Hi Sparky
We are still working out the exact particulars for the map sizes, balancing and all that. But realistically we won’t expect you guys to cater for ridiculous sizes like 100x100 map sizes.
We’ll most likely not go larger than 40x20, (20x20 per side) so for now you can consider that a max value until we finalize the exact numbers
20*20… I did feel a bit cramped… Hmm…
Catering for a 20*20 map will be a bit more complex. Theres a few things that I need to change in my bot now… nothing major. Even thinking a 10 * 10 map (Which I believe will be the end map) will be a bit more difficult than the current version. 20 * 20 will be amazing though.
Will you maybe please consider adding a size configuration option to the config.json file or similar, so that I can see how much smoke comes out of my bot at the bigger sizes?
+1 for making it configurable!
Will the size used in a tournament vary between matches, or will the size be static and be known beforehand 
Could we at least get a range of what these values can be.
2<=x<=20 1<=y<=20 or something like that?
For fairness, I assume that a whole tournament will have to be played at the same size. Varying between matches won’t give a fair comparison between multiple bots at all.
They could always do something like the best of three between each pair of bots, with each of the three matches being played at a different size - I think that would be quite interesting, and keep bot writers on their toes. They wouldn’t even have to release the set of sizes until submissions were in.
I do think that knowing a range would be helpful, though, and would also very much appreciate being able to change the board size when running the game engine.
There is also this quote:
I think there is a (very) over powered approach that could be used for this game, but which falls down if the size(s) for the tournament aren’t known in advance. (Hint: pretty sure, though not certain, that a NN can’t be trained for an arbitrary size board!) So that may be one of the things they are trying to prevent by not just having a fixed size or set of fixed sizes like last year.
Hi all,
We are currently working on making the map size configurable via the config.json file for local matches. As for the size that the tournaments will be run in, the map size will be predetermined and will stay fixed throughout the entire tournament. We will finalize a tournament map size in the next two weeks or so, but we are still deciding on a map size that will balance the various possible strategies, and any input from you guys will be appreciated.
Maybe keep an eye on the following topic in regards to any game configuration changes:
If Map sizes get larger will it bed possible to build more than one building per turn?
Building more than one building per turn would definitely allow for some crazy situations. Not sure if that’s a good or bad thing.
Branching factor would be squared. My poor bot. 
That could make the game very interesting, if you allowed to spend all the energy you have on your turn building multiple buildings
No, one would not be able to build multiple buildings in a single round. The reason for this stems down to the way that commands are processed and used.
Remember that on a bigger map, it would take missiles longer time to reach the other side (we might need to balance this in the coming week as well)
Come on! what if we change the command to have the the form
x,y,b;x,y,b;x,y,b;…
maybe in a later round? 
So any eta on when we will have the final map sizes?
https://forum.entelect.co.za/t/tower-defence-game-configuration/311/10
As mentioned in that thread, we will have final numbers for the map sizes on the 27th of May. If there are any balancing feedback, it will be really useful 
Hi, any update about the tournament map size?
Also very curious as to whether a final number has been decided upon?
Not yet, but it will happen before this weekend. So if you guys have some good ideas, there’s still time to weigh in
I suggest a 16 x 8 Map size or maybe a bit bigger. This will make it harder for game trees just of the sheer amount of possible moves.
I assume you not using a game tree then? which makes your suggestion a biased one in an attempt to weaken other players?