[Jay Random]

Originally Posted by CorsiHockeyLeague

This is a pretty ridiculous thing to say.

Let's say a team, say the Flames, has played 59 games and have 68 points. Let's say another team, whoever, has played 60 games and has 67 points.

Now let's say the Flames' remaining 23 games are all against a combination of the Blackhawks, Blues, Ducks, Lightning, Penguins and rangers. The other team plays multiple games against the Sabres, Oilers, Coyotes, Blue Jackets... you get the idea.

Who has the higher chance of making the playoffs?

But that is not the case here. That is not a significant factor.

Now assume that the other team has a cushy schedule with 17 of their last 22 played at home, and in the majority of those games, the team coming in played the night before and is on the tail end of a long road trip. The Flames, meanwhile, play most of their games on the road as part of brutal 3-in-4 stretches with back to backs involving long flights in between.

Who has the higher chance of making the playoffs?

But that is not the case here. That is not a significant factor.

Now assume that some time this afternoon, Brodie, Gio, Wideman, Monahan, Gaudreau and Hudler are all simultaneously struck by lightning and are out for the season. The other team is totally healthy.

Who has the higher chance of making the playoffs?

But that is not the case here. That is not a significant factor.

You know the answer. You know there are myriad factors that go into this.

Right. But your model, as described by your own website, uses Fenwick as the predominant factor. And you didn't even mention it in your long-winded baloney defence cited above.

In other words, you didn't even begin to address the actual criticism of your methodology. You blew smoke about a bunch of factors that have nothing to do with the point in question.

Basically, you are predicting that the Flames will miss the playoffs because their possession-proxy numbers are bad. But you would have predicted, based on those numbers, that the Flames would be dead last in the league this season, or close to it. You would also have predicted that the Oilers would be somewhere near the middle of the pack. When both these things consistently fail to happen, your only defence is to say that the difference is accounted for by luck. But in fact puck possession is not the sole determinant of success in hockey, and Fenwick and Corsi don't even measure puck possession.

Your model can't even predict the past. Why should we trust it to predict the future?

The only argument is how much relative weight to put on those factors to create the best predictive model you can - the one most likely to predict the final result or closest to it (since no model designed to predict the future in hockey is going to be even close to perfect).

Right. And you have put the heaviest weight on Fenwick, which is not a good predictor. Nor is Corsi.