For the 2018 Topcoder Open Marathon Match Round 3, the task required a good knowledge of university Mathematics, or statistics to be more precise. It turned out that many Topcoder members do have that knowledge and more than 60 people managed to achieve very close results. Marek Cygan did some analysis of probabilities a person taking a certain place on the final standings. These statistics show how close the solutions are and how hard the competition for getting to the finals in Texas is.
This time I decided not to participate, or at least not to submit – I had no compiled source code for this task to submit after all. To be honest I count this my behavior as if I went off to the track in a running competition, which I personally consider as not so good sportsmanship. Do not try to repeat unless you have to. As an excuse, there was no good idea in my mind during the contest, and that idea was that we have to trade-off money spent into analysis of the experts’ parameters – standard deviation and accuracy, and investing into the best deals being on the other side of the balance. So in the beginning we are clearly blind guessing the experts and it might be more profitable to take only positive-balance proposals, or proposals above some certain profit percentage. In this scenario we are losing the time however, as not all the experts are analyzed. As the time, or amount of rounds, is limited, this is probably not the best strategy.
What is the best strategy then?
The top scorers of the provisional leaderboard have posted their solution ideas on the forums. Check it out to get the knowledge from the best.
Congratulations to Errichto for winning the match and securing a spot in the finals. He showed that a transition from SRM participant to a Marathon Match participant is possible and can yield high results.