India runs a well-administered mathematics olympiad pathway. I have been through part of it and I have no complaints about how it is run. My argument is about the shape of it, and about what that shape costs the country in talent it never meets.

The numbers for the 2024-25 cycle, from the HBCSE olympiad programme and the associated reporting:

About 1,07,936 students appeared for the IOQM. Roughly 6,700 went through to the RMO stage. 1,099 were called for the INMO. From the INMO, 30 candidates are selected. Around 35 students in each subject go to the training camp at HBCSE, and six eventually represent India at the IMO.

One lakh to six. Any selection process for an international team of six is going to be brutal at the bottom, and that is not a criticism. The interesting question is what happens above the funnel, not inside it, because that is where the losses are largest and the fixes are cheapest.

One lakh is the small number

India enrolled over 24.8 crore students across school education in 2023-24 according to UDISE+. The higher secondary stage alone runs into the crores.

Against that, one lakh students sitting the entry examination is a participation rate well under one percent of the relevant cohort. The competitive stage of the funnel, the part everyone discusses, is doing less filtering than the step nobody discusses, which is the step where a student finds out the examination exists.

That first filter is not selecting for mathematical ability. It is selecting for information. Whether your school registers. Whether a teacher mentions it. Whether your parents have heard of it. Whether there is a coaching institute nearby that markets it, which is now a substantial part of how students learn the pathway exists at all.

None of those correlate with talent. Several of them correlate strongly with income and with being in a metropolitan area.

Class 11 is late

The pathway realistically engages students in classes 9 to 12, and effectively concentrates in 11 and 12, which collides directly with board examinations and JEE or NEET preparation.

Ask when mathematical ability becomes visible and the answer, from anyone who has taught young students, is much earlier. A child who is unusually good at reasoning about structure shows it at ten or eleven. What happens to that child in most Indian schools is that they are good at the arithmetic in front of them, they get told they are good at maths, and then the definition of maths they are handed is a procedural one for six more years.

By class 11, three things have already happened. They have been sorted into a stream. They are inside a preparation track with its own timetable and its own definition of what mathematics is. And they have formed a self-image about whether they are a maths person, usually from evidence that had nothing to do with mathematics.

The countries that consistently do well at this select much earlier and build a permanent structure rather than an annual examination. There are circles, clubs and correspondence programmes that a nine-year-old can join, that carry no selection pressure, and that exist to let a child find out whether they like the subject. India has some of this. It is thin, it is concentrated in a few cities, and it is mostly private.

What the structure selects for

An olympiad is a timed examination with a fixed syllabus and problems that are known to have solutions. I have written before about how the skills this trains diverge from research, and the same divergence shows up in selection.

The pathway rewards speed under time pressure, accuracy under pressure, and coverage of a defined body of technique. Those are real abilities and the students who have them are genuinely strong. They are not the same as the ability to sit with an ill-posed question for eight months.

So the funnel is optimising for one specific profile, and the profile is not the one that produces research mathematicians. This is fine as long as the funnel is understood as a team selection mechanism, which is what it is. It becomes a problem when it is treated as national talent identification, because then everyone outside it is being told something about themselves that the process did not measure.

The part that bothers me most

Consider the student who does well at the RMO, does not make the INMO cut, and goes back to board preparation.

There is nothing for that student. No society, no continuing programme, no correspondence course, no reason to keep going. The structure is built around an examination, and once you are out of the examination, the structure has no further relationship with you.

That is a very large number of people. A thousand students reach the INMO. Six thousand reach the RMO. One lakh sit the IOQM. Every one of them demonstrated some interest in mathematics beyond what school asked of them, and the system’s response to all but thirty of them is silence. A hundred thousand people who self-identified as interested, and no one on the other end of the line.

The countries with deep mathematical cultures have permanent institutions rather than annual events. A society you belong to, a journal that publishes student work, seminars, problem columns, an ongoing thing. The examination is one activity inside a community. Here the examination is the entire structure and the community is whatever informal groups students manage to build themselves.

What would change it

The examination is not the thing to fix. It works.

Entry needs to move down to primary and middle school, and it needs to be non-competitive at that stage, because the point is exposure and not sorting.

The pathway needs to be discoverable without a coaching institute in the middle. A student in a district town should be able to find out what this is from a school noticeboard.

And there needs to be something for the students who do not make the final cut, so that the ninety-nine percent of participants who are not going to the IMO have a reason to continue. That is where most of the actual mathematics in a country happens, and it is currently unserved.

None of this is a resource problem. Correspondence programmes, problem columns and student societies are inexpensive, which means the constraint is not money but the absence of anyone who treats it as their job. It is a design problem, and the design currently optimises for the six.

Six is a reasonable target for a team. It is a poor target for a country.

Related: