So, here's the breakdown. In 2007, the Supreme Court disposed of 61,845 matters. 5,054 of these were regular hearing matters. Out of these the Supreme Court released 1,488 written decisions (based on JUDIS). That year there was about a 29.5% chance that a disposed regular hearing matter would result in a written decision (at least one that would in any way live on in SCR or Manupatra or Indian Kanoon, etc.). And keep in mind, only published decisions have a realistic chance of impacting the law going. In 2006, about 24% of disposed regular hearing matters resulted in a released written decision (1146 out of 4762). In 2005, 19.6% (861 out of 4382). (Overall based on these years there seems to be about a 1.8% to 2.8% chance that a case that gets submitted during admission phase will eventually result in a released written decision, although these numbers are approximations because cases enter the system during different years).
Now, these numbers I cited so far I am fairly confident in because the disposal numbers are from the Court and can also be checked against the Supreme Court's annual report (and they are close, even if not quite a perfect match) and the JUDIS numbers can be checked against Manupatra and Indian Kanoon (and the numbers are close, although not a perfect match if you do searches in these years). When you try to look historically it's difficult to be as confident in the numbers, but is still worthwhile. The annual report lists disposal numbers for past years which I will assume are accurate (although sometimes they vary somewhat widely between subsequent years). Furthermore, when you do searches on JUDIS (J), Manupatra (M), and Indian Kanoon (K) for previous years you start getting fairly divergent results. For a bit of flavor look at this preliminary chart (apologies for the scrunching - the first number is the year, followed by the number of disposed of regular hearing matters that year, followed by the number of written opinions that came out that year based on legal database searches, with the last number being the % this is of the disposed of regular hearing matters):
Year Reg. Hearing Disposal Written Decisions %
1951 227 54(J)/80 (K) 24
1955 200 81(J)/126(K) 40
1960 1,271 273(J)/328(K)/320(M) 21
1970 2,569 258(J)/408(K)/507(M) 10
1980 2,433 239(J)/448(K)/511(M) 10
1990 4,348 405(J)/715(K)/769(M) 9
2000 4,320 714(J)/1548(K)/1769(M) 17
2003 6,905 714(J)/1324(K)/1147(M) 10
2005 4,382 861(J)/918(K)/858(M) 20
2006 4,762 1146(J)/1193(K)/1140(M) 24
2007 5,054 1488(J)/1522(K)/1439(M) 30
2008 6,240 2676(J)/2443(K)/1877(M) 43
The %'s are based on Judis numbers, and some years they would be quite different if you based them on Kanoon or Manupatra. I'm not sure what accounts for this variation, but perhaps someone can comment on it who might know. It is difficult to draw too many conclusions until this is clarified. It does seem in the last few years at least there has been a spike in both the raw number of judgments released and their % of overall disposed of regular hearing matters. (It should also be noted these are just general searches of legal databases by year so they would also pick up interim orders, but this should be a relatively small part of the overall hit load).
Even once we get a better handle on these numbers historically, do we know what an ideal % level or raw number would be for the number of judgments the Court writes and releases each year? For some Supreme Courts (like the U.S.) it writes and releases close to 100% of judgments for regular cases it disposes of, but it's caseload is obviously much smaller and it has a different function in American society and within the American judiciary. Written opinions provide accountability and arguably legitimacy, but they may not be needed/desired in many cases the Indian Supreme Court disposes of. However, what are all these cases that don't result in a written, released opinion? And what % are what? Are they dismissals? Are they withdrawn from court in a settlement that doesn't result in an opinion? Are they one-line orders not worth reporting? Are they orders just spoken in court? Are they cases that don't involve a point of law, and if so, what are they being decided on?
Once we have a better sense of what these unreported cases are we would then have better information to judge how many of them should result in written, released judgments that are then documented in SCR, etc. Also, it would give us a better sense of how, if at all, we want to keep track of them - perhaps, the Court should provide for the public some other type of record that indicates what happens to this part of its caseload. Alternatively, it might be decided that this isn't needed and would just be a waste of time.
Because of the Indian Supreme Court's high caseload it's an apex court that lends itself to/requires analysis through statistics in a way other Apex Courts might not. It's impossible for any one person to follow all of the Court's activities without in part using using statistics to break down it's caseload. This is not to say in any way you can understand what the Court does solely through statistics - you obviously can't; or that you can't be misled by them - you obviously can. The challenge, and the academic debate, in this situation I would argue is to determine which statistics are desirable to track, figure out what they mean, and then use this information to inform your decisions and arguments about where the Court should head.