This paper presents a model for the large scale bias of dark matter haloes, showing that the unconditional mass function alone is sufficient to estimate the large scale bias factor. The model is tested against numerical simulations of SCDM, OCDM, and ΛCDM cosmologies, and is found to accurately predict the bias relation for both massive and less massive haloes. The results show that haloes more/less massive than typical M* haloes at formation are more/less strongly clustered than predicted by the standard Press–Schechter mass function. The paper also demonstrates that the peak background split can be used to compute the bias relation using only the unconditional mass function. The model is applied to the GIF simulations, and the results show that the large scale bias relation can be accurately predicted using the GIF mass function. The paper concludes that the bias relation is scale independent on large scales and that the results are consistent with the assumption that the effects of stochasticity on the bias relation are negligible. The paper also discusses the implications of the results for galaxy formation models and the reionization history of the universe.This paper presents a model for the large scale bias of dark matter haloes, showing that the unconditional mass function alone is sufficient to estimate the large scale bias factor. The model is tested against numerical simulations of SCDM, OCDM, and ΛCDM cosmologies, and is found to accurately predict the bias relation for both massive and less massive haloes. The results show that haloes more/less massive than typical M* haloes at formation are more/less strongly clustered than predicted by the standard Press–Schechter mass function. The paper also demonstrates that the peak background split can be used to compute the bias relation using only the unconditional mass function. The model is applied to the GIF simulations, and the results show that the large scale bias relation can be accurately predicted using the GIF mass function. The paper concludes that the bias relation is scale independent on large scales and that the results are consistent with the assumption that the effects of stochasticity on the bias relation are negligible. The paper also discusses the implications of the results for galaxy formation models and the reionization history of the universe.