Analysis of Negative-Weight Events in Monte Carlo Generation with Next-To-Leading-Order Parton Distribution Functions

In a particle physics analysis, Monte Carlo simulation is one of the most important steps, as it allows for an estimate of the cumulative effects of a particle’s interaction with detector material, allowing an accurate estimate of event selection efficiency and background expectation. The CMS collaboration uses the parallel processing system CRAB to generate Monte Carlo events. Events go through four steps: GENSIM, DIGIRECO, AOD, and MINIAOD, each step compressing the events to a more workable amount of memory.The parton density function, or PDF, describes the distribution of particles within the colliding protons that is essential to generating Monte Carlo events. However, PDF's are determined by a complicated mix of theory and experiment, and different PDFs yield different results. Recent next-to-leading-order PDF's have led to the production of events with a negative event weight, representing events that should be subtracted from the analysis. This raises the disconcerting possibility that the total expected number of events in a channel could be negative, especially in the high-mass region of interest. We examine Monte Carlo contact interaction events for negative-weight events and determine their effect, if any, on the analysis.