Quote:
Originally Posted by Philly06Cup
I'm still unsure why there is a stigma with 'advanced hockey statistics.' These analytics are trying to REDUCE noise and form a more accurate picture of the statistics. Empty-net goals should count for less. Secondary assists should count for less. Puck possession is important. Inflated shooting percentages should be discounted. Surely we can agree with most of these premises and should try to create a better model than 'GOALS, ASSISTS, PLUS/MINUS.'
I remember when NBC was trying a new way of interpreting power-play efficiency. On their screens, they would display 'one powerplay goal in --:-- of powerplay time.' This methodology is way more precise than the conventional 'powerplay goals / powerplay chances.' So many PP chances are abbreviated by another penalty; I'm still not sure how a 5 minute major penalty is counted in terms of PP chances. The previous methodology also correctly identifies that teams who score quickly on the PP are likely better at the PP than those who take 1:59 of PP time to score a goal. Alack, NBC reverted back to the antiquated PP% method.
Advanced NHL statistics are simply BETTER than current NHL statistics. That's why they're called 'advanced.'
NB -- I was the guy who asked Burke that question. (I also asked him who would win the Stanley Cup)
|
The answer is super simple. Advanced stats, while valuable, are only able to tell historical data in hockey. Some people try to use advanced stats to predict trends, but because hockey is a series of random events, it's almost impossible.
Advanced stats are highly influenced by what line or team a player plays on, unlike baseball where advanced stats are not influenced by multiple other players.
It's not that advanced stats are bad, it's that they have limited value in hockey.