[Flash Walken]

Originally Posted by GranteedEV

So a sample size of one season is too small for someone to win an Art Ross trophy because it's too small to show if someone is a Hall of Famer? Differences in occurrence of different events can mean different sample sizes as useful.

You're talking about a sample size in which shot attempts were taken both ways. I'm not pretending that's huge but it's sure as #### bigger than a sample size in which zero goals were scored for either team. Goal scoring events are highly rare and more dependent on factors which skaters can't necessarily control.

Then why not leave it at that? Sample size inconclusive. Unknowable based on Jagr and Versteeg both being injured early in the season and other issues on the roster preventing them from getting solid icetime with Bennett? Isn't that the natural conclusion you should take?

Because one is a measure of process and one is a measure of results.

Corsi is a measure of results. You're using data from events that have already occurred to create the data set. You are measuring different results like shots to create substance where a measurement of results like goals and assists don't exist. You're saying a positive corsi over 16 minutes is a large enough sample size to conclude they were successful as a line. I'm saying getting more shots than the opposition over 16 minutes is not an illustration of success as a line, it's too small of a sample size. The poster who suggested they failed as a line may or may not be wrong based on the criteria he is using, but pointing to a positive corsi for over 16 minutes of play time is basically the ultimate form of statisical pedantry.

What is the raw data for their corsi rating? Did they get 2 more shots during their 16 minutes than they allowed? Is that substantive of anything?

Saying a line failed in a small sample size suggests the process was poor. But we see that no, the process was there. We don't know if that process was sustainable or bound to produce results, but we do know it was the kind of process you want from your "third line". This line was producing 58% of the shot attempts when the rest of the team was producing 45% of the shot attempts.

But again, 16 minutes.

If they were producing 58% of the shot attempts as compared to 45% for the rest of the team, but the rest of the team was producing goals in those 45% and the line generating 58% wasn't, that's an indication they failed as a line, is it not? Ultimately games are judged by goals and not corsi, right?

That's not what I'm using my post to state. I couldn't possibly tell you whether "1993 Jagr and 2010 Versteeg" would have had any results in that small sample size. I am telling you the following, and I'll even bullet point them out just for you:



- 16 minutes was an inadequate sample size to say this line failed regardless of any statistics, positive or negative.



- However, the underlying metrics of this line were strong - which is reason to have kept them together rather than split them apart after only 16 minutes together


- The supplied video of one of the games this line played shows the underlying numbers were accompanied by some high quality chances

Was the line 'split apart'? I don't even know. Do you know? What happened to the other two players on that line when they were separated? Were their corsi stats better or worse playing with Bennett, and does that reflect the tangible offense they produced?

Did Jagr generate more assists in 16 minutes playing with Gaudreau and Monahan or did Versteeg generate more offense playing 16 minutes with Tkachuk or Backlund? Where did these players go and what did they do away from that line? How can you know it wasn't a good idea to split the line up if you're not supplying the tangible results from when the lines were in fact split up?

- A larger sample size however, would not have assured truly high-end results, because over a large sample size, Jagr's shooting percentage had fallen off to 8.0% over his final 104 NHL games, contrast that with his shooting percentage he had the two year priors - 14.1% and Kris Versteeg was injured to the extent of only playing 24 games last season, none of which he was necessarily healthy for. However, had Jagr been the player he was two seasons ago and Versteeg been the player he was just one season prior, and Bennett playing exactly as he had, there is reason to believe that the potential for a strong line combination existed.

Like...what on earth does this have to do with anything? What if Bennett wasn't Bennett and was actually Ryan Getzlaf playing with 2011 Jagr and 1957 Beliveau?

- Put together, The underlying numbers and eye test suggest this line "not working out" has to be taken as a myth. They weren't together long enough to be evaluated even though the underlying metrics and eye test indicated they should have been kept together longer.

1. What happened to the players once they were removed from the same line with Bennett?

2. What happened to team results when they were removed from a line with Bennett?

3. What happened to Bennett when they were removed from a line with Bennett?

Really, you are simply being obtuse to my posts which is why you are not acknowledging their content and fixating on whatever detail supports your predetermined conclusion (which, of course, is why you made this thread).

No, I'm being sarcastic because I think these types of posts warrant it.