soireeculturelle
Registered User
- Jan 7, 2014
- 57
- 0
Was having water cooler talk with a coworker, and he was talking about a bet he made years ago with a friend about a certain 24-year old centerman putting up 69 points in a season by the time he is 25. Pulled up some comparables for him on hockey reference in terms of games played before age 25 and points per game up to that point. Seems like my buddy is in tough with his particular prediction.
Anyway, when it comes to predicting future results, what method do you find to be more accurate? 1) A weighed average of the players' past result, as proposed here: http://www.sbnation.com/nhl/2014/1/23/5333626/nhl-projecting-scoring-rates
2) or start a baseline prediction (league average for whatever group of player you are projecting - let's say PPG of centermen aged 24 who have played 4 NHL seasons) and then adjust from there. Was reading Thinking Fast and Slow and it seems that this is the approach that the author endorses since it better reflects regression to the mean in an individual.
Anyway, when it comes to predicting future results, what method do you find to be more accurate? 1) A weighed average of the players' past result, as proposed here: http://www.sbnation.com/nhl/2014/1/23/5333626/nhl-projecting-scoring-rates
2) or start a baseline prediction (league average for whatever group of player you are projecting - let's say PPG of centermen aged 24 who have played 4 NHL seasons) and then adjust from there. Was reading Thinking Fast and Slow and it seems that this is the approach that the author endorses since it better reflects regression to the mean in an individual.