Big upset in BYOB

A top algo was beaten by a trash can algo I submitted to the BYOB competition in Round 2.

🔜🔝#️⃣2️⃣ is struggling at 1740 elo and @Felix’ algo Felix_0.9.3_Advanced_Maze_2 is already at 2091 elo even though it was just submitted a few days ago.

In the above match from Round 3 against @n-sandersMazeRunner_1.1, you can see why I consider my submitted algo garbage. The idea of it could definitely be enhanced in defense because it is very inefficient, but I am not sure if it is the best idea to let bits decay.

Other notable defeats in the first three rounds are PythonSmasher3D_v1.1 and Track-6.2. The opponents were highly ranked algos as well, hence not to be considered upsets.

If you look at the recent leaderboard history, “Track” by @RuberCuber is one of the best performing algos lately (colored in orange and situated at the top). It was beaten by the exact algo that won the last competition (ActorV6).

1 Like

I find ActorV6 fascinating.
It clearly has strong performance in single-elim competitions, but on the global it’s sitting at 2073 elo in 19th place. And it’s record over the past 100 matches is 56-44. I don’t meant to discount being top20 at all, but it’s interesting that a 60/40 win rate in the global has translated well in the single-elim matches.

Edit: perhaps it’s explainable due to a relatively small sample size of single-elim matches.

Also, @876584635678890, I love your algo’s name :rofl:

1 Like