You raise an excellent point, and I don't disagree that aspects of the line were quite overbuilt. Nevertheless, there is still the fact that it was never completed to what the line was originally envisioned to be. For that, it's hard to fully know whether or not the line should have justified 6 car platforms given the time and the knowledge then. The line was built on the promise that it would be built as a 15 km line (instead of a 5 km stub) with development ensuing all around Sheppard. That was happening at the time, but the density was quite low regardless.
We as transportation engineers can't really predict what ridership patterns will be in the distant future, there are just far too many variables, many of which can't actually be quantified or predicted. One thing to consider with Sheppard is that if it was originally built as envisioned, the section between the STC and SY may see 150K PPD (so 75K PPD). Given that the line really only serves peak passengers and that there are no transfer points (ie no DRL, and no RER), those 75K passengers (round down to 50K passengers per direction since reverse and off peak commutes do exist) would only be traveling in the one direction over the 3 hr peak period, or an average of 17K passengers per hour in the peak direction. This already exceeds the capacity of light rail and quickly approaches the capacity of light metro (skytrain). We also have to remember that travel patterns aren't averages. The number of people traveling down a line during a peak period is bell curved, so while 17K is a fair average, the true number of people traveling at a time may be 13K passengers during the first hour, 26k passengers during the second hour, and 11K passengers during the final hour. It has the potential to peak at quite a high value. Obviously this is a gross oversimplification, but the point still remains that it made sense for the to plan for the future given what they knew at the time. Knowing what we know now, obviously it was grossly overstated, but the planners of the turn of the century had no idea the line would never be completed, that the developments would take 20 years to come, that zoning laws wouldn't change to reflect the existence of a subway, that additional bus routes wouldn't be incorporated into the line to serve the surrounding communities, or that commuters up in north York wouldn't be willing to give up their cars as easily.
The Yonge line wasn't built to fill the large 6 car trains we have today, in fact, it was built with smaller platforms which had to be expanded as time went on. It's harder to do that in a deep bore subway than it is on a trenched line. The line also did not nearly carry peak loads of 30K (or probably even 15K) passengers per hour per direction, it's taken over 60 years to get that line to 30+K PPD, so capacity was considered for the long term only. The bloor line is a bit different but for other reasons. The point still stands that the planners of the day could not have anticipated the changes that were happening to society and transit planning. With the knowledge we have to day, sure, we can criticize the scope of the project, but we should also consider the reasons for their size.
An interesting comparison to the Sheppard line will be the Eglinton crosstown line. Platforms there are being built at 90 meters, while Sheppards platforms have the potential to reach 140 meters. The catch is that the spacing of Sheppard's stations are a bit greater than those on the crosstown, so additional stations are built there in place of larger platforms. Whether one or the other is good or not (given Bessarion's performance, crosstown tech or skytrain with Eglinton stop spacing may have been the better choice overall) can be debated. I believe the crosstown is overall more expensive than a typical subway line per km because of the larger tunnel diameters and greater number of stations.
