How not to lie with statistics: the correct way to summarize benchmark results

@article{Fleming1986HowNT,
  title={How not to lie with statistics: the correct way to summarize benchmark results},
  author={P. Fleming and J. J. Wallace},
  journal={Commun. ACM},
  year={1986},
  volume={29},
  pages={218-221}
}
  • P. Fleming, J. J. Wallace
  • Published 1986
  • Computer Science
  • Commun. ACM
    • Using the arithmetic mean to summarize normalized benchmark results leads to mistaken conclusions that can be avoided by using the preferred method: the geometric mean. 
    View on ACM
    cse.unsw.edu.au
    Top 3 of 356 Citations
    Characterizing computer performance with a single number
    J. E. SmithCACM1988
    188
    How to assess and report the performance of a stochastic algorithm on a benchmark problem: mean or best result on a number of runs?
    M. Birattari, M. DorigoOptim. Lett.2007
    46
    Computer Architecture - A Quantitative Approach, 5th Edition
    J. Hennessy, D. Patterson1996
    1230

    Tables and Topics from this paper.

    Tables

    Explore Further: Topics Discussed in This Paper

    356 Citations
    Citation Type

    Citation Type

    Characterizing computer performance with a single number
    • 188
    How to assess and report the performance of a stochastic algorithm on a benchmark problem: mean or best result on a number of runs?
    • 46
    • PDF
    Computer Architecture - A Quantitative Approach, 5th Edition
    • 1,230
    • PDF
    Computer Architecture, Fifth Edition: A Quantitative Approach
    • 829
    Scene Text Recognition Using Similarity and a Lexicon with Sparse Belief Propagation
    • 186
    • PDF
    Software Measurement: A Necessary Scientific Basis
    • 576
    The memory fragmentation problem: solved?
    • 218
    • PDF
    Memory Systems: Cache, DRAM, Disk
    • 587
    • PDF
    Measuring forecasting accuracy : the case of judgmental adjustments to SKU-level demand forecasts
    • 99
    Scientific benchmarking of parallel computing systems: twelve ways to tell the masses when reporting performance results
    • 131
    • PDF