If yes then kindly share any example. I think in statistics a discrete random variable is that which cannot take on all values within the limits of the variable.
But Johnny technically is not right. Negative sign here only indicates (a concept) of the loss....(gain +, Loss -). The probability mass function, even in this case, is strictly positive.
I believe he is right. He could have simply given another example with dice, where if you have two fair dices whose outcomes is represented with X and Y random variables and you compute X-Y, what you get is a random variable, say Z which has a pmf
(of course the pmf would add up to 1 and the probabilities of outcomes of the random variable Z would be -5 to 5 and would all obey the laws of probability but perhaps I am misunderstanding your comments.)
Yes, because it depends upon the nature of trial or an experiment
for example in a game of wining and loss , for wining value of random variable is defined Rs: 2 and for loss it is defined Rs:-2 , so in this trial random variable can take values 2 or -2. Moreover negativity condition is only define for probability of a random variable not particularly that random variable.