Traditionally, capacity has been procured to meet system peak load with no formal regard for flexibility characteristics. However, this planning paradigm is shifting, with increasing variable energy resource penetration, to consider system flexibility requirements in addition to peak capacity. Flexibility is broadly defined as the capability of system operators to respond to changing loads net of intermittent renewable generation. It can be provided by energy storage, but also by gas-fired generation, enhancement of existing resources, responsive loads, new or redefined ancillary services and operational rule changes. We show that traditional cost of new entry (CONE) capacity planning metrics fail to fairly represent the net cost of different types of flexible capacity resources. We contribute to existing literature by using a mixed integer linear program to (1) optimize the dispatch of three bulk energy storage technologies and a conventional combustion turbine (CT) and (2) co-optimize their dispatch as price-takers first in day-ahead and subsequently in real-time energy and ancillary service markets. We then define a "flexible capacity" CONE and demonstrate, that even without premium "pay-for-performance" payments, bulk energy storage can be cost competitive with a CT as a flexible resource in today's markets. (C) 2014 Elsevier Ltd. All rights reserved.