As this year's IOI is around the corner, I have been considering where I stand and what to aim for this year. I am writing this blog to share some of my thoughts related to this, as well as collect some opinions and experiences from other IOI participants.
Two metrics to quantify skill
First of all, I would like to talk about how I think of skill when it comes to OI, as I will be mentioning it later. Each competitor's skill is defined by two metrics: untimed skill and timed skill. I think of these similar to how CF problems are rated: someone's untimed skill level is the difficulty level at which they'd be able to solve 50% of problems given a very long time (say 5 to 10 hours, essentially unlimited). Timed skill is defined similarly, but under the time constraint of 100 minutes per problem (the average time that can be allocated per problem at IOI). I believe that practice targetting each of these aspects is important to improve IOI performance.
The dynamics of gold medalists
Inspired by E869120's blog on winning a gold medal at IOI, I have put much thought into what it takes to achieve each kind of medal. When it comes to gold medalists, it seems like there are two main classes: those who were able to be fairly confident (>70% chance) about winning a gold medal going into IOI, and those who were confident about winning a silver, and had some chance (15~70%) at a gold medal. Just for a very rough estimate, I would say that those who rank 16th to 60th in expected performance fall under the second category, which I believe includes myself for this year.
For this group of competitors, playing the safe strategy of spending a roughly equal amount of time on each problem and grabbing as many subtasks as possible would be able to secure a silver medal. However, when it comes to the race for a gold, it is much more dependent on the problemset for that year, which naturally introduces quite a large element of chance. With this line of thought, for these last two months of practice, I aim to maximise my chances (even if it ends up just being 20 or 30%) of winning a gold medal at IOI 2026.
The dilemma of how to practice
Setting this goal has put me into a dilemma: part of me thinks that focusing on my untimed skill is more important to be able to widen the scope of subtasks I could potentially solve. On the other hand, I am able to pass the gold cutoff on most IOI sets given unlimited time, which makes me wonder if I should instead improve my speed.
If you are a former IOI gold medalist, I would like to know more about your experience practising. Specifically, I would like to know your answers to the following questions:
How many hours a day do you practise on average? If it's inconsistent or irregular, feel free to elaborate more! (e.g. 3 hours a day during school days, 8 hours a day during intensive IOI camp)
Do you participate in any structured training? (e.g. at school, national training camps etc.)
What problems do you usually do as practice, and at what point do you give up to read the editorial (if applicable)? (refer to my answer in comments)
Thanks for reading!
