Quote:
Originally Posted by Hockey-and_stuff
It's not a flaw, it's simply a constraint of the available data.
All NHL teams play 82 games. Not all NHL teams play playoff games, and not all playoff teams play the same number of games.
These models are just numbers, and sometimes you have to make concessions to make the data more consistent across the board.
There will always be players that beat the
projections and there will be players who fall short. The hope with the model is that as you accumulate historical data the variances get smaller and your model works with a higher degree of accuracy.
Sent from my Pixel 7 Pro using Tapatalk
|
If the constraint results in poor or misleading results, because it is eliminating key or valuable information, then yes, it is a flaw.