Which openings lead to which types of endgames

I was having a conversation with a friend of mine recently and I was saying "I need to get better at rook endgames. Don't you find them tricky?" and he was like "Huh? I hardly ever play endgames!"

We figured out one factor was I tend to trade down much more than he does. Another factor that just hit me is I almost always play the Caro-Kann, and he doesn't. With good play from both sides, many Caro-Kann games seem to reach a rook and pawns endgame.

What are the types of endgames that the other openings tend to reach?

First, it sounds like you trade too much.  Do you leave pawn tension on the board or trade pawns ASAP without thinking beyond whether or not it drops material?  Same with pieces.  Do you just trade to trade?  Initiator of any trade loses a tempo.

That said, I have seen the following a lot in OTB game by myself or KassySC:

Accelerated Dragon Maroczy Bind - Double Rook Endings, often with 7 pawns each, White no d-pawn, Black no c-pawn.

Exchange French - Knight Endings

Berlin - Various combinations of single minor piece Endings (Same color Bishop, pure N, or N vs DSB favoring White, N vs LSB typically equal, OCB either equal or favoring Black, and B vs N typically favoring Black - all of those, White having the first named piece, Black the second) and R+Minor Each.

Rook endgames are generally the most common endgames in most of the openings

You might take the results of those that happen to track them and compare to an overall endgame frequency table...

https://www.chess.com/blog/NimzoRoy/endgame-percentages

With possession of this database or an even bigger one it must be possible to identify the most common openings and the frequency with which they lead to those categories of endgames.

Once I have played 100 Caro-Kann games that reach an endgame I will compare my results against the standard endgame percentages.

Rook endgames are the most common endgames from any opening. A rook covers 14 squares regardless of where on the board it stands. Queens, Bishops, and Knights are much more powerful if they are in the center. So putting a queen, bishop, or knight in the center is usually enough to compel the opponent to trade it. That leaves a rook ending. Now in a rook ending one side will benefit from a trade into a pawn ending and thus his opponent will avoid that. That is one reason why rook endings are most common.
Another reason is that most pawn endings, knight endings, same color bishop endings and queen endings with an extra pawn are won. Rook endings with an extra pawn are often draw. So the player having one less pawn seeks salvation into a rook ending.

There's even an old saying to the effect that "all Rook endgames are drawn". That's something of an exaggeration. of course.

#7
Yes, but the better player always win them.

Actually, not a true statement at all!

Yes, if you take every game in the database, Rook Endings are the most common, but he asked on an opening by opening basis.  Rook Endings tend to be quite rare in openings with a symmetrical pawn structure with exactly one open file, like the Exchange French, Exchange Slav, certain lines of the Petroff, etc.  They get traded off on that one open file, ESPECIALLY the Exchange French (I know because I face it often here as Black and probably 2-3 times OTB with Black).

For the Exchange French, it is almost always some kind of knight ending.  Either a pure knight ending or something like BN vs NN, the Bishop usually being the White LSB or the Black DSB.

As far as being the pawn up, you are right in most cases, but wrong in one.

Pawn Endings - pawn up usually wins

Knight Endings - pawn up usually wins

SCB Endings - pawn up usually wins

Rook Endings - pawn up wins in certain circumstances, not others 

Here is where I differ from you.

Queen Endings - Quality means more than Quantity.  You can OFTEN draw down 2, 3, or 4 pawns!  All it takes is 1 monster pawn and quantity means nothing.  I recall a game I played over the board YEARS AGO (do not recall the exact position) where I was Black, it was a Queen ending, but White had a pawn either on e6 or d6, his Queen was aiding the pawn on something like e6, d6, e7, or d7.  He only had 3 total pawns on the board to my 6, but all my pawns were on the 5th, 6th, or 7th rank (needing to get to the first rank to promote).  There was no way to outrace or even stall his advanced pawn, and so I had to settle for the draw by perpetual check of the White King.  I was up 3 pawns, but had to settle for the draw.  This is not unusual at all in Queen Endings.  Quality over Quantity.

The other case where pawn count can be irrelevant is B vs N.  Now here, 3 pawns will win, but 1 might not, especially if:

1. The side with the Knight has the extra pawn but the side with the Bishop has the more advanced King and his pawns on the color of his Bishop to build a fortress.

2. The side with the Knight has the extra pawn, but he has a backwards pawn or blocked pawn at the base of a pawn chain on the color square on the Bishop.

3. The side with the Bishop has an extra pawn, but one of his pawns is a Rook pawn with Opposite color promotion square of the Bishop, then all the side with the knight has to do is eliminate the other pawns and even sacrifice itself for the second-to-last one, leaving Bishop and Rook Pawn.  For example, the following:

(Having problems with diagrams today)

White: Kd4, Nd3, Pb2, Pg3

Black: Kd6, Bb7, Pa6, Pf5, Ph7

Now, this specific position may be a win for Black, I have not analyzed it, but if White can force a trade of 1 set of pawns, a for b or f for g, even if indirectly (Knight or king attacks a Black pawn while Bishop or King attacks a White pawn), then White can sacrifice the Knight for the other non-h-pawn if the King can get to the corner.  I have drawn this way down a pawn with the N in B vs N endings more times than I can count.

So yes, Rook Endings are not the only high-frequency drawable endings.  Q endings and B vs N endings often can be as well.

Btw, it seems to be an interesting question.

#9
This position you gave

is a draw indeed. It is a case of good knight versus bad bishop. With the black bishop at c7 it is a win for black.
Rook endings are often a draw with 1 or even 2 pawns extra for one side.
Indeed openings with 1 completely open file usually lead to rook trades along that file and thus not to a rook ending.
This is a paradox: openings with a drawish reputation like French Exchange, Petrov, Slav Exchange avoid the reputedly drawish rook endings.

Take an extra step, as in Thriller's post, and identify by variation/pawn structure also. Otherwise the results won't be as clear.

It strikes me as a rabbit hole, first get % from Caro-Kann, then wonder "maybe in the advance variation is different than the exchange variation?" and then "maybe the bronstein leads to different endings than the Korchnoi?" and then before you know it you are wondering what effect move 10 has on the ending and are working with tiny sample sizes that can't produce reliable answers.

But don't get me wrong, thriller's analysis is very good and useful.

I appreciate it.

You can compare it to the larger sample size in a database. Of course, the quality of samples is also a question.

You can go down a rabbit hole, a little or a lot. You could find something, but you may not.

Perhaps not a bad idea. Maybe even earlier? Move 3 too early? Move 2?

For example:

1.d4 d6 2.c4 e5 3.dxe5 dxe5 4.Qxd8+ Kxd8

Should White consider what might happen in an endgame? Has Black?

#13
Take a large data base and look at which endgames result from which openings by ECO code. That should answer the question.
You have 50 classes of openings A0x to E9x.
You have 15 classes of endgames
Q-Q, Q-R, Q-B, Q-N, Q-P, R-R, R-B, R-N, R-P, B-B, B-N, B-P, N-N, N-P, P-P
Your 50 x 15 matrix with numbers of games provides the answer.

Black actually scores over 50% in this line.  The problem is the c4-pawn gets in the way of White's pieces, particularly his LSB.  Black gets free development and the King can go to c7.  Against 1.d4 d6, White should play 2.e4!