Confused about Classful/Classless IP addressing question VLSM IP addressing.

Not sure if this is the right sub to go to, but I was just wondering if someone could simply explain this concept since I have a networking exam soon and I'm confused.

So the firstly is subnetting. The tutor gave an example:

"Network: 172.16.0.0 (65534 Hosts)

Subnet mask: 255.255.0.0

There could be a lot of wasted hosts here, so what if we took that B class network, and divided it into a bunch of C class networks e.g.

Subnet 1 - 172.16.1.0

Subnet 2- 172.16.2.0

Subnet 3- 172.16.3.0

Subnet mask: 255.255.255.0"

Right this part makes sense, however, the only bit I'm confused on is that it said divide it into a whole bunch of C class networks, but how is it a C class network when the first Octet is 172? (making it a B class network)

This is a separate question, but later on he goes on about VLSM and the formulas to work out how many hosts and subnets you get (I assume a subnet is a network?). What I was wondering is that he said 252 (11111100) was most commonly given out by ISPs because there's 2 IP address, but what about 254 (11111110)? How many IP address would there be for a 254 in the subnet mask?