Then again you remove the choice to sacrifice a human and this simplifies the problem by a great deal
Thanks a lot for your explanation, it is really helpful. Do you the name of these kind of randomized method ? Is there a theory behind ?
It’s called “Monte Carlo”. You try random strategies in the giving time and you keep the best.
Hey, we would have loved to reveal the source code of top players as we know it really helps to improve, however, this would spoil the thinking for those who want to try the game, as we will place it into the game session!
Similar, but different
For anyone interested in knowing how the winner proceeded… Thanks eldidou for sharing!
I tried what MaximeC says
[quote=“MaximeC, post:27, topic:1089”]
An algorithm that was enough to get 100% was: go to the nearest human that I can reach before a zombie and stay on it.[/quote] And so I almost reached the upper third of the leaderboard. YEAH!
Well, the amount of points is crappy, but who cares?
This was almost exactly my strategy, I didn’t have the
repeat X times for my strategy however. I came in 81st place. Did you move towards the zombies position, move towards the zombie’s position next turn, or did you predict where the zombie will be when you get close to it? Also, for the “move to random position”, was the position a random point within the Ash’s circle of movement, or on the circumference of his circle of movement.
When CvS will be available as optimization puzzle?
You will receive a notification when the game will be available, i heard the team is quite busy on big feature right now
Amazing stuff, thanks for sharing!
As for the suggestions for future contests, I agree with what eldidou said – except for the last one. I think it is important to keep the completion criteria for the ranking. The best strategy handles 100% of cases in the first place. Here, the test cases were pertinent enough: I too was stucked at 95% for some time and overcoming this was actually fun.
Relaxing completion criteria could easily produce various unfair results. For example, my one submission got more points with 95% than what I could achieve with 100%. Taking it to extremes, you could win by solving just one case (e.g. a particular strategy for the CG case would do).
We received interesting suggestions regarding the validators but that requires some development first. As SaiksyApo said, we are currently working on really big features and that will come with it.
Do not expect to see the puzzle before next week.
I moved toward the zombie position next turn, which was slightly better than zombie current position (I didn’t tried other interception method).
The position of the random points was in an area of radius ~2500, so sometimes it was inside the circle, and sometimes it was on the circumference.
I agree with alex2cf on the 100% thing.
I live programming,i eat programming i dream programming i play programming…Programming is a big fun
This challenge is not in the game section. When will it be there?
ikr, want to get 1000 more points?
For more information about when will it be released, I’ll quote @MaximeC:
As SaiksyApo said, we are currently working on really big features and that will come with it.
When you’ll hear some news in the Log Book (click for a direct link to that topic), then you will find the challenge on the game page within few hours (depending on some priorities that we have when we release new features).
To complete what Yohann said, we have decided to postpone the release of the new features for several reasons and as a consequence, the availably of Code vs Zombies as a puzzle has been delayed a bit.