- Thread starter
- #1
Darkstone42
Oh.
- 35,085
- 2,053
- 173
- Joined
- Apr 19, 2010
- Location
- Tucson, AZ
- Hoopla Cash
- $ 1,000.00
- Fav. Team #1
- Fav. Team #2
- Fav. Team #3
So after building up significant hype for this little project of mine, I have finally completed it. That's right, the Resident Nerd has once again lived up to his name, but this time has done so in a fashion that is [hopefully] beneficial to the board.
To begin, I'll rehash how I came across my little equation. I entered the points per game (hand calculated, no less =P) of every player with a sufficient sample size over the past year and that player's salary into an Excel document. I then plotted all of the points in a scatter plot. Then I fit a linear trend line (it's much easier to analyze and makes more sense to me from a business standpoint anyway) and took the equation from there.
It is important to note that the scatter in this chart was huge. The R^2 value was 0.6553, so the fit wasn't very precise. But I believe that the enormous sample size (319 data points) accounts for this shortcoming. That is to say, I believe there are enough data points that the discrepancies nearly evenly balance out, and that there are about as many overpaid players as underpaid players. Also, since this accounts for 319 NHL forwards, we should also assume that the salary value returned for a certain player's ppg assumes that player is average in every other aspect of the game. So, for example, a player with x ppg who is an above average defender should make more than a player with x ppg who is an average defender, all other things equal.
I also decided to force the trend line through zero. Now, I understand that the league minimum is about 550K, but players making that sum are probably not getting paid for their offense. I figure a 0 ppg player adds exactly zero value to his team, and that player must possess some other ability (face-offs, defense, agitation, physicality, fighting, etc.) in order to earn consistent playing time. Besides that, a GM knows he must pay a player 550K, so if a player's ppg puts him under that, and the GM still wants him, he knows what he's going to be paying. Additionally, as I mentioned in the "Coming Soon" thread, forcing the line through the league minimum produced about a 20% decreased in the line's slope. That was way too much.
So without further adieu, here is the equation:
y = 5.3293x
where x is the ppg and y is the player's salary in millions of dollars.
Now you can have fun with it, evaluating your teams' signings of forwards this off-season. Remember that this applies only to forwards.
The next step of this work, I think it's important to note, is to work backward and add specificity. That is to say, given time and opportunity, look at the percentage of points a player scores as goals and find out how that affects salary. Then work from there to determine a numerical value that could be assigned to forwards for their offensive proficiency versus league average. But that will likely take months, so don't get excited for that just yet. =P
Thank you all for your kind attention. =)
To begin, I'll rehash how I came across my little equation. I entered the points per game (hand calculated, no less =P) of every player with a sufficient sample size over the past year and that player's salary into an Excel document. I then plotted all of the points in a scatter plot. Then I fit a linear trend line (it's much easier to analyze and makes more sense to me from a business standpoint anyway) and took the equation from there.
It is important to note that the scatter in this chart was huge. The R^2 value was 0.6553, so the fit wasn't very precise. But I believe that the enormous sample size (319 data points) accounts for this shortcoming. That is to say, I believe there are enough data points that the discrepancies nearly evenly balance out, and that there are about as many overpaid players as underpaid players. Also, since this accounts for 319 NHL forwards, we should also assume that the salary value returned for a certain player's ppg assumes that player is average in every other aspect of the game. So, for example, a player with x ppg who is an above average defender should make more than a player with x ppg who is an average defender, all other things equal.
I also decided to force the trend line through zero. Now, I understand that the league minimum is about 550K, but players making that sum are probably not getting paid for their offense. I figure a 0 ppg player adds exactly zero value to his team, and that player must possess some other ability (face-offs, defense, agitation, physicality, fighting, etc.) in order to earn consistent playing time. Besides that, a GM knows he must pay a player 550K, so if a player's ppg puts him under that, and the GM still wants him, he knows what he's going to be paying. Additionally, as I mentioned in the "Coming Soon" thread, forcing the line through the league minimum produced about a 20% decreased in the line's slope. That was way too much.
So without further adieu, here is the equation:
y = 5.3293x
where x is the ppg and y is the player's salary in millions of dollars.
Now you can have fun with it, evaluating your teams' signings of forwards this off-season. Remember that this applies only to forwards.
The next step of this work, I think it's important to note, is to work backward and add specificity. That is to say, given time and opportunity, look at the percentage of points a player scores as goals and find out how that affects salary. Then work from there to determine a numerical value that could be assigned to forwards for their offensive proficiency versus league average. But that will likely take months, so don't get excited for that just yet. =P
Thank you all for your kind attention. =)
