[CorsiHockeyLeague]

Tons of stuff on this page that isn't just questionable but objectively wrong... why people don't educate themselves a bit before forming strong opinions is always baffling to me. I mean you can have a skeptical viewpoint about what the stats are telling us but if you don't even understand what they're saying it's kind of fruitless.

Originally Posted by Samonadreau

Away from Tanev he probably had easier opponents.

Nope, even after you adjust for opponents it was still better. I think Tanev was just done in Vancouver.

Originally Posted by djsFlames

Hughes is bad defensively.

Even worse than Darnell. So hey, the Oiler apologists have that to cling to.

Arguably - if you average out the past 2 years, then Hughes is better defensively. Either way they both come out nearer the bottom of the league than the top, though, so it's not exactly a good argument to win.

Originally Posted by Gordies Elbow

The primary arguments when defending are around assumptions. Some of the assumptions added to these models are subjective (e.g. what constitutes a turnover, what is a proper zone entry.)

True in the case of a turnover. I'm not sure there's much subjectivity in a zone entry. The puck crosses the line or it doesn't.

Some make assumptions about specificity (e.g. shot location makes a shot harder or easier to save.) These may seem obvious, but are they statistically significant given the small numbers of minutes of performance?

As discussed earlier, there are thousands and thousands of shots of data from everywhere in the offensive zone. So the answer is yes - it's very clear that shot location influences save percentage, and by how much.

Descriptive models can be tested against current performance, predictive against future events. In order to be validated, they should be tested. Are any of these?

Yes. Or at least, the guy who created the one we're talking about says he's tested it, and I don't know why he'd lie. Dom L from the Athletic runs his model thousands of times and then evaluates its performance against the results and against other models at the end of the year to determine whether or not it'd make money if you used it as a betting tool.

Originally Posted by Gordies Elbow

As to the shot attempt rates, if the shots aren't converted to goals and by proxy, to wins, does it matter?

The question doesn't make sense. Of course they're converted to goals and goals to wins. We know how often this happens to a high degree of certainty. There is lots of data to support this.

To argue your assumption - are "complete" players better than "incomplete" players in all cases?

No, because the object is still to score and prevent goals. If you can score so much that it doesn't matter that you suck at preventing goals, you'll still be a positive impact player. See: Draisaitl, Leon.

That being said, goal prevention is more valuable than goal scoring because of how the NHL's standings work - the OT point incentivizes prevention. Every analytical model I am aware of overrates the importance of scoring (especially goals) because they're more reliably predictable from year to year. So that needs to be taken into account (and it's acknowledged by the people who do the modeling; the guy who does the model Jfresh uses says his work probably overrates goal scorers and underrates people who make "dangerous passes", which is why MacKinnon is rated below Matthews).

Originally Posted by Shazam

These advanced stats aren’t models. There’s no predictive analysis. People could even take past data and model it forward and match it against past real data but I’ve never seen that done.

It's done all the time. Literally everything in this post is wrong. Amazing.

I’ve seen some terribly naive analysis done (mostly by the Nation) where they attempt to find things out like when players start to decline but it was done all wrong.

Point to it and explain why it's wrong?

Originally Posted by PepsiFree

His career-high is a 50 point pace over 82 games. Every other defenceman with that kind of production and that kind of contract is extremely good defensively. Every other defenceman with that kind of production and that kind of defensive ability is being paid closer to $4-5M.

And even then, some of those guys are probably being overpaid because their scoring doesn't offset the plays they allow the other way - scoring gets you paid, being responsible doesn't, even though the opposite is the way it should be in terms of winning. See Adam Pelech.

Originally Posted by edslunch

Sure, but IIRC most advanced hockey stats have an r-squared of around 0.3 which is considered a weak correlation statistically, so clearly the models miss a lot of context too. All I’m saying is that these analyses would be more complete if you include the actual results.

Actually, the model we're talking about has an R2 of .888, although that's at the team level. It's obviously going to be less accurate at the individual player level, but that doesn't mean it doesn't have any utility. It just means that if you use it as a predictor of next season, it's going to be close in many players' cases, a bit off on some others, and wildly incorrect on a few.

Originally Posted by CNeufeld

I like that Edmonton has collected 3 of the bottom 11 players on that list... Is that wrong?

C

Gets considerably worse if you take into account the fact that Cody Ceci is probably going to join them now that he's not on the Pens - he was terrible in Ottawa and Toronto. It was posted earlier in the thread, but it's still the projected opening night lineup...