one is proving the inductive rule that: p(n) implies p(n+1)
the other is proving that there exists some number i for which p(i) is true. This, combined with the proof of the inductive rule, allows you to prove that p(n) is true for all n >= i.
In simpler terms the two steps for this problem are:
1. There exists some number of horses where all horses are of the same color. This is obviously true when you have 1 horse.
2. The second step is that you have to prove that if you have some number of horses n, and they are all of the same color, then if you add another horse all the horses are of the same color.
He's saying there is some set of n+1 horses, n of which are of the same color. Then he makes the (false) claim that removing one horse at random leaves you with n horses, which therefore must all be of the same color. This is incorrect because the assumption is that there is a set of n horses, not "if you have n number of horses they are all the same color.
The base case of n=1 is just fine. It's adding a randomly colored horse to a set of horses that doesn't always result in an n+1 set of like-colored horses. He doesn't even explain the illogical step correctly, so the whole thing is a mess.
> He's saying there is some set of n+1 horses, n of which are of the same color.
> This is incorrect because the assumption is that there is a set of n horses, not "if you have n number of horses they are all the same color.
You have this exactly backwards. The assumption you say is not being made is exactly the one that is being made (and is a step that occurs similarly in every inductive proof.)
one is proving the inductive rule that: p(n) implies p(n+1)
the other is proving that there exists some number i for which p(i) is true. This, combined with the proof of the inductive rule, allows you to prove that p(n) is true for all n >= i.