A stack of previous year papers sits on almost every aspirant's desk at some point, usually solved once, checked against an answer key, and then set aside. This gets roughly a tenth of the real value out of what is genuinely the closest thing to primary source material a preparing candidate has access to, because a previous year paper is not just practice, it is direct, unfiltered evidence of what a specific exam's specific examiners actually consider important enough to ask.
A mock test simulates the exam. A previous year paper is the exam.
This distinction matters more than it sounds. A mock test, however well constructed, is someone's informed guess at what might be asked, built from a syllabus and general pattern awareness. A previous year paper is the actual, real record of what a specific state's judicial service exam has genuinely asked in the past, which means patterns you find in it, a subject that recurs across several years, a particular provision that keeps showing up in a specific form, are not guesses about tendency, they are documented fact about that exam's actual history.
Read across years before you solve, not after
The instinct is to grab the most recent paper and start solving. A more useful first move is to lay out three to five years of papers for your target exam side by side and simply read through the questions, not answer them yet, looking for what repeats. Which subjects appear in every year's paper versus which show up occasionally. Which specific topics within a subject keep recurring in slightly different phrasing. This reading pass, done properly, takes under an hour and tells you more about where to weight your remaining preparation time than a month of generic subject-wise study would.
Solve under real conditions, not comfortable ones
A previous year paper solved at your own pace, with your notes open, over two relaxed sittings, teaches you almost nothing about your actual exam-day performance. The value of these papers as a diagnostic tool depends on solving them the way you will actually sit the real exam, same time limit, same absence of reference material, ideally at roughly the time of day your actual exam will be held. This is uncomfortable and is exactly why most aspirants avoid doing it properly, and exactly why the ones who do it get a genuinely honest signal the rest are missing.
The gap between your answer and the model answer is the actual lesson
After solving, the useful work is not confirming you got the right general area, it is comparing your answer's structure, precision, and completeness against a strong model answer or a topper's copy if you can find one, and noting specifically where the gap is. Did you miss a provision the model answer included. Did you spend too many words on a point worth few marks. Did your structure bury the actual conclusion an examiner would be looking for. This comparison, done honestly across several previous year papers, builds a much clearer picture of your actual exam-readiness than a raw score ever will.