casmith07
Attorney-at-Zone
- Messages
- 31,538
- Reaction score
- 9,312
So the mock draft discussion got me thinking, particularly with stuff like FanSpeak -- what kind of logic do they use? What kind of modeling algorithms are they using in order to make the selections? Weights, etc?
All of this being said, I decided to pull from some of my old mathematics instruction in college as a systems engineering major, and have begun constructing a Monte Carlo simulation in Microsoft Excel using a normal distribution for the data.
This is a work in progress. I will use this thread to post updates and eventually will post the ultimate results of the "mock draft."
Ideally, I will set up the sheet to run the simulation approximately 10,000 times and save the data from each run in a separate sheet. This will allow for a graphical distribution of each player to see where they are likely to be selected in the upcoming draft. Using this mathematical data (based on scout, web scout, and other rankings I pull from data sources online) you can then use the "Monte Carlo Big Board" to plug in players around their rankings or picks with different teams in a mock draft based on team need or team philosophy.
I hope to have something ready to go by April. That way it'll be a useful tool that can be played with for the month leading up to the real deal on May 8th.
If anyone has big board or scout ranking data, particularly in Excel format, please let me know. It would be very useful for increasing the number of data sets, thereby adding more variance in the model and achieving a more realistic normal distribution rather than only using the four or five data sets that I'm trying to input into the model itself now (Top 400 prospects...so a lot of work).
Sorry for the wall of text. Hopefully this makes sense to someone, or seems interesting to someone.
All of this being said, I decided to pull from some of my old mathematics instruction in college as a systems engineering major, and have begun constructing a Monte Carlo simulation in Microsoft Excel using a normal distribution for the data.
This is a work in progress. I will use this thread to post updates and eventually will post the ultimate results of the "mock draft."
Ideally, I will set up the sheet to run the simulation approximately 10,000 times and save the data from each run in a separate sheet. This will allow for a graphical distribution of each player to see where they are likely to be selected in the upcoming draft. Using this mathematical data (based on scout, web scout, and other rankings I pull from data sources online) you can then use the "Monte Carlo Big Board" to plug in players around their rankings or picks with different teams in a mock draft based on team need or team philosophy.
I hope to have something ready to go by April. That way it'll be a useful tool that can be played with for the month leading up to the real deal on May 8th.
If anyone has big board or scout ranking data, particularly in Excel format, please let me know. It would be very useful for increasing the number of data sets, thereby adding more variance in the model and achieving a more realistic normal distribution rather than only using the four or five data sets that I'm trying to input into the model itself now (Top 400 prospects...so a lot of work).
Sorry for the wall of text. Hopefully this makes sense to someone, or seems interesting to someone.
