Colton's 2014 Mock Draft Compilation, v. 1.0
- Thread starter colton
- Start date
A bit above average, I'd say. But not as far above average as I had hoped.
But that's only the data for a single year. Really I'd need to do the analysis every year for a while to get better statistics.
D
Deleted member 365
Guest
I no how two spell.
I think you need a vacation. You seem like your stressed out. Go out with you're family and half a good time. I wood hit up San Diego. Theirs a beach they're and there very friendly.
Stifle Tower
Punch Bowl Re-Filler
I thought he meant "my complaints."
Irregardless, well done!
<colton baiting...the new favorite sport of Jazzfanz>
jjscap
Well-Known Member
Great work @Colton. I like the fact that you used only the most recent mock drafts.
I made a compilation myself too. It includes Chad Ford's 8.1 too but lacks some of your mock drafts. I am dying to find out what you think of this site: nbadraft.dahoops.com
I made a compilation myself too. It includes Chad Ford's 8.1 too but lacks some of your mock drafts. I am dying to find out what you think of this site: nbadraft.dahoops.com
Cool! I haven't seen another compilation that comprehensive. Did you write some scripts to pull the numbers off of the mock draft websites? Or how does yours work? I'm ashamed to admit that I just entered numbers manually into a spreadsheet.
Also, a problem I see with yours is that where a player isn't ranked, that doesn't affect the average. That artificially makes the player ranked more highly than he should be. A non-ranking should be a negative thing, not a neutral thing. That's one of the best parts of the instant run-off voting that I use... it can handle that situation very naturally. I set all of the non-ranked players in a given poll to be a "61", so they are all tied and all lower than the lowest ranked player in the poll. (Since it just compares relative ranks, there's no difference between voting for three players by indicating 1, 2, 3, and by indicating 1, 2, 61.)
jjscap
Well-Known Member
Hey Colton,
Please don't mind my spelling and grammar mistakes, English is not my first language
Well yes, I wrote a script with Pyton's Beautiful Soup to scrape the data from the mock draft pages. While this is the long-term solution that makes things automatic, I had just a few days to put the whole thing together. And these websites are usually coded terribly, are a bit tricky to scrape. Also I had to add some extra mock drafts suddenly, the clock was ticking. So for now, it is half script, half manual. But for the 2015 draft it'll be full automatic collecting the data from about 25 mock drafts and running some math on its own.
Unranked players gave me some headache. The problem is some analysts did not publish their second round picks. Then I first thought about calculating the averages assuming an unranked player is picked #35 if he is not in the top 30, which would be pretty close to his actual ranking if they published the 2nd round. Then I can't just show 35 in the sortable column of the respective mock draft. This is ugly. Besides ignoring is not that bad if you think about it. I mean my player made the top-30 and usually has something like 17,21,24,20,22... then unranked. so 31 as you say, or 35. This particular mock draft is misbehaving a bit, don't you think? There are statistical models that eliminate black sheep in such cases. I did just that. Well in the end I simply decided to revisit this matter another time.
In your original post you mentioned IRV which seems to take care of this problem elegantly. There is even a Python script doing just that, might help solve unranked players problem.
Please don't mind my spelling and grammar mistakes, English is not my first language
Well yes, I wrote a script with Pyton's Beautiful Soup to scrape the data from the mock draft pages. While this is the long-term solution that makes things automatic, I had just a few days to put the whole thing together. And these websites are usually coded terribly, are a bit tricky to scrape. Also I had to add some extra mock drafts suddenly, the clock was ticking. So for now, it is half script, half manual. But for the 2015 draft it'll be full automatic collecting the data from about 25 mock drafts and running some math on its own.
Unranked players gave me some headache. The problem is some analysts did not publish their second round picks. Then I first thought about calculating the averages assuming an unranked player is picked #35 if he is not in the top 30, which would be pretty close to his actual ranking if they published the 2nd round. Then I can't just show 35 in the sortable column of the respective mock draft. This is ugly. Besides ignoring is not that bad if you think about it. I mean my player made the top-30 and usually has something like 17,21,24,20,22... then unranked. so 31 as you say, or 35. This particular mock draft is misbehaving a bit, don't you think? There are statistical models that eliminate black sheep in such cases. I did just that. Well in the end I simply decided to revisit this matter another time.
In your original post you mentioned IRV which seems to take care of this problem elegantly. There is even a Python script doing just that, might help solve unranked players problem.