Offshore arctic pipelines must be designed with adequate reliability against damage or rupture due to the effects of ice features that gouge the seafloor. A common design approach is to bury the pipeline sufficiently deep to avoid contact by a gouging ice keel of a target rare return interval. The effects of sub-gouge soil displacements on pipe stress or strain are also assessed. It is implicitly assumed, in this traditional approach, that ice keels have infinite strength and momentum, so that gouge depths are not limited by ice keel failure and direct contact of ice on pipe results in rupture. However, many first-year ice features may not be strong enough to gouge the soil to the extreme gouge depths. Considering the high cost of incremental pipeline burial depth, it may be desirable to account for limits imposed on extreme gouge depth by soil resistance. Introducing these limits, however, gives rise to additional uncertainty and requires the development of a reliability framework to assess the consequences of direct ice contact on the buried pipe. This paper presents a methodology for evaluating the probability of ice gouging non-performance of a pipeline while explicitly accounting for the ability of the seabed soil to limit ice gouge depth based on the relative strength of soil and ice keel. Non-performance is defined either as ice contact on the pipe or exceedence of some level of acceptable pipe strain when contacted by ice. A three-step approach is followed through the use of nonlinear finite element analysis to estimate pipeline capacity, the statistical characterization of demand from ice gouging events, and the development of a reliability framework combining demand and capacity. Parameters governing the capacity of the pipeline are identified, the sensitivity of the pipeline reliability to various parameters is investigated, and the impact of allowing ice contact on pipeline reliability is presented. Relative merits and risks of allowing ice contact are discussed in the context of pipeline reliability. The presented results and discussions are believed to be of significant help in the development of burial depth criteria for future offshore arctic pipelines, and certainly in the reassessment of existing ones.