My goal for Romi has been 70% vs Hamsters 0.2 ever since Hamsters 0.2 was released. Romi started out at about 52% against Hamsters 0.2 and has made choppy progress. About a month ago (give or take, I have really no idea) Romi 32 bit scored 69.5% and I thought that just one more minor tweak and I will have it. No way!
Everything that I tried only made things worse. 
My frustration meter was pegged.So, a few days ago (give or take, I really have no idea) I got a new idea. The first expression of this idea resulted in Goatka, a half joke half serious attempt at an improvement. But, like Tony says, "it has goat powers." Therefore, I have been working hard to improve upon the idea.
It was not going well and I was about ready to give up, when I then noticed a bug that had crept (that's what bugs do) into the code. What a miracle finding that bug was!
A quick time control, 80/1, match is proving this out.A little history is needed here. When I first wrote RomiChess I tested almost exclusively at 80/1 repeating and got very good results. As modifications were made the 80/1 performance started to suffer and so, I started testing at 1+1, then 2+2 and finally at 4+4. RomiChess NG has never scored above 55% at 80/1 time controls vs Hamsters 0.2.
To make a long story short. The 80/1 match is not over yet, however, the score is Romi +45 -16 =9, 71%!
If Romi is doing this well at 80/1 then I know that she will do better at 1+1 or higher! 
