Halfway mark statistical analysis of top 20 point producers

[lomiller1]

This is inane. It doesn't matter if PP time is limited, we're looking at point shares here and a disproportionate number of points are accumulated on PP.

And PP points are primarily a function of the players role and how good the PP is. At best we could say maybe their PP is overachieving or underachiever and they should get a few more or a few less points, but the sample size is so small players/teams will be all over the map wrt to "expected"values so the noise you inject may overwhelm any benefit.

The OP captures the most important thing to capture is who gets the points on their teams PP because that probably won't change much unless someone gets injured.

 

[42]

If we look at the oiSH% of the top 5 point finishers in the last 3 seasons, we can see that the highest it's been is 12% and that 12% is a statistical anomaly with the average in the high 10%

Name, points, oiSH% in last three seasons:

2017/18

1. McDavid, 108, 10.5
2. Giroux, 102, 12
3. Kucherov, 100, 10.3
4. Malkin, 98, 10.3
5. MacKinnon*, 97, 11.9 (played 74 games)

2016/17
1. McDavid, 100, 10.9
2. Crosby, 89, 10.1
2. Kane, 89, 11.3
4. Backstrom, 86, 10.2
5. Kucherov, 85, 10.8
5. Marchand, 85, 9.3

2015/15
1. Kane, 106, 11.2
2. Benn, 89, 10.4
3. Crosby, 85, 9.6
4. Karlsson, 82, 9.6
4. Thornton, 82, 11.1

Even Crosby has never had an oiSH% higher than 11.3% when playing in at least 75 games.

Now scoring is up this year but not that much higher that we could expect the current oiSH% to keep over the entire season. Rielly is not going to finish the year with 16% nor Marner with 15% nor Rantanen with 12.9 % , to name just a few of the highest numbers in the top 20. Hence the expectation that these players will regress the most.

There is also the expectation that the top finishers will not average an oiSH% higher than the average of last 3 seasons by 1%. When the season is done, I expect the top 5 finishers to have an average of maybe something in the low 11%, accounting for the higher scoring this season.

 

[VoluntaryDom]

you can find oish% on the PP just saying

 

[VoluntaryDom]

And PP points are primarily a function of the players role and how good the PP is. At best we could say maybe their PP is overachieving or underachiever and they should get a few more or a few less points, but the sample size is so small players/teams will be all over the map wrt to "expected"values so the noise you inject may overwhelm any benefit.

The OP captures the most important thing to capture is who gets the points on their teams PP because that probably won't change much unless someone gets injured.

and yeah this, PP production is SOOOO systems dependent

 

[42]

you can find oish% on the PP just saying

Yes, I can. I'm still debating, though, which would be a more accurate predictor, PP career oiSH%, which tends to fluctuate a lot more from year to year than EV oiSH% or year-to-date PP production, which is what I used.

 

[VoluntaryDom]

Yes, I can. I'm still debating, though, which would be a more accurate predictor, PP career oiSH%, which tends to fluctuate a lot more from year to year than EV oiSH% or year-to-date PP production, which is what I used.

yeah its hard to say. i might look at PP oish% since start of last season or PP production in the same span to avoid an unsustainable scoring rate in a small sample (eg Huberdeau)

 

[The Apologist]

Tavares will get 28
Rielly 24
Marner 37

Seems like some pretty big drops in productions. Sounds like the Leafs offence may completely tank in the second half?

I think someone forgot to tell Mitch he's supposed to be falling off?

 

[42]

I think someone forgot to tell Mitch he's supposed to be falling off?

There is still half a season to go. Don't worry, he will fall off.

 

[The Apologist]

There is still half a season to go. Don't worry, he will fall off.

Im sure you're right. I fully expect expect JT to end up with only 22 points in 42 games too.
Only makes sense.

 

[ImNeverWrong]

Did Sean Monahan die in your fancy stats prognosticator?

 

[42]

Im sure you're right. I fully expect expect JT to end up with only 22 points in 42 games too.
Only makes sense.

Well, let's judge projections at the end of the season instead of after a game in which Marner scores 2 goals, and half the season still unwritten.

And yeah, I have already acknowledged that Tavares' prediction is too low, for reasons i went into and won't repeat here. But I feel much more confident about the Marner regression prediction.

 

[42]

Did Sean Monahan die in your fancy stats prognosticator?

Read OP. It clearly states that I only considered the top 20 point producers at the time of writing. You can figure out why Monahan isn't on the list.

 

[The Apologist]

Well, let's judge projections at the end of the season instead of after a game in which Marner scores 2 goals, and half the season still unwritten.

And yeah, I have already acknowledged that Tavares' prediction is too low, for reasons i went into and won't repeat here. But I feel much more confident about the Marner regression prediction.

Im pretty sure theyre both too low but yes, we'll see.

 

[Talain]

They're probably basing it off of last season's trends, when Judas basically mailed it in during the second half.

 

[The Apologist]

Been a while, how about an update?
Tavares predicted to end up at 68, currently at 62 with 25 games to go. 89 pt pace
Rielly predicted at 68, currently at 55 with 25 to go. 80pt pace
Mitch predicted at 87, currently at 67 with 25 to go. 97pt pace.

