JEE Main · Method and backtest

How far to trust it

Tested on 2022–2026, the 30 chapters Lumi ranks highest held 53% of a shift's questions, against 37% for 30 picked at random. A plain all-time count of past papers did just as well (53%). Past papers tell you where marks have been; recency weighting adds little on top, and nothing here predicts the questions.

Top 30 chapters, hit rate: 92.7%; random 75.5%
Top 20 topics, hit rate: 66%; random 34%
Top 100 ideas, hit rate: 8.0%; random 3.3%
Chapter rank correlation: 0.68; all-time count 0.67

Sources

What went in

Each question comes from an official JEE Main Paper 1 question paper published by NTA (CBSE for 2017), and each answer from the final answer key. Lumi analyses a shift only when both exist. Of the 174 shifts held from 2017 to 2026, 42 are in: 3,465 questions, of which NTA dropped 14. No 2018 or 2019 shift is in this analysis. The coverage table lists every shift held and why it is or is not here.

Every question was tagged with its chapter, topic and the exact idea it tests, plus format, skill and difficulty. Of the 3,465 questions, 1,696 fit an idea in Lumi's JEE taxonomy closely and 1,180 fit it reasonably. 589 had no close idea; they count for their chapter but carry no topic or idea. No answer key needed an erratum. Difficulty is Lumi's rating, not NTA's.

Chapters follow NTA's 2026 syllabus. Chapters it no longer lists are kept, marked as out of the syllabus, because their older questions are still useful practice. Syllabus checked against Syllabus for JEE (Main) - 2026, Paper 1 (B.E./B.Tech.), NTA, linked from jeemain.nta.nic.in on 2026-10-04.

Backtest

Tested on years it had not seen

For each year from 2022 to 2026, Lumi ranked chapters, topics and ideas using only earlier years, then scored the ranking on every analysed shift of that year, next to three simpler ways of picking. The score is the share of a shift's 75 questions the top picks held.

History against guessing

22% vs 11%

The 20 topics asked most in earlier years, counted with no weighting at all, held 22% of the next shift. Twenty topics picked at random held 11%. Most of what past papers can tell you comes from simply counting them.

Weighting against counting

23% vs 22%

Letting recent years count more added 1.1 points for the top 20 topics, about 0.8 of a shift's 75 questions. For the top 30 chapters it was 53% against 53%: no difference. It did better on exact ideas (12% against 9% for the top 100), but 100 ideas still held only about 8.9 of 75 questions.

Four ways to pick, side by side

Share of a shift's questions held by each method's top picks, averaged over the analysed shifts of 2022–2026. A random pick is the expected share for that many chosen blind. A gap under one question of a 75-question shift counts as a tie. Against the all-time count, the weighting was ahead for 0 of 3 chapter lists, 2 of 3 topic lists and 3 of 3 idea lists.

Top 15 chaptersTie · 0.0 Q
- Weighted estimate30%
- All-time count30%
- Last year only28%
- Random pick19%
Top 30 chaptersTie · 0.0 Q
- Weighted estimate53%
- All-time count53%
- Last year only50%
- Random pick37%
Top 45 chaptersTie · +0.2 Q
- Weighted estimate72%
- All-time count72%
- Last year only68%
- Random pick56%
Top 20 topicsTie · +0.8 Q
- Weighted estimate23%
- All-time count22%
- Last year only22%
- Random pick11%
Top 40 topicsAhead · +1.1 Q
- Weighted estimate40%
- All-time count38%
- Last year only38%
- Random pick22%
Top 60 topicsAhead · +1.7 Q
- Weighted estimate53%
- All-time count51%
- Last year only51%
- Random pick32%
Top 50 ideasAhead · +1.2 Q
- Weighted estimate8%
- All-time count6%
- Last year only6%
- Random pick2%
Top 100 ideasAhead · +2.5 Q
- Weighted estimate12%
- All-time count9%
- Last year only9%
- Random pick5%
Top 200 ideasAhead · +1.9 Q
- Weighted estimate19%
- All-time count17%
- Last year only15%
- Random pick9%

List	Weighted estimate	All-time count	Last year only	Random pick	Weighting vs count
Top 15 chapters	30%	30%	28%	19%	Tie · 0.0 Q
Top 30 chapters	53%	53%	50%	37%	Tie · 0.0 Q
Top 45 chapters	72%	72%	68%	56%	Tie · +0.2 Q
Top 20 topics	23%	22%	22%	11%	Tie · +0.8 Q
Top 40 topics	40%	38%	38%	22%	Ahead · +1.1 Q
Top 60 topics	53%	51%	51%	32%	Ahead · +1.7 Q
Top 50 ideas	8%	6%	6%	2%	Ahead · +1.2 Q
Top 100 ideas	12%	9%	9%	5%	Ahead · +2.5 Q
Top 200 ideas	19%	17%	15%	9%	Ahead · +1.9 Q

Year by year

Each line is one graded year, averaged over its analysed shifts (four a year before 2024, eight or nine after). Across these 15 comparisons the weighting was ahead of the all-time count 5 times, tied 9 and behind 1.

Top 30 chapters

Against the all-time count: ahead in 0, tied in 4, behind in 1 of 5 years.

2022Tie · 0.0 Q
2023Tie · −0.6 Q
2024Tie · +0.8 Q
2025Tie · +0.6 Q
2026Behind · −1.1 Q

35%60% of the paper

Top 100 ideas

Against the all-time count: ahead in 4, tied in 1, behind in 0 of 5 years.

