Please subscribe to the official Codeforces channel in Telegram via the link: https://t.me/codeforces_official. ×

tzuyu_chou's blog

By tzuyu_chou, history, 2 months ago, In English,

https://www.youtube.com/watch?v=9vJRopau0g0

tl;dw

50k ppl were asked to complete a coding challenge. When there were points lost for wrong submissions, only 52% passed. When there were no penalties, 68% passed. Thus, codeforces contests are bad for training because of the penalties. Don't do contests to get gud.

 
 
 
 
  • Vote: I like it  
  • +58
  • Vote: I do not like it  

»
2 months ago, # |
  Vote: I like it +8 Vote: I do not like it

ya,it feels so motivated after watching the video,there should be no penalty and learning is the most important but the truth is,everyone wants to push others and become first :D

»
2 months ago, # |
  Vote: I like it +13 Vote: I do not like it

This is why USACO is so good

»
2 months ago, # |
  Vote: I like it +56 Vote: I do not like it

The Guessing Game Effect — I'll Just Keep Submitting Shit Till I Get It Right Because Why Not.

Not in all problems, but sometimes you want to penalise people for trying to squeeze through.

»
2 months ago, # |
  Vote: I like it +78 Vote: I do not like it

The main statement "shift the focus from falling into pits to saving the princess, to stick with a task and learn more" makes sense. The video has nothing much to do with the statement, and I find the video stupid. Also, I disagree with your conclusion that CF penalties are bad. I disagree so much that I'm even thinking whether you wrote that ironically.

Mark Rober (the guy in the video) created an experiment, where the goal for a participant was to do a coding challenge, not to learn something. If for some reason I want to the same or e.g. juggle a ball 10 times, and every time I fail somebody tells me "you're slightly worse than others, 5 points penalty for you", of course I'm less likely to finish the challenge. The same happened there.

The experiment could only examine one's motivation, not necessarily the speed/efficiency of learning, so it has nothing to do with people wanting to do competitive programming, or wanting to learn for an exam efficiently. Thus, the title of the video is wrong and/or stupid.

He shows a lot of projects/videos that he made and I agree they are cool. Other than advertising, his point was that one project was harder to be done, and I see no connection to the rest of the video or the main statement. I think that his projects are cool, though.

I did some extra research, what wasn't easy because Mark makes his videos unlisted. I managed to find the video where he announces the challenge (link), and also one announcing results and explaining the hidden experiment (link), and the game itself (link). The first important thing is that he promises an ipad for a random person. It's obvious some people thought they need a good score to win that (having a score suggests that). The second thing is that the game requires quite a few attempts on average (because of its nature). The average number of attempts is around 9 in no-penalty version. So with penalties you would get from score 100 to score 55 on average. Isn't that unmotivating? One of Mark's points is that those were virtual points that don't matter and nobody would see them later. Participants didn't know that. Mark also examined type of messages displayed (positive, neutral, negative) and the difference was significant too. A real conclusion should be: motivation matters.

And CF contests are for comparing yourself with/against others. You should upsolve for practice and learning, and there are no penalties there.