I discovered Condingame.com a few days ago and, as an amateur/self-taught programming, I’m enjoying it a lot. So, many thanks to those who work behind the scenes to make it up and running. That said, I have a question: is it a possibility that some of the programmers I play with in the “Code clash” section are bots or automata of some kind? I ask, because at times they post their reply in just 15-30 secs, which is a rather impossible reaction time even for a fast-as-light programmer (can anyone even read the assignment in such a short time?).
Should what I’m saying be true, I think it should be avoided in some way, as it makes harm to the spirit of a honest “competition” - how can one enjoy clashing against allmighty machines?
No. It’s just people playing unfair.
Because the Clashes are so frequent, there are not enough puzzles for them to be unique so people just copy&paste the solutions they already know for the puzzle as quickly as they can.
There really isn’t a great solution to that, I guess.
I have to believe what you are writing, nevertheless I find it amazing that someone can identify a puzzle onscreen, look for the proper solution in his/her archive, then copy the solution in memory, go back to his/her browser, paste the solution from memory and push the “submit” button in as little as 15-30 seconds. Super fast indeed!
I’ve been told that there are indeed fully automated bots in Clash by now. I haven’t played it for a long time because even with manual copy&paste people playing I din’t enjoy it anymore.
Now that I know there are actual bots, it’s even less interesting to play (more interesting to write a bot ;oP)
What’s your typing speed? 15 seconds to invoke five functions is easy with little practice. Users who have played >1000 clashes probably have seen most of the ones in the pool and have already completed the clash before they’ve already seen it, just by looking at the example would already be enough to recall the logic.
Not to mention that most of the clashes are either trivial or ridiculously convoluted, underspecified - and also poorly tested for that there’s not much point trying to cover edge cases, so sometimes people would submit code that doesn’t even pass the sample tests and still get 100% if they know such a clash has that vulnerability. And there’s a lot of them in that category.
Honestly, if you want proper speedcoding experience, you should use a site tailored for that; the pool for clash of code problems is quite small and repeats itself regularly, and they’re also not tailored for correctness. Shortest is just code golf, except the site’s language choices are so small that there are only a few competitive options (ruby, perl, bash…), and reverse mode are black box problems where if you’re seeing it for the first time and having to figure it out, you’ve already lost to people who’ve already seen it at least five times before.
Cheating by copy-pasting for the 15-seconds range is preposterous - these are usually the people you’d see who are actually competent in speedcoding, and at the sub-20 seconds range it takes more time copy-pasting than actually solving a problem yourself. Cheaters using primitive techniques usually cash in at 20-25 seconds consistently, and it’s very easy to recognize them because their performance is consistent with almost no deviation and they don’t share their code. (This knowledge is based on my experience with dealing with cheat detection on an internet chess server)
As for what you can do about them, there’s not much really, I’ve reported many cheaters and codingame has done nothing about them. Remember that this site is a recruitment tool, the administrators don’t really care about your experience more so than they can keep their sponsors happy.
Oh, and about cheating bots - I’d love to see one at work, the day where a cheat engine goes around to clear everything in 2 seconds is the day I can sleep without having to cry myself to sleep. As to my knowledge however, the claim that anyone has had the time and effort to automate cheating for that gamemode is unbacked and ridiculous, because bots can hash a problem statement and poll from a tiny database, there’s no way it’d take it more than ten seconds.
Let me clarify things quickly so newcomers like @AldoBaldo (hello ) don’t get a false image from quite an unfair leap.
First, we do care about the experience of players here. I’m a player myself (not of Clash of Code though). Second, yes, we’re a business; we sell recruitment tests to companies. This has almost nothing to do with this platform which is free for developers. Furthermore, last year, we have implemented a captcha mode to prevent bots on Clash of Code.
Honestly, I don’t get why people play so much CoC (once they have played them all twice) and kind of ruin the experience of others. Our system is far from perfect; that’s for sure. If that was up to me, Clash of Code wouldn’t give any point for the leaderboard; perhaps that would help.
However, improvements on CoC are not in our top priorities right now. I hope you can understand that.
not sure about 15, but 30 seconds seems reasonably realistic to me - I just joined CodinGame this week and I’m pretty sure I did my 1st or 2nd CoC in under 30 seconds. I am not a bot. Though, granted, I’m a professional developer with an interest in CS education, so quite a bit of practice with simple concise puzzles.
It was a “reverse mode” puzzle (in fact, all CoCs I’ve tried were “reverse mode”), and I did it in Python 3. In languages like Python, it’s quite possible to bust out a solution in a short amount of time, as long as you remember the syntax for doing what you want.
) don’t get a false image from quite an unfair leap.