Small update on the chess engine stuff
The Emulator/Adapter setup seems to be a mixed blessing.
Most of the engines aren't any fun to watch playing matches because they don't output what they're thinking. The exceptions is the MM5 by Ed Schröder. It outputs the evaluation, search depth, and a two move current variation. Cool
The fact that these engines don't output their thoughts is inconvenient in another way as well: the emulator isn't perfect. Sometimes an engine crashes (MM4 and Rebel 5), and then the GUI just sits there, waiting for a move, until the time is up. Rebel 5 even crashes the entire emulator when you unload the engine.
Roma32 (which doesn't output anything apart from its moves) seems to be the strongest engine in quick testing, closely followed by MM5.
I'm curious to see how engines like the MM5 and Roma32 fare against the Turing engine. That one implements Turing's
Paper Machine. (Strange. Chessbase has an article about it in 2017, but the engine, created in 2004, can only be found through the way back machine. I have had it archived for at least a decade though
)
Turing, on a current day computer, can run up to 8 moves deep, just as MM5 can (and probably, Roma32). The difference is that Turning's chess knowledige is obviously very, very limited, as it was written before even the first computer was invented
I wonder what its results will be against MM5, and how MM5 (1990) will perform against something like Fritz 5.32 (1997).