We analyze the crush-down collapse of One World Trade Center (1~WTC, North Tower) in the framework of the National Institute of Standards and Technology (NIST, 2005) collapse hypothesis. The main feature of crush-down is that a moving part of the building - the top section - falls onto the stationary base, and absorbs the mass in the way. We extend the Ba\v{z}ant-Verdure-Seffen (BVS) model of crush-down~(\bv, 2002; Seffen, 2008), where we split the crushing front in two, one at the core and to other at the perimeter of the building. We fit the BVS and the split-front crush-down model to recently published roofline motion data (MacQueen and Szamboti, 2009), to find detailed variation of crushing force $\FC(Z)$ in the storeys 97 and 96, and the average crushing force $\mean{F^C}$ in the remainder of the impact zone (storeys 95 to 93) and in the base below for the remainder of roofline data (storeys 92 through 87). We show how within the NIST hypothesis and the BVS model $\ec{}$, defined as $\ec{} = \mean{F^C} / \FC(0)$, requires a correction factor of 1/6 to match the data. We construct a Controlled Demolition (CD) hypothesis which avoids this and other correction factors through two assumptions: $(i)$ the top section is twice as massive as what it appears to be, where its core stretches initially down to the 75-th storey; and $(ii)$, the collapse starts with the wave of massive destruction which annihilates the core below the 75-th storey and separates the top section from the base below the impact zone, following which the top section falls to the ground opposed mostly by the perimeter columns, which strength is approximately a third of the total strength. Within the CD hypothesis we achieve excellent agreement between Ba\v{z}ant-Verdure model of crushing force and the data.