 

[42]

Been a while, how about an update?
Tavares predicted to end up at 68, currently at 62 with 25 games to go. 89 pt pace
Rielly predicted at 68, currently at 55 with 25 to go. 80pt pace
Mitch predicted at 87, currently at 67 with 25 to go. 97pt pace.

I'll do an update at the end of the season. As far as Tavares goes, I know the prediction was much too low for reasons I already stated. Rielly and Marner may end up higher than predicted but not by much. We'll see at the end of the season. Only 7 weeks to go.

 

[The Apologist]

I'll do an update at the end of the season. As far as Tavares goes, I know the prediction was much too low for reasons I already stated. Rielly and Marner may end up higher than predicted but not by much. We'll see at the end of the season. Only 7 weeks to go.

Ive taken your correction to Tavares into account and wont bug you about him anymore (although any 'stat' that uses information based on team stats for a team the player no longer plays for is inherently faulty to begin with).

Regarding Mitch and Morgan, what would you consider 'not by much' so that we can come to some sort of agreement on how accurate this stat actually is.
For Mitch, over/ under of say 92? (splitting the difference)
Morgan, using the same five point spread (even though hes a d) would you say O/U of 73 is fair?

 

[42]

Ive taken your correction to Tavares into account and wont bug you about him anymore (although any 'stat' that uses information based on team stats for a team the player no longer plays for is inherently faulty to begin with).

Regarding Mitch and Morgan, what would you consider 'not by much' so that we can come to some sort of agreement on how accurate this stat actually is.
For Mitch, over/ under of say 92? (splitting the difference)
Morgan, using the same five point spread (even though hes a d) would you say O/U of 73 is fair?

I'll take the under on both.

 

[The Apologist]

I'll take the under on both.

Alright. Sounds like a deal. I would be downright shocked if you're right on either.

 

[The Apologist]

Update:
Mitch at 82 with 15 left. Your prediction would have him getting 5 more in the last 15. He needs 10 more to kill your under prediction.

Riles has 64, so 4 away from your prediction with 15 to go. 9 away from my generous over under (you took the under).

As an aside, i think your Tavares prediction has gone from 'off' to ridiculously low.

 

[OiledGun]

Update:
Mitch at 82 with 15 left. Your prediction would have him getting 5 more in the last 15. He needs 10 more to kill your under prediction.

Riles has 64, so 4 away from your prediction with 15 to go. 9 away from my generous over under (you took the under).

As an aside, i think your Tavares prediction has gone from 'off' to ridiculously low.

You’re really invested in proving this guy wrong eh?

 

[The Apologist]

You’re really invested in proving this guy wrong eh?

Not so much just this guy but the whole theory of it. The fact that they were so certain the numbers were right just makes it more fun.
It's like the whole Corsi crowd when you ask them the actual correlation between Corsi and winning....

 

[North Cole]

Update:
Mitch at 82 with 15 left. Your prediction would have him getting 5 more in the last 15. He needs 10 more to kill your under prediction.

Riles has 64, so 4 away from your prediction with 15 to go. 9 away from my generous over under (you took the under).

As an aside, i think your Tavares prediction has gone from 'off' to ridiculously low.

"I won't bug you about Tavares"

Comes back his bookmark 3 weeks later....HOW BOUT TAVARES M8.

 

[authentic]

If we look at the oiSH% of the top 5 point finishers in the last 3 seasons, we can see that the highest it's been is 12% and that 12% is a statistical anomaly with the average in the high 10%

Name, points, oiSH% in last three seasons:

2017/18

1. McDavid, 108, 10.5
2. Giroux, 102, 12
3. Kucherov, 100, 10.3
4. Malkin, 98, 10.3
5. MacKinnon*, 97, 11.9 (played 74 games)

2016/17
1. McDavid, 100, 10.9
2. Crosby, 89, 10.1
2. Kane, 89, 11.3
4. Backstrom, 86, 10.2
5. Kucherov, 85, 10.8
5. Marchand, 85, 9.3

2015/15
1. Kane, 106, 11.2
2. Benn, 89, 10.4
3. Crosby, 85, 9.6
4. Karlsson, 82, 9.6
4. Thornton, 82, 11.1

Even Crosby has never had an oiSH% higher than 11.3% when playing in at least 75 games.

Now scoring is up this year but not that much higher that we could expect the current oiSH% to keep over the entire season. Rielly is not going to finish the year with 16% nor Marner with 15% nor Rantanen with 12.9 % , to name just a few of the highest numbers in the top 20. Hence the expectation that these players will regress the most.

There is also the expectation that the top finishers will not average an oiSH% higher than the average of last 3 seasons by 1%. When the season is done, I expect the top 5 finishers to have an average of maybe something in the low 11%, accounting for the higher scoring this season.

What's Marners scoring placement then compared to now?

 

[The Apologist]

"I won't bug you about Tavares"

Comes back his bookmark 3 weeks later....HOW BOUT TAVARES M8.

Listen, I never said I was sane.

The whole list is quite frankly terribad.