Towards cosmological constraints from the compressed modal bispectrum: a robust comparison of real-space bispectrum estimators

Higher-order clustering statistics, like the galaxy bispectrum, can add complementary cosmological information to what is accessible with two-point power spectrum. While standard way of measuring bispectrum involves estimating a value in large number Fourier triangle bins, compressed modal approximates as linear combination basis functions and estimates expansion coefficients on chosen basis. In this work, we compare two estimators by using parallel pipelines analyze real-space halo measured suite $N$-body simulations corresponding total volume $\sim 1{,}000 \,h^{-3}\,{\rm Gpc}^3$, covariance matrices estimated from 10,000 mock catalogs. We find that yields constraints are consistent competitive analysis: for bias shot noise parameters within tree-level model up $k_{\rm max} \approx 0.06 \, (0.10) \,h\,{\rm Mpc}^{-1}$, only 6 (10) necessary obtain equivalent estimator $\sim$ 20 1,600 depending bin width. For have implemented pipeline Markov Chain Monte Carlo first time, discuss detail how parameter posteriors robust to, or sensitive several user settings pipeline. The highly efficient compression achieved catalogs available allows us quantify our depend mocks used estimate functional form likelihood.