1. #1
    subs
    subs's Avatar Become A Pro!
    Join Date: 04-30-10
    Posts: 1,412
    Betpoints: 969

    NBA push chart smoothing

    Happy New Hangover

    I've just got a DB for the NBA and spent a while cleaning it up. it goes back to 1995.

    so for my push chart (which i need fairly accurate) i like some simple smoothing techniques.

    can any1 help...

  2. #2
    Peregrine Stoop
    Peregrine Stoop's Avatar Become A Pro!
    Join Date: 10-23-09
    Posts: 869
    Betpoints: 779

    learn some applied statistics

    there are no set rules for this type of thing anywayz

    or ...

  3. #3
    Justin7
    Justin7's Avatar Become A Pro!
    Join Date: 07-31-06
    Posts: 8,577
    Betpoints: 1506

    What sorts of problems are you having? What results do you see without smoothing that you don't accept as accurate?

  4. #4
    subs
    subs's Avatar Become A Pro!
    Join Date: 04-30-10
    Posts: 1,412
    Betpoints: 969

    I am already doing an applied stats course online with stattrek- but i thought that asking the community here for their experiences with this project would be beneficial, and because there is no single answer, may even start a debate.

    as you guys would know well, there are some 0%... i mean near the edges. also there are some very small sample sizes in places. there are also some numbers that look wrong....

    also i want to look for subsets, home/away, totals etc. (that are better teased than ATS) so this will dilute my charts even more and make smoothing "by eye" a very long and messy job.

    so because i do not know where yet, like a do in the NFL (according to my, and others' commonly available numbers) i feel like i need to make the whole chart, edges et all, as tight as i can without wasting too much time. maybe something i only use when the numbers do not look good... how do you even 0.25% kelly without trusting ur numbers?

    any feedback much appreciated.


  5. #5
    Justin7
    Justin7's Avatar Become A Pro!
    Join Date: 07-31-06
    Posts: 8,577
    Betpoints: 1506

    Subs,

    If you are getting 0% push rates for many numbers, I have to wonder how large your sample size is. How many games are you using?

  6. #6
    subs
    subs's Avatar Become A Pro!
    Join Date: 04-30-10
    Posts: 1,412
    Betpoints: 969

    the NBA set goes back to 1995 all games i believe. i mean around the edges that i have 0% on a particular MOV

    i have been advised to merge 3 or so NBA DBs as a good starting point. i imagine the NBA plays i am looking 4 r harder to find than NFL 1's and not as common.

    BTW today is a good day - enjoy!
    Last edited by subs; 01-03-11 at 04:04 PM.

  7. #7
    Justin7
    Justin7's Avatar Become A Pro!
    Join Date: 07-31-06
    Posts: 8,577
    Betpoints: 1506

    When you get to spreads above 15 or so, you might check to see if the push rates are all similar. I would expect they are (or treat them as if they are).

  8. #8
    subs
    subs's Avatar Become A Pro!
    Join Date: 04-30-10
    Posts: 1,412
    Betpoints: 969

    thanks Justin7

  9. #9
    Peregrine Stoop
    Peregrine Stoop's Avatar Become A Pro!
    Join Date: 10-23-09
    Posts: 869
    Betpoints: 779

    Quote Originally Posted by subs View Post
    the NBA set goes back to 1995 all games i believe. i mean around the edges that i have 0% on a particular MOV i have been advised to merge 3 or so NBA DBs as a good starting point. i imagine the NBA plays i am looking 4 r harder to find than NFL 1's and not as common. BTW today is a good day - enjoy!
    0% obviously doesn't pass the sniff test
    Don't forget to think in order to follow set rules

  10. #10
    subs
    subs's Avatar Become A Pro!
    Join Date: 04-30-10
    Posts: 1,412
    Betpoints: 969

    Roasthawg rulz - any1 need a DB, message RoastHawg - really!

    or go to gambling DBS dot com.

    so far can't recommend it enough. BTW i had to ask him, he did not come touting anything....

Top