Actuall,y having a closer look at the analysis (now I'm more used to how to work it) I think stockfish sees the above position as very drawish, leading to a draw in 20 moves when I play stockfish against itself.
This position is similar parity, but it takes 112 moves to reach a draw when stockfish plays against itself:
Have you looked at whether or not it works better to switch the colors around or leave them the same, or is it more balanced the way it is
Have you looked at whether or not it works better to switch the colors around or leave them the same, or is it more balanced the way it is
Both of the postions I've shown are 0.0 in the analysis tab or thereabouts (it seems the anaylsis changes when you move it to a different postion and move it back again and I don't understand why).
It's possible it is more balanced with switched piece colours but I don't think this engine is going to tell me that. I could check to see if it's less drawish, though.
OK, stockfish seems a little...confused about this position, but sees it as balanced if you leave it (initially it records a 0.5 advantage for black or something but changes to a 0.01 advantage for white).
I guess the fact stockfish is confused probably means it's less drawish, which is good. I'll test that.
Or at least, it did. I put the position right back into stockfish and it thinks it's +0.64 for white now. I think this position really confuses stockfish.
Playing stockfish against itself on this position gets me checkmate for black in 98 moves.
So perhaps it is less drawish than the white horde positions, but also perhaps a tad less balanced. I dunno.
I realised using stockfish, I hadn't been leaving srockfish enough to fully analyse it. I'll try these positions when giving stockfish the time to fully analyse each move and see what happens.
OK, after a more patient engine analysis, the first postion is technically an 87-move draw, but every move from move 15 onwards is a white king move: i.e. it's just the engine finding creative ways to avoid a draw by repetition.
So yeah, I think it does seem position 1 is too drawish.
A more patient analysis in position 2 leads to stockfish checkmating with white in 69 moves.
Position 3 is a draw in 32 moves according to stockfish.
Based on engines alone, position 2 is looking the best, because there's a relatively long game and it's not too drawish.
But it does depend on how transferable engine evaluations are to human play in such an unorthodox position.
I have now analysee two more positions, playing full depth against each other.
In this position:
Black checkmates in 98 moves, despite it giving an initial +0.5 rating for white.
I had to save the full game for this one, I only had a quick look at what the game was doing (as I was more interested in which position is more balanced) but it includes a knight promotion which sticks.
I have already done this on the main varients forum, but no-one replied, and also I have updates.
So basically I am trying to make a version of horde that is playable on chess.com custom daily chess games.
This means it needs a white king, and I also recently I found you also must remove the back row of pawns; because without this chess.com will not recognise the FEN.
Since the original thread, I have messed about on the chess.com analysis with the engines turned on. According to stockfish, I think this is the most balanced position:
Surprisingly, this is actually the same number of pawns as standard Horde, if the king is to count as a 'pawn'. I guess the disadvantage of being checked cancels out the extra powers of the king pretty closely.
But just because stockfish says it's balanced, doesn't mean it actually is. Stockfish isn't neccessarily good at such a non-standard position.
My question is: how balanced is this arrangement, and what would a more balanced version of Horde (using chess.com daily setup rules) look like?