During the design of complex telecommunication systems one can run across a number of stunning questions. These questions are mainly motivated by the goal to achieve performance values in real system implementations that are as close to the theoretic optimum as possible whilst still maintaining reasonable complexity. But how far can one get and what amount of complexity is reasonable? The conjunction of these two questions can be answered neither solely by information theory nor by estimation theory alone. Both of them have to be put into relation to answer these questions.When both views are brought together a bunch of new questions comes up, which this thesis tries to address:. Systems using multiple antennas at transmitter and receiver promise extraordinary capacity gains. But what amount of this theoretic capacity can be realized? Where are the limits?. The rate of change in the channel, i.e., number of independent channel realizations, can be identified easily as a key parameter. But how does the structure of the channel influence the former the achievable capacity gains in addition to this?. Ergodic capacity or mutual information is an interesting quantity to characterize channels but outage probabilities and in actual realizations even block error rates are at least as important as channel capacity. How can they be characterized?. Speaking of outage probabilities and block error rates: How does the amount of redundancy inherent to some transmission schemes influence the answers to the questions above?. Is a system with an iterative ARQ scheme that adds information until error free decoding becomes possible a practical way to achieve rates close to theoretic capacity?. Going from link level to system level: How do multi-user scenarios influence the above questions? Can the consideration be extended to systems with scheduling? And can the concept of smart scheduling and dumb antennas be evaluated using the same techniques as before?. Are there more advanced techniques or is synchronized detection really the ultimate method?