We explore the factors associated with historical IPO failures by developing an IPO failure prediction model that includes accounting information as well as proxies for the role of information intermediaries and other IPO deal-related characteristics. We document statistically significant differences in failure models applicable to nontech versus high tech IPOs, and these structural differences are largely driven by accounting-based proxies for firms' investments in intangible assets, operating performance, and financial leverage. We also develop parsimonious, predominantly accounting-based, strictly out-of-sample (i.e., no hindsight) IPO failure forecasting models for each of the two sectors. We find that our forecasts are negatively associated with one-year post-IPO abnormal returns. A pseudo-hedge strategy of going short (long) in high (low) failure risk portfolios yields returns of economically significant magnitudes over the one-year horizon, and is robust to alternative returns methodologies. Further results suggest that IPO long-run returns anomalies may persist, but they take different forms for high-tech and nontech IPOs. Copyright University of Chicago on behalf of the Institute of Professional Accounting, 2007.