I’m in silver, i’m surprise I can only use trust, point distangle and angle as parameter.
I want to use n+1 checkpoint, opponant distance/angle.
Am I missing something, or we can use it in gold or more?
I am really close to legend league I need less than 1 scoring point to beat the boss can anyone see how to make my code better thank you. I can pm the code to anyone who can help
You can map checkpoints after your first pass to get better results for subsequent laps. I think as long as you are swinging a little wide and turning early and not missing checkpoints, the boss is not that bad. I did retry this because I wanted to make a GA with a little less complications than legend league has, and I was surprised to find this as well.
Proper Blocking is very important. Probably won’t take much to squeak out an extra point if you improve your blocker a little. Collisions work in fractions of a turn, so don’t just go by the initial values that are output. Break the collision down into fractions of a turn and you will find your blocker doing a much more amazing job.
Thank you I will try it
Hi guys,
i’m new to this game. bronze atm, they just added collision.
I really need some advise to get past boss 4.
No matter how hard i tried i seem to never reach the boss score.
I tried different things with no significant results.
Any advise on what to focus on or some hint?
ty
What is the best technique to track the opponent and block it in front? Thanks.
You’re just outputting the string ’ 100’ every time, not the thrust variable.
its been over a month since i’ve gone from wood 1 to bronze and I still can’t figure out what to do about the collisions.
Push other with one pod, and keep the first pod doing the race
Hey there. Handling collisions is something I haven’t had a lot of success with. I am in Gold league. I boost of the start line with pod1. I have the same code to reach targets, but I pod1 targets the checkpoints and I use pod2 to target the future location of the lead enemy pod.
Is there a way to check current thrust,the one in the game information only shows the initial thrust.
You can save the value of the last lap in a global variable
I agree.
One of my usual strategies is to create a Game object. I can track the current state of the game board in that object and store anything about past gameboards that I would like. I can do the same thing with any other objects I use to model the game. In your case I would have a pod object and one of the properties might be lastThrust.
can anyone help me get started, using python. I do not know how to get to the next target.
Hi all,
I’m possibly missing something, but there’s mention of using this to learn more about neural nets. However, the game state doesn’t persist between games - so how can we ‘train’ our net?
Apologies if this is a daft question!
Cheers!
You may refer to the discussion here: https://www.codingame.com/forum/t/neural-network-ressources/1667
I am in Wood 2 and cannot figure out how to implement the angles. (I am coding in C)
Is there any way to rank down to silver, or to play with the silver rules?
I jumped from bronze to gold as I was not even satisfied with my code…
Now I have to do multi agent thingies and genetic algorithms, which do not interest me atm, I would prefer optimizing my code for one agent.
No, there is not a way to go back to an earlier league after promotion.