating equal to yours?? When we played this game we both were around 1330 in blitz. I guess if he had so poor endgame skills it means that he was superior in other aspects of the game.

Even if what you say is 100% true--what does it gain you to play a ending where your opponent cannot even mate you with K and R vs  your lone K?

I am guessing it was just the satisfaction of drawing the game and showing your opponent what a poor player he was?

That is ok, I guess--maybe your opponent learned something about chess?

Yes it's exactly this - satisfaction and thrill of drawing in a lost position. Do you suggest that I should have resigned instead? Would you resign a theoretically lost position if there is a tiny chance that your opponent doesn't have a skill to win it? 

I just learned the Lucena Postion, which, had I known it a month ago, could have led to a win in a game I ultimately lost because of poor endgame technique. How does one queen the pawn? (I've found it doesn't matter what black does here. With proper technique, you can queen the pawn.

Reminds me of this one time I accepted a rated challenge from someone hundreds of points below me and... let's just say he played out of his skin and not make invidious assumptions. We reached an ending with K+R+P each and I was looking at a massive rating hit for drawing against such a low-rated player. However, I figured out the one line to give him trouble if he defended poorly, pushing my pawn until he had to give his Rook for it. With better play he wouldn't have had to but he must have thought I had no way to win with only a Rook. He was wrong. Rating rescued!

Gil, to answer your question--yes, in the circumstanes you had i would have resigned.  this does not mean you were wrong in not resigning--we each have our own reasons for what we do or what we would do. 

I know that they are K+R+P vs K+R endings and I never reach them.

insighked2

thanks for showing the Lucena Position. Am sure a lot of players learned from that.

Is 1 Ra2 an alternate way to win?

It is. The Lucena method is the only way to win with a pawn on the b or g files.

THEY DO

I pulled it right out of Silman's endgame book.

You should study everything. It is pointless to study only endgames, you will get destroyed in the opening or middlegame. 

I understand this idea and why people agree with it. But my games are often so one-sided that by the time an endgame arrives no technique is needed to win other than simply not blundering. In fact it's said that most amateur games are ended by blunders. Notice here that no one ever says that these blunders only happen in a specific part of the game.

And it's not just some random idea that floated around that learning endings first isn't a hard and fast rule. GM Jan Gustufsson and GM Peter Svidler have said that endings aren't as big a deal as people make them out to be. They are important for sure but I don't think it's practical to learn any but the most basic for a time. I've heard GM Hikaru Nakamura often talk about how CMs are mostly good at tactics and not much else. And it's pretty obvious when you think about it. If you blunder in the opening or middlegame, you probably won't have much solace in the ending anyway. Anyway I've already shared my opinion before that tactics are the most important thing for amateurs by far and yeah... endings are important in chess but I don't see all but the most basic as practical.