NoGo is a young game. The first mention that I could find is on Sensei's Library's page on Anti-Atari Go, Version 1, dated April 1, 2005, by John Moore. He claims: "We've tried this variant a few times in the Stupid Room on KGS". This version of the document does not specify whether suicide is forbidden, but we can assume that this is the same game as current NoGo.
One advantage of NoGo is that each valid NoGo game is also a valid (if very bizarre) Go game. This means that all the tools for developing Go programs can be used. For example, GoGui can be used as a GUI, and games can be played online on KGS (as long as the losing player resigns instead of counting the final score according to the Go rules...).
In January 2011, this game was selected as the featured game for the 2011 Combinatorial Game Workshop at BIRS.
Two NoGo tournaments wer held throughout the week: a human tournament on a 7x6 board, and a computer tournament on a 9x9 board. My MSc student Fan Xie won the human final against Tristan Cazenave of Paris.
The computer tournament featured six programs. Entries were written by Jean Mehat, Tristan Cazenave, JP Grossman and Olivier Teytaud. BobNoGo and FxNoGo, our programs based on Fuego, rounded out the field. The format was a double round robin, with each program playing once as Black and once as White. BobNoGo won all its games, and FxNoGo took second place. In a playoff between the human and the computer champion on a 9x9 board, BobNoGo defeated Fan Xie.
A series of computer games tournaments sponsored by the Taiwanese Association for Artificial Intelligence (TAAI) this year included a NoGo competition. Eight programs took part and played five rounds of two games each. The undefeated winner was Coldmilk, followed by PSABR with 7 wins. MoonGo and DJNoGo shared third place with 5 wins. The TAAI 2011 webpage contains both a report on the NoGo tournament and all game records.
Four programs registered for the olympiad NoGo tournament, but only three played: TAAI winner Coldmilk, and the two Fuego-based programs BobNoGo and FxNoGo. BobNoGo won ahead of FxNoGo. See the results on the ICGA webpage.
In the tournament games, BobNoGo was happy most of the time, with the evaluation slowly climbing. The most difficult game was against FxNoGo playing white. The program was below 50% for a long time, then slowly climbed up. Fan was very unhappy when his program played move 29 at C1, instrad of building an "eye" with B2. Running 10 million simulation searches on my laptop, BobNoGo's winrate for move 28 is 0.488, while for move 30 it is 0.493. At move 34, the evaluation finally climbs above 50%, then jumps to over 53% for move 36. Possibly, the game was decided around this time.
In BobNoGo's other three games, the evaluation was above 50% (except the first few moves with white). In all four games, the solver announced the proven win pretty late, when the Monte Carlo win rate was already above 90%. I think the solver can still be improved a lot. I am not sure how much the solver is worth earlier in the game, i.e. what is the effect on the Monte Carlo search if it knows that some deep lines are wins, or losses.
CMNoGo is Coldmilk NoGo.
Held July 1, 2012. Four programs, double round robin. BobNoGo was the undefeated winner.
BobNoGo was finally dethroned at the 2013 Olympiad and took third place. HappyNoGo was the winner. See the results. A report is: W.-J. Tseng, Jr-C. Chen, P. She, I-Chen Wu, "HAPPYNOGO Wins NOGO Tournament", ICGA Journal (SCI), Vol. 36(3), September 2013.