I was reading the article on H2H vs. cumulative with interest. I primarily play H2H (both Roto but more recently points based). That said, I have wondered about how to apply the randomness. I have concluded as you did that players that are simply hot/cold some weeks are part of the game. There is no predicting when a player is going to be hot or cold, and quite frankly I am not certain if the “consistent” player or the “hot/cold” type player is better. Overall I agree that it is not worth trying to follow and integrate into the stats.

That said, I am not sure if I agree with your Chase Utley example of an injury prone player. As I start going down the draft, I think a valuable strategy can be picking some high value stars susceptible to injury. The key is wanting a higher average points per game played. Unlike the hot/cold player, I know when my Chipper Jones type is on the disabled list. While they are on my bench, I can slip in a replacement level player and still get my points/counting stats.

To account for this, a useful option would be at least know the number of games assumed that the player will play as part of the projections. An even better option would be the ability to sort/rank based upon expected points per game.

Instinctively it seem that would be easy (stats divided by games), although while writing this sentence it occured to me that there would likely need to be some option to filter out based on a minimum number of at bats. Otherwise, there could be some very obscure minor leaguers or pinch hitters jumping out in the rankings. In addition, there could be complexities regarding non-counting roto stats that do not appear with simple points.

Just my thoughts. ALways searching for that magic tool and your program gets as close as any I have found. Keep up the great work. Looking forward to next season already.

mjfhokie