Post 17.04.2011   # 1
Subject DumbTicTacToe

I want to make a compaint about the task 400.

It was not nicely explained what is statistical chance. I thought this means to find the probability that the next player on move will win, keeping in mind that every state has different probability to come in and the probability is sum of probabilities of all states for which the next player wins.

I coded my solution pretty fast and then I was trying to find my "bug" for 1 hour and 10 minutes and I couldn't find it, and after I saw the code that passed the test cases I saw from the correct code that I see the statistical chances differently. So, the code that passes the test case has this definition of statistical chance 

number of wins of the next player/total number of different games

I think it was not well explained in the task what is statistical chance. Because every computer played random I assumed uniform probability in making move in each of the empty field in each state.

I totaly ruined this competition because of the wrong view of the definition of statistical chances. I think it should have been better explained. I didn't spend time on the fifth task which I knew how to solve, because I wanted to find my "bug" and to submit the 400 task first...

I hope in next competitions, the tasks would be better explained and not ambiguous.

igorkulev is offline Reply
Post 18.04.2011   # 2

I agree with you Igor, i too wasted ~40 minutes of my time on the 400 thinking i had a bug . But i also think the task was relatively easy and if the part of the statement where the probability is mentioned was better explained, the statement could have been called a pseudo code instead of a statement  ))). Really , the solution was in the text itself ))  . So, in my opinion you needed quick fingers,and some good luck in order to solve it , both of which are part of programming contests ))

vasja is offline Reply
Post 18.04.2011   # 3

In my opinion, problems should be written to be as understandable as possible. That is why on topcoder tasks where Euclidean distance is mentioned, you see in the bottom definition of Euclidean distance.

Here the term statistical chances was used and not defined. If you google on these two words you won't get a single result containing these two words in order. The mistake was that the "statistical chance" was not defined in the text. I don't agree that programming contests should contain tasks which are ambiguous so that the first one to find the correct meaning will succeed (the luckier one), because this is not a lottery, this is a contest where contestants show their skills.

Another thing, I interpreted the problem like I think it was right because there is no probability in the formula (count the positive outcomes/ total number of outcomes). These are not chance of winning. Especially when given that each computer plays at random indirectly assuming uniform probability for next move. In the definition of the statistical chance which is the "right" one, it is not important whether the computer plays at random or not, as long as the probability for each move is bigger than zero. Why is given the information that the computer plays at random? This is confusing also.

I won't continue with discussion and I really hope that on the finals there won't be ambiguous taks and everything will be explained in detail.

igorkulev is offline Reply
Post 18.04.2011   # 4

I have to agree with igor. Me and deluxd asked for an explanation on the arena chat but we were told "it'd take hours for us to fully explain the example" by the admins which is not correct in my opinion.

I solved 100,200,300 done at 13.00 so I had 1 hour and 10 minutes to figure 400 out which I couldn't.

boro is offline Reply

Please login to post reply.