
if you deploy weekly and pentest annually, you have 50 untested deploys for every test. think about that.
ship weekly = 52 deploys/year.
annual pentest = 1 evaluation/year.
51 of those deploys are untested at the time of the next test. and the next test will only find what's currently exploitable, not what existed for 6 months in between.
[the math on this gets ugly](refer link):
1/ annual model MTTD: ~180 days
2/ sprint-cadence model MTTD: ~14 days
3/ bugs caught early: 6× cheaper to fix
annual pentesting is compliance theater for any team shipping more than once a month. change my mind.