Sperical Case [Shrapnell] was a cast iron shell filled with iron balls with a small bursting charge that was supposed to be ignited by a fuse. Alas there were a number of premature detonations due to the friction of the balls and faulty fuses. This was only solved by Boxer in the 1850s. Spherical Case was excellent against targets at known distance and especially defensively. Spherical case only exploded when at the distance required if there was no problem with the fuse.
"So why would a cannister round have burst?"
This has been the discussion. It has been concluded that cannister of thick tin plate [French / Gribeauval Design] can did not burst in the gun tube but split in flight or less likely upon hitting the ground. The balls being distributed by the spinning of the cannister [tumbling as it is not ballistically stable]. The bottom of the tin being heated by the charge and so the solder fails and then along the seams. The balls had enough kinetic energy to richochet to some distance again depending upon the ground.
This explanation I think holds the key to the objection of cannister being rather hit and miss. It was not always the horrendous killer as given in wargames rules. Now the construction of other cannister was different and there may be the action of base plate acting as a piston to expell the contents of the tin hence a shorter range for this form.
Cannister made up of balls in a cloth bag had a very small range as the container was consumed before leaving the barrel as explained by De Scheel (1800; p78). This could be called light graprshot as it used 36x 1.5in balls for the 12-pdr. It should not be confused with Naval Grapeshot of 9 balls e.g. 9x 3-pdr shot for a 24-pdr.
It is a very complex subject but I think there is a much better understanding now of the process. I hope that has answered the question.