A last comment since this fun while I am playing a game and waiting for moves.

Even though I have never really liked ponder moves with dedicated computers because of their small horizon. Even with 20 Mhz you are lucky to get a search depth of 5 ply most of the time in middle games. So when it ponders and moves instantly I always felt it could avoid some of its instant move disasters if it had have calculated the next ply but missed that opportunity.

Using CB-Emu this becomes however an even bigger risk. So for example
Computer A now gets only 10 seconds to ponder. Its search now instead of 1000 positions at 30 seconds becomes 333 positions at 10 seconds. Whereas the 3x computer remains at 3000 positions. So what started as a 3 to 1 advantage now because of pondering is an almost 10 to 1 advantage while they ponder.

So I think there is a bigger handicap then we sometimes realize?