This paper is an exploration of subprime mortgages over the cohorts from 2000 through 2006, especially those prior to 2004. In particular, this study contrasts subprime originations during the “boom years” of 2004-2006 with originations during an “early period” of 2000-2002. We develop a counterfactual technique to determine how originations during the early period would perform in a different environment, namely, the environment faced by originations in 2004, 2005, and 2006. We find that representative originations during the early period of 2000-2002 would not have performed significantly better than representative originations in 2004, 2005, and 2006. This result is robust to counterfactual exercises for originations with different LTVs. We conclude that mortgages of early cohorts were no less vulnerable to the environment faced by cohorts of 2004-2006.