Popular opinion: CF ratings are unrelated to your IOI results. CF contests differ significantly from IOI contests.
❌ False!
Unpopular opinion: A contestant's Codeforces rating is more predictive of their final IOI result than their score on Day 1 is of their score on Day 2.
✅ True!
A higher CF rating, with a gap of at least 50, correlates with an 84% chance of achieving a better placement at IOI. The relationship between IOI placement and CF rating is strongly connected. The Pierson Correlation Coefficient for these two factors is -0.813. At the same time, the Pierson Correlation Coefficient between the scores of day 1 and day 2 of IOI is 0.804.
A side note for IOI participants: If you performed poorly on the first day, just think of it as a single bad Codeforces contest—you can always do better next time! You can likely compensate on the second day. Likewise, if you did a great job on Day 1, don't relax too much; the next day is a new challenge. The 0.804 coefficient shows a strong connection, but just as CF rating!
Correlation between CF rating and IOI placement
Here are the statistics on CF ratings for each medal:
| Medal | Minimum Rating | Median Rating |
|---|---|---|
| 🥇 | 1991 | 2450 |
| 🥈 | 1693 | 2303 |
| 🥉 | 1318 | 2106 |
Additionally, I've discovered an interesting correlation: participants' ratings one year before IOI also correlate with their final ranks; the Pierson Correlation Coefficient is 0.70.
Correlation between CF rating a year before IOI and IOI placement
Based on this insight, I’ve trained a model that can predict your placement at IOI. If you’re curious about your predicted placement for IOI 2026, leave a comment! Note that as time passes, the model can determine your place more accurately. Feel free to comment again in six months to see the updated prediction.
You can check out a similar blog post for IOI 2024 here.
I predicted results for some people in the above blog.
| Predicted Place | Actual Place |
|---|---|
| 49 | 67 |
| 104 | 71 |
| 148 | 156 |
| 80 | 126 |
| 174 | 121 |
| 107 | 274 |
| 254 | 290 |
| 183 | 209 |
| 180 | 194 |
| 59 | 88 |
Update: If you want to make the data more complete, please comment if you know someone from this gist.
Not sure what the source of your data is from. But I'm pretty sure that it is incomplete. Maybe you would like to merge it with the data from the IOI 2025 Teams blog (though it will take a lot of time). Still, thanks for making this blog! It's always fun to see the correlation between CF and IOI each year.
For sure, it's incomplete! I used the blog you mentioned for sure, but even with a lot of scripts and manual work, I was able to connect only 205 people with their CF profile.
What you said made me think of making a gist. I added names I don't know their CF profile, there. If you know someone, please comment.
By using CP HOF and some personal knowledge, I found 18 of the missing handles. Hope this helps! Thanks again for making this blog!
Thank you! That's a lot.
I think CPHOF has data for most contestants.
Can you predict my place?
160
hmmm not bad actually.
Without researching my identity and my place in the IOI, can you predict my place this year based on the last time I played Codeforces? I am curious how accurate this is.
64.
While trying to change the code to calculate this (as my code predicts your place in IOI 2026 by default, not IOI 2025), I figured out that because of a few 3000+ rated people in the model, my model predicts ranks worse than they should be. For example, people who should be predicted around 10 are predicted around 40. I'll try to fix this later.
Given that, your actual place should be something like 25.
I ama Kristjan Lepp 164. Can you predict for me
And 278 Rain Martin Ohlau is Rainmar
A Pearson correlation of $$$0.813$$$ means the linear model explains about $$$66\%$$$ of the variance.
Whats my predicted placement, i want to see if I over-performed on the virtual contest I did.
We need 2026 edition now
Sure, just wait 31017600 seconds.
https://ioi2026.uz/
Can you predict me, please?
273rd
Predict my place, please
publish the data!