I'm having a hard time understanding the solution to the following problem from Dan Stefanica's book "A Primer for the Mathematics of Financial Engineering":
Call options with strikes 100, 120, and 130 on the same underlying asset and with the same maturity are trading for 8, 5, and 3, respectively (there is no bid-ask spread). Is there an arbitrage opportunity present? If yes, how can you make a riskless profit?
To model the situation, we assume $x_1, x_2, x_3$ units of the three options are purchased (where the $x_i$ may be negative as well). Then, the cost of the portfolio is $8x_1 + 5x_2 + 3x_3$, which we want to be negative for an arbitrage opportunity to exist. So we get our first equation
$$8x_1 + 5x_2 + 3x_3 < 0$$
Then we model the payoff of the portfolio at maturity as $$C(T) = x_1 max(S(T) - K_1, 0) + x_2 max(S(T) - K_2, 0) + x_3 max(S(T) - K_3, 0)$$. This gives us a piecewise description of the payoff -
\begin{equation} C(T) = \left\{\begin{array}{lr} 0, & S(T) \le K_1 \\ x_1(S(T) - K_1), & K_1 \le S(T) \le K_2\\ (x_1 + x_2)S(T) - x_1K_1 - x_2K_2, & K_2 \le S(T) \le K_3\\ (x_1 + x_2 + x_3)S(T) - x_1K_1 - x_2K_2 - x_3K_3, & K_3 \le S(T) \end{array}\right. \end{equation}
Each of these should be non-negative if an arbitrage opportunity is to exist. If we set $S(T) = K_3$, then we get $(x_1 + x_2)K_3 - x_1K_1 - x_2K_2 \geq 0$. Substituting the actual numbers in, we get $30x_1 + 10x_2 \geq 0$.
Now, here comes my confusion - the solution states that from the payoff description, we can also come up with a third condition $x_1 + x_2 + x_3 \geq 0$. I'm not sure how to arrive at this condition based on the current information we have. Any thoughts?