Is the Maze Algorithm still an effective strategy? I mean, the static maze algorithm has gotten me to top 40, but I have trouble improving it. This is probably because given how the defenses are configured, your attack options are really limited. Even with simulations that I am doing, it doesn’t really help. Is this a bottleneck of the Maze layout? I have to ditch the Maze layout all together to improve with an adaptive algorithm?
No, definitely not.
There is multible examples of good maze algos in the top 20 and one of my secondary algorithms is one aswell (would be ranked around 8.)
Does your maze algo have the option to adaptively switch sides? That should improve it quite a bit and there are other things that make mazes better, but you’ll have to figure some out on your own
Heres a replay of my maze algorithm against Aelgoo (not the best Aelgoo though):
Now I don’t recommend waiting for like 3 rounds to place anything, I do that out of other reasons.
Also @RuberCuber has used mazes for quite some time now, maybe take a look at his track or guide series?
I hit top ten at one point at time with static maze.
I got to #15 after adding some gun detection and proactive defenses to my static maze algo, if you’re at #40 that’s probably you are probably fighting against a lot of corner gun algos.
Static maze algo’s are really easy to develop from the starter kit, but to get into the top 5 one probably needs to make everything dynamic and search for the best moves.
I found success with maze algos early on, and managed to improve on it for quite a while in my Track series, it performed pretty well in season 1 and even won the recent sentdex competition. Currently in season 2, the meta has shifted and my Track algos aren’t performing as well (currently at #17 on the leaderboard, but shifts between #10 and #20), probably due to many algos having layouts that counter mazes. Maze algos are definitely easy to make and can be decently effective against non-adaptive algos, but to get into the top leaderboard you will have to use adaptive strategies like switching sides, delaying building the front wall, or even switching layouts entirely (my Track algos do that when facing ping cannons).
How long was that ago?
How long ago? Months ago
Static maze is great practice to get you started, as there are realy a few things to improve on.
After that dynamic mazes (switching sides, protecting corners, sending pings) are the next great thing.
Dynamic kill all noobie` strategies as they crush the front 2 lines are really efficient int resources.
Most advanced algos just don’t build on the first 2 lines
wefw ranked around #13, has done really well with a standard maze. A string of losses for me as sending out too many scramblers which work well vs Demux but not against mazes. its really important to open holes in the maze as a front line emp attack across the board is disaster for a Demux setup. I saw ryand popping out some temporary front line attacks which were really effective, maybe something along that line could break the demux domination of the leaderboard. A fully dynamic algo somehting like aelgoo should be theoretically be better that a predictable defense line.