Pair trading on the German intraday power market is a commonly used risk-averse, heuristic trading strategy. However, due to myopic decision-making and a lack of foresight, the profit obtained is far from optimal. By comparing this strategy with the ex post optimal solution (ie, a strategy with perfect foresight), we show on a set of 15 selected days from 2020 to 2022 that the predictive information lets us generate on average more than five times as much profit by excessively buying and selling the same contracts for a trading interval of five minutes. Another problem with pair trading in practice is the possibility of unbalanced auction wins. We show that an unbiased loss of up to 10% has a negligible impact on the obtained profit. In contrast, we also show the value of frequent optimization updates by simulating strategies with only sporadic participation in the market. While this is hardly beneficial for the pairtrading strategy, the ex post optimal profit increases on average by 30% when the time between two trades is halved.