Bonn1997 wrote:meloshouldgo wrote:Lottery numbers cannot be predicted using statistical models. Each lottery is a new event, the past outcomes have no impact on it. On the other hand, human performance around team sports can be better analyzed using statistical models. How a player played in the last few years is s good predictor of how he will play next year.
Exactly but I don't think you'll convince him at this point. There's a huge difference between purely random events and events that can be predicted pretty well but imperfectly.
Bleacher Report 2015-16
= within 5 wins/losses of prediction
= more than 5 wins/losses of prediction
Hawks prediction 50-32, Actual 48-34
Celtics 43-39, Actual 48-34
Nets 29-53, Actual 21-61
Hornets 30-52, Actual 48-34
Bulls 51-31, Actual 42-40
Cavs 54-28, Actual 57-25
Mavs 32-50, Actual 42-40
Nuggets 24-58, Actual 33-49
Pistons 37-45, Actual 44-38
Warriors 60-22, Actual 73-9
Rockets 56-26, Actual 41-41
Pacers 39-43, Actual 45-37
Clips 57-25, Actual 53-29
Lakers 25-57, Actual 17-65
Grizz 52-30, Actual 42-40
Heat 46-36, Actual 48-34
Bucks 42-40, Actual 33-49
Timberwolves 22-60, Actual 29-53
Pelicans 46-36, Actual 30-52
Knicks 28-54, Actual 32-50
Thunder 59-23, Actual 55-27
Magic 27-55, Actual 35-47
76ers 20-62, Actual 10-72
Suns 37-45, Actual 23-59
Blazers 25-57, Actual 44-38
Kings 35-47, Actual 33-49
Spurs 55-27, Actual 67-15
Raptors 48-34, Actual 56-26
Total somewhat accurate predictions - 8
Total not even close predictions - 22
Percentage of accuracy = 8*100/30 < 27%
Average margin by which predictions were off = Add all margins / 30 = 240 / 30 = 8 wins/losses
Knicks 2016-17 prediction 38 wins. This means we should not be surprised if we win 30 to 46 games. Oh wow great. I could have predicted that in my sleep.
Is this your definition of imperfect prediction, but "predicted well"??? GTFOH.
Nixluva is posting triangle screen grabs, even when nobody asks - Fishmike. LOL
So are we going to reference that thread like the bible now? "The thread of Wroten Page 14 post 9" - EnySpree