In an earlier paper (Skoog and Ciecka, 1998), we unmasked and clarified the methodology embedded in the worklife expectancy tables produced by the accounting firm Coopers and Lybrand (by Charles Sherfey) under contract with the Association of American Railroads (AAR) in 1995. In addition, we provided a set of exact-age worklife tables based on a similar, but improved, methodology; and we provided a worklife table based on a Markov process model (increment-decrement methodology) of railroad workers. Those tables were based on data from the Twentieth Actuarial Valuation (TAV, 1997) of the US Railroad Retirement Board, and the increment-decrement table was additionally based on 1992-93 transition probabilities. Here, we update our Markov table for railroad workers by using mortality and retirement data from the Twenty-Second Actuarial Valuation (TSAV, 2003) and transition probabilities used by Skoog and Ciecka in their recent work on years of labor market activity (2001a, 2001b, and 2002). We provide a new table of AAR type worklives based on the TSAV.