Overall Results

• Confounding Factors

  • No statistically large effect in weeks 1-3
    • Taken as a whole
    • Taken week by week

• Attack Graph

  • Good methodology

• Deception Works

  • Deception always did better
    • Increased time for progress in the attack graph
    • Attackers went down the wrong parts of the graph
    • Attackers believed they succeeded when they failed

Notes: