Almost no predictive model for any industry can identify all the statistically relevant variables. However, that doesn't mean the identified and statistically significant ones should be ignored either.
To make an analogy on it, if you are driving to work today, you most likely do everything you can with the variables you have control over which are most likely to lead to success of efficiently arriving to work on time. Some notable variables would be departure time, speed, route, driver and vehicle condition, etc. However, there are unknown variables that can intervene and play a major role in the trip (other drivers, traffic, cops, road conditions, etc.) that are generally unaccountable to control methods.
- jfkst1
So what do you do if you want to make a statistic of time it takes to get to work?
Do you average out your times over a period? Do you note down effects that made the trip longer?
You can collect all data & make the statistic accurate by noting down each event that transpires daily & form an accurate mathematic summary of the voyage. Then with all collected data you can form an estimation of the time it 'will' take before you leave - that is, if you have a traffic report of what it's like on that day, is it raining that slows people driving, accidents, etc.
You can predict with confidence what you your travel will involve. There are variables that can not be taken into account - blow tire which may occur 1 in every 100000 voyages but that can be taken as chaos.
The analytical data that I have seen in regards to the NHL is no where near as accurate nor does it detail much of the important data that affects that particular result - player Corsi is a perfect example. The data is there to be collected it just hasn't been as yet. It appears through comments I've seen, that the stats people are waiting on the smart chip for puck & player movement. They don't have to, the data is all there & can be amassed by the naked eye - how many times a player touches the puck, what zone the possession takes place, the shots taken by player RHS LHS, the goalies saves technique & goalies GA - goals can be a variable due to the same shot not scoring every time but the data would be interesting to see.
The NHL is getting on board & many bloggers already are but to say people that have doubts dont get it is wrong. I appreciate the data as it is but still don't take it as serious as most.
In science they can predict through mathematical algorithms, to this point no one is predicting anything with the advanced stats but I have heard them called science.