2 ms·

That difference only matters for discrete probability, and has the problem that the empty interval X<0 is impossible to represent.

• nequo 4y ago
  Yes, this is relevant for discrete distributions. Why is X<0 impossible to represent? If the support is {0, 1, …, k}, then Pr(X < 0) = 1 – Pr(X >= 0) = 1 – Pr(–X <= 0) = 0.
• wnoise 4y ago
  There are distributions that are mostly continuous but also have point-masses.