Portfolio theory has found its way in numerous applications for optimizing the electricity generation mix of a particular region. Existing models, however, consider typically a single time period and correspondingly do not properly account for actual dispatch constraints and energy sources with a non-dispatchable, variable output. This paper presents a portfolio theory model that explicitly distinguishes between installed capacity (power), electricity generation (energy) and actual instantaneous power delivery. This way, the variability of wind power and ramp limits of conventional power plants are correctly included in the investment optimization. The model is written as a quadratically constrained programming problem and illustrated in a case study. The results show that the introduction of wind power can be motivated to lower the risk on generation cost, albeit to smaller levels than typically reported in the literature. This wind power deployment further requires the need for sufficiently rampable technologies, to deal with its fluctuating output.