We examine how accurately the price-earnings, price-to-sales and market-to-book ratios of past comparable firms can predict the same multiples of IPO firms in Taiwan. The results show that the first two multiples predict IPO values fairly accurately. For the price-earnings (PE) models, forecasted earnings do not provide better firm-value estimates than historic earnings, implying the former has a dubious quality. The historic earnings PE model supplemented with growth information outperforms the forecasted earnings PE model even more, further confirming the quality problem of forecasted earnings. Overall, including growth and profitability as additional variables increases the forecasting power for IPO values, suggesting additional pricing adjustments are required on these two prospects.