What is the most difficult winnable endgame to learn?

I'd say K+Q vs K+R

Otherwise it is mainly K+R+P vs K+R+P when it is a winning position demonstrable by a strong bot

Is it more difficult than KQRRBBNNPPPPPPPPvKQRRBBNNPPPPPPPP?

I can usually do it in less than 40 moves against stockfish.

If you're talking about KBNvK, I can almost always do it in the shortest number of moves against Nalimov (or at any rate could when I last seriously practiced it).

We're talking about endgames ?

otherwise K+Q+P vs K+Q+P or K+N+P vs K+N+P should be more difficult but there's not much use or interest in learning them because even Gms don't seem to be able to find methods to get techniques to master them IMO

How many pieces constitute an endgame?

I don't really know but when there are no longer enough pieces forcing us to shelter our King and we can play our King it's an endgame I think.

So is KBBvK an endgame for both sides or just one?

It is position is an endgame or not and yes KBB vs K is an endgame.

Yeah I don't even think mine counts because stockfish for some weird reason always runs the king to the edge immediately lol, so it's no challenge. I think my best was around 37 starting from a really weird position and my worst was 41. I need to study Queen vs Rook more, so hard to remember all the possible tactics.

@Alchessblitz

Is that not contradicting itself?

I tried to explain simply in my own way starting with "I don't really know" after if you show me positions I can tell more easily if this position is an endgame or not.

I don't really know how to define what an endgame is but I know when it's an endgame for a lot of cases

There's Stockfish and Stockfish of course, but I've played it also against 8, 11 and 12 and it they were all pretty accurate (not perfectly) given anything more than ridiculously low think time. I don't believe accuracy is the best defence in this endgame, but it's what you get from Nalimov anyway.

QvR I would say is distinctly harder than KBNvK on average.

Neither do I. There are a lot of conflicting definitions.

I just take the view of the endgame tablebase creators that all positions are endgames, but some have more pieces than others, which, when they're closely matched usually makes them more difficult to learn.

Which begs the question of what I mean by closely matched of course.

For the purposes of the thread, the relevant question would be how does OP define an endgame, but he doesn't say - which leaves his question difficult to answer.

Yes, to a knight (on the sixth move). Possible candidate for your under-promotion thread?

Maybe lol, the underpromotion itself isn't that big a deal, but the endgame is!

Well, the under-promotion is necessary, but I wonder how many people would spot the mate in 549 OTB. Could be a candidate for hard to learn as well.

If their opponent doesn't play perfectly, it could drop to a win in 20 by the 3rd move..

Could be said of a large number of positions. It could also drop to a win in 20 by the opponent.

The assumption of the question, I think, is that you would learn the endgame to the extent you could play it against someone who could also play it (probably, in that case, Syzygy).

I would be interested in knowing why one move wins because of a 100-200 move sequence after that, if I could process and remember it all lol.