Fall Challenge 2020 - Bugs & Questions

Of course I did scroll, but the statement looks exactly like the old one since I can’t find the instructions about the new spell to acquire new ingredients.

In addition to that, if I run the code I get this error,

provided invalid input. Expected 'BREW <id> | CAST <id> | REST | WAIT'

Got ‘OPPONENT_CAST 85’

but if I print input, I get this one

Expected 'BREW <id> | CAST <id> | REST | WAIT'

Got ‘[‘75’, ‘BREW’, ‘-1’, ‘-3’, ‘-1’, ‘-1’, ‘16’, ‘0’, ‘0’, ‘0’, ‘0’]’.

Where does that OPPONENT_CAST 85 come from?

Hello,
I often see a bug and the final score 1st place/2nd place does match real result

It’s very frustrating when first in score, but we get only 2nd place for no reason…

At the end of the game the number of ingredients in both inventories that are of tiers 1-3 are added to both scores.

4 Likes

those error messages are because you are printing something that isn’t the output instruction.
The command to print debug messages should be included in a comment in the starter code.

2 Likes

Thank you for the tip, I am going to look into it and see if I can figure it out.

Hello.
Does anyone know why my result is not taken into account?
Step 59

The tooltip on the timeline says “timeout”, your bot probably just crashed.

1 Like

Sometimes, I lose games in the arena because of timeouts. When I try to reproduce the game in IDE, there is not timeout. Even if I try it several times.
After putting some error logs, it seems the timeout comes from my search loop, in which I have put the necessary break conditions to avoid this.

Can somebody shares the strategy for making it to Gold and Legend league? Anybody used MCTS technique?

https://www.codingame.com/forum/t/fall-challenge-2020-feedbacks-strategies/187846

2 Likes

I also had a lot of timeout issues. It made me nervous !

After three days of tearing my hair off with java and its damned GC, I have decided to recode everything in C…

After an afternoon of typing C and debugging, the timeouts were still there… It was a bad surprise !

On Ocean of code I had set the timer to 49ms and it was fine. But here, it was definitely too much…

In the arena, everything went upside-down, the timeouts were stricking randomly but each time I sent the game parameters in the IDE, my bot was cruching the opponent who crushed me in the arena… Even lowering the timer to indecent values was not enough…

Maybe I wasn’t the only one to see that, but from one submit to another with the same code, the difference was huge. With my last code, my rank varied from 10 to 200.

The silver boss is very strong (maybe too strong), improving the code to beat him was fun, but took me a lot of time. To much time to find a strategy that could bring me to the gold league… It is the first time I could not beat a silver boss ! I am so angry I will stop coding and start selling beer !

This morning, I was crushing him in the IDE, but in the arena, I had so many timeouts I barely reached the 10th place. I was insuling him in front of my screen each time I saw my bot ranking several dozen places from him !

I was angry at him because my rank sucks ! But nevermind, each challenge makes me better, and despite my deception, I’ll be there in may !

One last thing : in case of crash, the error messages are not displayed in the new IDE console. Some timeouts were hiding segmentation faults and it was very hard to debug. Have you changed something in the console ?

I don’t blame the coding game team, I know you did the best that could be done so far. Your servers answered to several thousand players and you have certainly spent a week under pressure !

This challenge was bigger the the others, and most of all it was great ! Thanks the CG team for this great moments.

PS : Before I forget to talk about it, for a long time I wonder how some hardcore Codingamers have managed to rank java bots in Legend League… I fear it will remain a mystery to me…

(Sorry if this message is a little bit off-topic…)

3 Likes

Will it be available for practice? How do I get my previously submitted bot’s code?

Yes, it will be available in the normal bot programming section. I don’t know how long it will take, but if I remember correctly it was already available at the last contest a few hours after the end.
If you want to have a look at your code now you can look at the past contests section under “view report” of the fall challenge: https://www.codingame.com/contests/finished

Edit: It’s available now: https://www.codingame.com/multiplayer/bot-programming/fall-challenge-2020

1 Like

FYI, we’ve removed some bots from the leaderboards. Like for the Spring Challenge:

  • either alt accounts.
  • or bots who shared the same code (or most of it). We kept the best ranking ones (or the ones we were sure they were the origin).
2 Likes

Hi .

There is an “issue” with the leaderboard :
I submitted a new code 30 minutes before the end of the competition.
I finished 1st in bronze league, and unable to access silver league because the fights were still ongoing, and the elligibility for new league happened only minutes after the end of the contest.

With the challenge now available as multiplayer, I re-submitted my code and I just won 600 places in silver league.

No need to correct this now, but it would be nice if next challenges did not have this issue.
(but I would gladly accept a correction in this challenge :wink: /me cries in bronze league in the challenge. )

looking at this game of your last battles you were 0.02 below the boss. Promotions always complete, even after the contest deadline.

1 Like

I double checked and I’m pretty sure you were never eligible to Silver. Even if you ranks first amongst players in leaderboard, you need to be above the boss to get promoted.

You score was 38.4876401331549 , the boss is 38.5107248639532. That’s why.

2 Likes

This is a common source of confusion on the chats. Could you consider providing the full(er) value as a hover tooltip?

The scores are displayed rounded to the nearest 0.01. There should be no confusion in this case.

We’ve had a case where the scores, rounded to 0.01, were the same.