With core counts as well as heterogeneity on the rise, the sequential components of applications are becoming the major bottleneck in performance scaling as predicted by Amdahl's law. We are therefore faced with the simultaneous problems of occupying an increasing number of cores and improving sequential performance. In this work, we specifically focus on improving the energy efficiency of sequential algorithms through the n-way programming model. In previous work, we introduced the n-way programming model which seeks to exploit the algorithmic diversity present in certain computations in order to speedup or improve the quality-of-result. The core idea behind n-way parallelism is to launch a number of instances of a key computational step and benefit from either the algorithmic diversity present in the algorithm or the diversity in algorithms available to express the computation. In this paper, we propose to combine metrics measuring the algorithmic progress of a computation with metrics measuring the energy expenditure to compute an efficiency metric. This metric can then be used to quickly pick which instance of a n-way computation is the most energy efficient and cull the inefficient ones. The evaluation of our idea on sorting benchmarks shows that our technique is promising.