Evaluation of 2011 Dobber Projections

[OceanMon]

I did a analysis of Dobber's numbers from last year.

Approximately 10 players he had projected numbers for did not produce any numbers at all.

There were also a number of players that produce substantial numbers like Sykora, Henrique, and Mcdonough that were not accounted for in the projections.

The following statistics only apply to the players that Dobber had projections for and did indeed play at least one game.

# of projections = 491

Within 5 Games Played = 42%
Within 10 Games Played = 61%

Within 10 Penalty Minutes = 45%
Within 20 Penalty Minutes = 74%

Within 5 Points = 21%
Within 10 Points = 41%
Within 15 Points = 59%
Within 20 Points = 74%

[ericdaoust]

Interesting. The number projections are tough in some cases because some guys just do not play the full season in the NHL and when they do play it's just a handful of minutes per game.

[OceanMon]

That's a good point.

Of the players that had GP projections within 5 games, the point projection accuracies within 5 pts and 10 pts were 30% and 56%, respectively.

Similar numbers were obtained for players with GP projections within 10 games.

[robmyatt]

That's a good point.

Of the players that had GP projections within 5 games, the point projection accuracies within 5 pts and 10 pts were 30% and 56%, respectively.

Similar numbers were obtained for players with GP projections within 10 games.

That's pretty damn good relative to other guides from what I've read.

[Ross The Boss Palmer]

60% of the time, it works everytime.

[93213]

Considering how unpredictable injuries are these % are pretty amazing.
Not that I really needed to be sold on buying the guide again this year, but this is very reassuring.

[OceanMon]

I would be interested to see how he did on more a points/gp type analysis...

Within 0.05 points per game = 19%

Within 0.10 points per game = 37%

Within 0.20 points per game = 69%

[Ritchierich]

I assume alot of the players towards the end don't amount to much or ever get drafted. Do you have the numbers on just the top 150 for points? That would be interesting.

[OceanMon]

I assume alot of the players towards the end don't amount to much or ever get drafted. Do you have the numbers on just the top 150 for points? That would be interesting.

For the predicted top 150 point getter projections:

Within 5 pts = 22%
Within 10 pts = 38%
Within 15 pts = 53%
Within 20 pts = 67%

[OceanMon]

If someone has projection data from another source, it would be interesting to compare their success rate against Dobber's.

[Dobber]

Here are the full numbers, year over year:

# Pts Off Within 5 Within 10 Within 20 Exact
656 11/12 avg 12.4 33.1% 57.8% 89.0% 3.2%
722 10/11 avg 11.6 31.6% 56.4% 84.1% 2.8%
638 09/10 avg 11.1 35.7% 58.8% 83.4% 3.0%
708 08/09 avg 11.9 29.7% 54.5% 83.2% 3.1%
738 07/08 avg 12.6 32.0% 54.2% 80.1% 5.6%
659 06/07 avg 13.1 27.2% 48.1% 78.8% 2.7%
609 05/06 avg 15.1 30.3% 52.5% 86.1% 5.2%
556 03/04 avg 13.2 27.0% 50.5% 82.6% 2.2%
542 01/02 avg 12.2 29.9% 52.9% 83.9% 3.1%
541 02/03 avg 11.6 31.8% 55.8% 85.4% 2.8%
541 00/01 avg 11.6 26.5% 54.1% 82.6% 2.0%

[Dobber]

But seriously - buy the guide for the content and the thought-process.

And I agree with eyemissgilmour - I should start tracking points per game track record instead.

[OceanMon]

Even if I found better projections, I'd buy it just to give a little back for the great ramblings all year.

[BleedCopperAndBlue]

All this talk about the guide has me PSYCHED for tomorrow! Already purchased it along with my prospects guide which was amazing.

[hockeyrulez69]

60% of the time, it works everytime.

hahahaha classic! I tried giving you rep for that, but it says I need to spread it around first