Abstract: Events with short lived progenitors such as core collapse supernovae (including LGRBs) should track the available star-formation and therefore looking at where they deviate from the star-formation is a powerful tool for probing their formation processes.  We compare the metallicity of the galaxies hosting LGRBs with those hosting Type Ic-bl, and Type II supernovae as well as the metallicity distribution of star-formation in the general galaxy population of the local universe to find strong evidence that LGRBs have an intrinsic preference for occurring in low-metallicity environments.  This requires looking at the star-formation in the universe in a way not biased by how it is associated into galaxies but instead with the metallicity of each galaxy weighted by the galaxy's star-formation rate.  We find that three quarters of the LGRB host population have metallicities consistent with the lowest metallicity tenth of the local star-formation and conclude that a low metallicity environment must be a fundamental component of the evolutionary process that forms LGRBs, despite the existence of some LGRBs in high metallicity environments.  Thus, while not a strict cutoff, we find that per unit star-formation an LGRB occurs about 30 times more often at metallicities below 8.3 on the KK04 scale (or about a third solar) than above.  Knowing this we can then relate the LGRB rate to the cosmic star-formation rate, and show that in low-metallicity environments one LGRB occurs for approximately every 40 Type Ic SNe with the broad-lined spectral features which denote the sub type of Ic supernovae associated with LGRB explosions.  Recent improvements in the population of LGRB host galaxies with measured metallicities and host masses allows us to investigate how the distributions of both these properties change with redshift and we find the metallicity distribution of LGRB host galaxies shows a curious lack of evolution.  This is at odds with the general evolution in the mass metallicity relation, which becomes progressively more metal poor with increasing redshift.  Using the evolving mass metallicity relation we can predict the expected metallicities of LGRB hosts given their known masses and redshifts and we find that the the expected metallicities also do not evolve, thus requiring the LGRB host galaxy mass distribution to increase with redshift so as to preserve the LGRB metallicity distribution as the mass metallicity relation decreases.  We also find that the metallicities estimated from mass and redshift are about twice as metal rich as the population with actually measured metallicity values, and this resolves much of the difference between our claims that the LGRB formation metallicity cutoff is at about a third solar with the cutoff value of approximately solar claimed in Perley et al. 2016 in favor of the former.  As LGRB hosts do not follow the general mass metallicity relation, there is no substitute for actually measuring their metallicities!