This paper presents a new methodology to adaptively shift the dependability and security bias of the transmission system protection logic depending on the system state. Contingencies such as generator outage, line outage, and line fault can induce stressed system condition. Under stressed conditions, protection logic should be more biased toward security. Phasor measurement unit information is used at the system protection centre to estimate the system state. A data-mining model known as random forest is utilized to accomplish the task of state assessment. Two different protection logics are used to make the final relaying decision. The first protection logic is the existing distance relay and the second protection logic is based on wide area information. Both the protection logics are connected though logic gates to make the final relaying decision. The performance of the proposed scheme is validated on the IEEE-39 bus New England system and 246-bus Indian Northern Regional Power Grid system. The test results indicate that the proposed algorithm can help in shifting the dependability-security bias of the protection logic adaptively, which will help in mitigating cascaded outages in the power transmission system. © 2010-2012 IEEE.