I have seen a good deal of analysis regarding Barry Bonds reaching 600 and 700 home runs and surpassing Mays, Ruth, and Aaron using the Bill James "Favorite Toy." The Favorite Toy is a great, fun tool. However, I have some reservations with using it exclusively for predictions.
I feel that it sees things too linearly to be an effective analysis tool. Basically, it takes the number of years remaining based on age and the number of home runs in the last three years, with each year weighted appropriately, calculates the expected home run total for the player and then determines the likelihood of reaching a certain preset threshold. I just feel that the expected home run totals are calculated way too uniformly for all players.
Barry Bonds is an exceptional home run hitter and to base his expected home runs based on his age without taking into account his performance level is too limiting. A player who hits 30 HRs at age 38 will not "age", I believe, the same as a player who hits 5 at 38. The 5-HR player will reach, or already has reached, replacement level and will not be long for the baseball world. A 30-HR player generally has farther to go. I know that Favorite Toy takes into a account the player's performance level but it assumes that that level has no bearing on the player's reamining playing time.
I am trying to devise an expected home run total for Bonds (and strikeout total for Randy Johnson) that unites the idea of the Favorite Toy with a weighting for a player of Bonds caliber. I have a plan for deriving the data and just have to find time to do it. I'll keep you posted.