Well … as we discussed before the contest, a 4h sprint will be pretty anoying if the submit are still this long at the start of the contest. Right now everyone is just looking at his idle last battle panel. I wonder if my submit will finish before the end of the sprint The winner will be the first to submit (when it was possible to reach bronze fast).
By the way there’s a huge bug with the mana. If you print PASS; SUMMON ID you can summon anything you want. EricSMSO is using this bug to destroying everyone. I really hope codingame will check/fix that kind of behavior BEFORE giving any prize.
I think the 4 hour sprint is a great idea. Just, next time, start everyone in bronze and without a possiblitly to advance to silver. So everyone has all the rules, and after 4 hours everyone battles everyone.
Big thanks to the developrs of LoCM, its a great contest.
The 3 wood league were a perfect step by step introduction to the rules, but I feel that in a short period of time like a “sprint” format, spending too much time trying to be promoted is a waste of time. Maybe make it a 1 wood league only to filter out the “I submited the default code” bots and god striaght to bronze with all the rules. I feel the sprint is more of a “hardcore mode” and doesnt need as much hand holding as normal contests. Those wood leagues are perfect for the marathon tho.
Also, being a wednesday, it means its pretty much a European contest. I get that most people in here are Europeans, but me being in the Eastern American time zone, it means a contest on a wednesday afternoon. Maybe do it on a saturday or sunday next time ? I had my 1h lunch break to do something and thats it, which makes the multiples wood leagues even worse.
On a positive note: the game is fun and I’m hyped for the marathon !
I really like the idea of sprints, short deadlines kick me into gear and stop me from being too lazy to implement a thing I thought of!
Not sure if 4 hours is too short - longer would have been nicer, but on the other hand, in general, I’m more likely to be able to clear an arbitrary 4-hour chunk of time in my schedule than, say, an 8-hour one.
That said, in practice, I think the servers have to be a lot more efficient/responsive before a sprint will be fun for me rather than frustrating. In particular:
While I actually rather like the idea of leagues, and incrementally adding rules, even in sprint contest, maybe there can be some other algorithm for advancing past a boss in a league? Especially in wood leagues, the ranking seems superfluous and wasteful - don’t make people compete with each other, just make them fight several battles each against the boss to determine whether they’re past this particular tutorial level. It was particularly sad watching hundreds of bots, all clearly better than the boss (boss was rank ~130), running tons of battles against each other before they were allowed to advance to the next league, at which point it all gets resent and they once again run tons of battles against each other.
I think that requests which have a human waiting on the other end (e.g. running my code in the IDE to test whether I’ve made a syntax error) should have WAY higher priority than automated stuff (e.g. running the battles to determine ranking), and ideally should be sandboxed/isolated from the automated stuff somehow. I submitted my last bot a few minutes before the end of the contest, after testing it just once, and spent the rest of the time trying to re-run code in the IDE and failing (it got to the end of the pseudo-loading bar, then froze, then gave me some server error in a red box).
It turned out that my last submission has some bug on it that makes it time out most of the time, which is super frustrating, because each time I ran a test in the IDE (which never hit that bug) I was beating bots ranked way higher than me. So I probably have some really stupid error in there that I could have fixed, if I could only test my code, and would have done much better.
I ended up reloading the ranking and “last battles” quite a bit, which was really slow (and probably also really hard on the server), when I really only needed a couple of pieces of information - Where the boss was ranked (and whether I was already ranked above it), and what percentage of battles have completed for me. Ideally, there would be much cheaper calls/ an easier-to-reload interface for that sort of information. At least one time, I gave up waiting and resubmitted when I shouldn’t have (I was already ranked way above the boss even on my previous submit, which was about halfway through, so all I did was slow down when I’d get promoted).
similarly, just loading the interfaces for things like IDE enemy selection slows down soooo much when there’s too much ranking activity.
Anyway - The LoCaM game rules also seem really fun and interesting from the 4 hours I’ve spent with them; though I’m not sure if this type of game (lots of fiddly numbers and two stages to take care of right off the bat) is the ideal candidate for the first ever sprint - it took me just about an hour to make my first submit, though granted, it turned out to be a good one (some robust greedy selection stuff that served me very well up until that stupid syntax error, whatever it was, I hope I never find out ).
I also wanted to say that the timing for a 4 hours sprint is pretty bad, anybody on the Asian side it’s too late as they have work/school the next day morning and in America they are at work/school. So yea, that was a European contest.
But as I see the comments here, maybe it was a good Idea to do that first sprint when no one is available, probably that codingGame would have been ddos’ed
In any case, I suggest for future sprint like this to start the contest on a Saturday noon UTC. In this way, everybody between middle America to western Australia could participate in a “normal” day without having to skip school, work or sleep.
After that, I’d say it would be way better to leave the game available when the sprint is finished. Just block the submit until Sunday noon UTC for the ranking to complete and then let the marathon start officially. Anyway, anybody that did the sprint could improve its code in a local arena while waiting for the marathon to really start.
I look forward for the marathon and for the next sprint, hoping I could make it.
nice CC, the Sprint is fun, thx aCat and radekmie.
But the submit time… I am really frustrated that my Bot what not prepare for the bronze league the submit time kill the wood league rules
4hour after the end of the Sprint my bot do wood2=>wood1=>bronze and stuck around 350place cause of new rules…
like others I hope somthing change for the next sprint (one rule for the whole sprint seem the shortest way)
Hi, here are my feedbacks
First of all : BeberLeNewbie rules the contest : first french guy
when I asked a shorter contest, I didn’t think of a 4h contest ; better a 48h or 72h contest.
Anyway ; this sprint was a huge success : 800 players in 4h ! With the performance issues we had, we can expect more than that.
So keep the idea to have shorter contest (shorter than 10 days) ; long contest doesnt mean more people.
4h contest is great, but this only depends on the infrasctrure capabilities you can have. If you have problems, do 24h ou 48h.
For the contest itself : nice one, I Like the draft mode and we didn’t have yet a card contest, so it is a good contest.
+1 same for me. I got 3 or 4 submits in and got promoted to bronze at 1:30AM…
This format will only work with either faster submits or only battling the boss to reach higher wood ranks.
(Maybe as suggested before only have one wood level to filter out the default-code-submits)
The true reason I used SUMMON on everything is that I didn’t understand the mana rule at first reading
I kept this behaviour because I had only 1 hour 1/2 available for this contest (2 submissions). By the way I don’t think it helped me in any way. All I get is a bench of “invalid moves”.