Reverse engineer NHL Player Value Formula

• Status: Closed
• Penyertaan Diterima: 2

I have a set of hockey data from the current NHL season for each player. This include stats like assists, blocks, plus minus, etc. In addition I have a computed "PLAYER VALUE" for each player.

The player value is calculated by comparing the performance of a player to the other the players in the league in his position.

I need you to reverse engineer this value and come up with a formula for computing it using the following 6 stat categories:
- assists
- blocks
- faceoff wins
- hits
- penalty minutes
- plus/minus

The position of each player is also provided and is important for the comparison. The working solution should being able to re-create the values from the raw data. I believe the formula works by comparing each stat to all the players in the league who play that position, and then coming up with a value based on how much better or worse the player is.

Note: all 6 categories are considered "good", this means the higher the number, the higher the player value (even penalty minutes).

Here is the sample data to use in your calculations. The far left column, "PLAYER VALUE" is the expected output of the function, and the remaining columns can be used as inputs for determining that value.

Preferred solution format is in one of the following programming languages: PHP, Python, Ruby, Javascript

BONUS: If successful, you will be offered more work to improve upon this formula.

Experience with hockey pools and/or NHL stats is very desirable!

Note: the first person to create a formula matching all data will be awarded the prize.

Papan Penjelasan Umum

• 7 tahun yang lalu

Is this open or resolved? (because it still says "open")

• 7 tahun yang lalu
• 7 tahun yang lalu

If it is open:
1. can you supply more than one list? PPLAYER VALUE is not 0 sum and seeing it calculated on a different set would really help figure out what it's based on.
2. Do you have any guesses wherether people in multiple positions (like "LW/RW") should compete in both fully, in both partially (probably 50% infulence on group score) or each combination is scored in a group of it's own.

• 7 tahun yang lalu
• 7 tahun yang lalu

Dude, I know you've seen this multiple times. It's impossible (and neither of the applicants submitted it) to find out the general formula without
a) a second set of data.
or better yet
b) a source of the data (it's likely to update fter games, so we won't need to bug you)

• 7 tahun yang lalu
• fzvinicius
• 7 tahun yang lalu

Done. Accuracy = 99,9%

• 7 tahun yang lalu
• aniagarwal97
• 7 tahun yang lalu

Sir, Do you need a script or you want just a formula?

• 7 tahun yang lalu
• 7 tahun yang lalu

Hi. That sounds great, please submit as soon as you can!

• 7 tahun yang lalu
• 7 tahun yang lalu

Hi, I am willing to work on your project using Python. Best Regards,

• 7 tahun yang lalu