In Western chess, we have accepted that computers outplay humans, full stop. (Some might say we’ve resigned themselves to this fact. Please hold your applause; I’ll be here all week!)
Shogi (Japanese chess) is quite a different beast. Western chess revolves heavily around material balance. If you’re down a Pawn, you’re expected to have a significant advantage in time or position. Being down a Knight, or even two Pawns is hopeless in a typical position. In Shogi, material is not as pressing. In fact, unlike chess, if you’re down in material, you probably want to exchange pieces.
That’s because of the coolest feature of Shogi…piece drops. In Western chess, a captured piece is out for good. In Shogi, when you capture a piece, it can be returned to the board under your control, with few limits. There are many more opportunities for positional exchanges of pieces.
With regards to the computer, dropping pieces greatly expands the number of possible moves, reducing the effectiveness of brute force searches. The board is always full, so there is no “endgame” with just few pieces. In Western chess, computers can use endgame tablebases to play positions with few remaining pieces perfectly…if they even have to play that far.
But these differences are starting not to matter, as computer processing power and improved programming have finally defeated a current Shogi master.
I’m declaring the year 2020 as the over/under for a computer to play Go at a master level. And Go is NOT easy to tell computers how to play at all.