No idea. Here's a try with 19.5
You have to make 15 captures, and you need at least 2 moves worth of non-captures. So without any checks there's 17, so 19 and a half is not so far off. But I feel like there should be a faster way. Maybe using the rooks in some way?
18.5, still not sure what fastest is though
Good try, but not even close.
Chess problem maestro Sam Loyd devised the fastest way.
thanks for answers. very amusing games. does anyone have any knowledge of how fast a " reasonable " game might end with two kings ?
Wow, 17 moves!
This is actually a very useful question and solution for me because I'm working on the math of the "denudation rate" of units on a chessboard, so AussieRookie's post seems to determine the fastest possible rate of that parameter. I intend to post my results within the next few days. Thanks for all contributors to this thread.
thanks for answers. very amusing games. does anyone have any knowledge of how fast a " reasonable " game might end with two kings ?
edit...
"The fastest possible" and "reasonable" conflict for sure. What's reasonable to you may not be reasonable to someone else.
In real games where there's still fighting in the endgame most don't last much longer than 60-70, but some openings you exchange a lot early. So I'd guess around 30-40 moves for what I'd see as a reasonable game.
I just looked at two different games on 365chess that ended in draws, probably played by masters, plus the Sam Loyd composition as an extreme case, which resulted in three data points for denudation rate:
(#1)
http://www.365chess.com/view_game.php?g=3641503
Ruy Lopez (tends to lose units slowly)
32 initial units - 18 final units = 14 units lost
14 units lost in 41 moves = .314 units/move
(#2)
http://www.365chess.com/view_game.php?g=2441668
Austrian Defense (tends to lose units quickly)
32 initial units - 18 final units = 14 units lost
14 units lost in 25 moves = .560 units/move
(#3)
http://www.chess.com/forum/view/general/earliest-possible-king-v-king-ending
Sam Loyd (the most extreme case; theoretical limit)
32 initial units - 2 final units = 30 units lost
30 units lost in 17 moves = 1.76 units/move
(#1)
Ruy Lopez
41 moves until draw
32 initial units - 18 final units = 14 units lost
14 units lost in 41 moves = 14/41 = .341 units/move
(#2)
Austrian Defense
25 moves until draw
32 initial units - 18 final units = 14 units lost
14 units lost in 25 moves = 14/25 = .560 units/move
(#3)
Sam Loyd composition
17 moves until draw
32 initial units - 2 final units = 30 units lost
30 units lost in 17 moves = 30/17 = 1.76 units/move
x moves * denudation rate in units/move = 32 initial units - 2 remaining units = 30 units lost
Solve:
x = 32 / (denudation rate in units/move) = 30 units lost
(#1)
Ruy Lopez
32 units / .314 units/move = 102 moves per game until only 2 kings remain
(#2)
Austrian Defense
32 units / .350 units/move = 91 moves per game until only 2 kings remain
(#3)
Sam Loyd composition
32 units / 1.76 units/move = 18 moves per game until only 2 kings remain
One problem is that this rate is not linear, although I believe the constant involved in describing the curve *is* constant, which is the problem on which I'm working now.
Does anyone know what the earliest possible time the game can end in a king v king position?