2022Tie · 0.0 Q
2023Ahead · +1.8 Q
2024Ahead · +2.2 Q
2025Ahead · +4.5 Q
2026Ahead · +4.1 Q

0%20% of the paper

Weighted estimate
All-time count
Last year only
Random pick

Chapter weights: how close

	Rank correlation	Average miss
Weighted estimate	0.68	0.29
All-time count	0.67	0.31
Last year only	0.56	0.33
Same weight for every chapter	–	0.40

Rank correlation with the next year's chapter mix, and the average miss in questions per 25 a subject.

Top 30 chapters, hit rate by year

	2022	2023	2024	2025	2026
Weighted estimate	96%	92%	95%	90%	91%
All-time count	95%	93%	96%	89%	91%
Last year only	92%	87%	94%	87%	93%
Random pick	87%	84%	74%	66%	67%

Share of the 30 chapters that came up in a shift. Almost every big chapter comes up, so even a random 30 scores high.

Chapters: top 30, hit rate

	Hit rate	Covered
Weighted estimate	92.7%	52.7%
All-time count	92.6%	52.7%
Last year only	90.4%	50.4%
Random pick	75.5%	37.1%

Topics: top 20, hit rate

	Hit rate	Covered
Weighted estimate	66.0%	23.4%
All-time count	62.8%	22.3%
Last year only	62.6%	22.1%
Random pick	33.7%	10.8%

Exact ideas: top 100, hit rate

	Hit rate	Covered
Weighted estimate	8.0%	11.9%
All-time count	5.8%	8.5%
Last year only	6.3%	9.4%
Random pick	3.3%	4.7%

Sensitivity

Does the answer depend on the settings?

The chapter estimate, rebuilt four other ways and compared with the one Lumi uses.

Change	Rank correlation	Same top 30	Largest move	Moved most
Weight years by how many shifts they have	0.994	29 of 30	0.13 Q	Units and Measurements
Leave out 2017 (one offline paper)	0.998	30 of 30	0.07 Q	Haloalkanes and Haloarenes
Let old years fade half as fast (half-life 6 years)	0.996	29 of 30	0.11 Q	Application of Derivatives
Ignore the 2026 syllabus cuts	0.992	27 of 30	0.36 Q	The s-Block Elements

Every variant ranks the chapters almost the same way (rank correlation 0.992 or higher). The chapter order on these pages does not hang on one choice of settings. Ignoring the syllabus cuts moves the most (The s-Block Elements), because it puts dropped chapters back in. "Largest move" is in questions per 25 a subject.

Within a year

Do shifts in one session look alike?

Each 2024, 2025 and 2026 shift was held out in turn and predicted from the rest. Averages over the three years.

Topics: top 40

Built from	Hit rate	Covered
Earlier years only	57%	41%
Earlier years + the same year's other shifts	62%	45%
The same year's other shifts only	59%	43%

Random pick: 31% hit rate.

Chapters and ideas

Built from	Chapter rank corr.	Top 100 ideas hit
Earlier years only	0.59	9.1%
Earlier years + the same year's other shifts	0.65	9.9%
The same year's other shifts only	0.65	7.3%

Random pick: 2.8% for ideas.

Shifts in the same session share their topics far more than their exact ideas. Adding the year's other shifts to the history lifts the topic hit rate from 57% to 62%, and barely moves ideas (9.1% to 9.9%). It does not reveal questions. Over every pair of analysed shifts, the chapter mix of two shifts in the same year correlates at 0.46 (116 pairs), against 0.33 for shifts in different years (745 pairs).

Calibration

When it says 60%, does it happen 60% of the time?

Every predicted chance, grouped, against how often the unit actually came up in the shift.

Chapters

Predicted	Came up	Cases
5–10%	42%	12
10–20%	6%	34
20–30%	19%	81
30–50%	48%	659
50–70%	81%	1,296
70–90%	94%	574
90–100%	100%	13

Topics

Predicted	Came up	Cases
0–5%	9%	285
5–10%	17%	681
10–20%	20%	1,444
20–30%	31%	1,369
30–50%	43%	1,800
50–70%	67%	509
70–90%	80%	25

Exact ideas

Predicted	Came up	Cases
0–5%	2%	19,001
5–10%	3%	11,819
10–20%	8%	3,276
20–30%	13%	341
30–50%	27%	33

For chapters and topics the stated chances mostly run low: likely chapters come up more often than predicted. For exact ideas they run high. Read the chances as a ranking, not as exact odds.

See the shifts by year, starting with JEE Main 2026.

Questions

About the method

Can past papers predict JEE Main questions?

No. They show where marks have been. In the backtest the top 100 exact ideas held about 8.9 of a shift's 75 questions, and in 2026 39% of questions tested an idea not asked in any earlier shift Lumi analysed.

What is the weighted estimate?

Each chapter or topic's share of past shifts, with recent years counting more (a three-year half-life) and the 2026 syllabus applied, scaled to 25 questions a subject. In the backtest its top 30 chapters held 53% of a shift, the same as a plain all-time count (53%).

Why only 42 of 174 shifts?

Lumi analyses a shift only when NTA's full question paper and final answer key are both available. It covers 42 of the 174 JEE Main Paper 1 shifts held from 2017 to 2026 (3,465 questions), not every shift.