# NBA - Relative Point Spread Rankings (03/10)

Page 3 of 3  1 2 3
Author: [NBA Betting] Topic: NBA - Relative Point Spread Rankings (03/10)
fresh4lyph PM fresh4lyph
Joined: Sep 2010
Posts: 2668

Veteran
#51
Posted: 3/14/2012 10:52:29 AM
si1ly PM si1ly
Joined: Oct 2008
Posts: 12061

Banned
#52
Posted: 3/14/2012 12:15:22 PM
Woops, thanks termayt, Houston did only have 2.8 points of value. Ill adjust the record accordingly.
Posted using a mobile device.
mmeiselman731 PM mmeiselman731
Joined: Dec 2011
Posts: 128

Prospect
#53
Posted: 3/16/2012 2:50:24 AM
I have a similar system, and I found that using full season numbers is starting to producer worse results. I tweaked my formula to weight more recent performances at a higher rate. Have you considered this?
jamdat21 PM jamdat21
Joined: Mar 2012
Posts: 16

Prospect
#54
Posted: 3/16/2012 3:51:17 AM
QUOTE Originally Posted by si1ly:

These are the updated relative point spread rankings as of March 10th, 2012.  I've included in this thread separate power rankings for home and away.  We're deep enough into the season that teams have played a statistically relevant amount of games both at home and away.  This model works the same as before except the power rankings are derived using only home splits and only away splits respectively.  This means you no longer need to adjust for home court advantage: just take the differential between the home team's PSR and the away team's PSR and you have your true line (less adjustment for intangibles).

These numbers are the result of a complex formula that gives different weights to the metrics that most accurately predict the outcome of a basketball game (the big 4 are eFG%, TOR, ORR, FTR among other things).  While I'm not willing to divulge the specific formula I've come up with I am willing to explain a little about how it works.  This formula effectively converts each of the aforementioned stat categories into a separate value that represents the number of points that a team is better or worse than their competitor (specific to each stat).  I do this by using the differential between the offensive and defensive rates for each of these stat categories.  I then apply a relative weight to each of these differentials depending on how significant an influence that stat has on determining the winner.  This also normalizes the numbers so they can be compared from one team to another.  eFG% is the most important - it has about 45% relative weight in my model.  ORR and TOR each have about 18% weight and FTR about 7%.  There are a few other metrics that I look at, but don't have more than 5% weight.  Add up all the individual stat point spread adjustments and you get your total point spread ranking for a team.

hi can u explain this to me whats those tor eFG im new here so can anyone explain it thanks so much
hubag737 PM hubag737
Joined: Mar 2012
Posts: 26

Prospect
#55
Posted: 3/16/2012 9:23:04 AM
could you explain these stats? i can t read the complete sentences. cant be seen on our screen.

61-31
>2 45-23
>3 31-20
>4 24-16
>5 15-10
5-6

>1 53-28 (65.4%)
>2 40-21 (65.6%)
>3 28-18 (60.9%)
>4 22-15 (59.5%)
>5 14-9 (60.9%)
>6 5-6 (45.5%)
si1ly PM si1ly
Joined: Oct 2008
Posts: 12061

Banned
#56
Posted: 3/16/2012 2:28:13 PM
eFG% = effective field goal percentage
TOR = turnover rate (% of possessions ending in a turnover)
ORR = offensive rebounding rate (% of rebounds grabbed on the offensive glass)
FTR = free throw rate (free throws attempted per field goals attempted)
si1ly PM si1ly
Joined: Oct 2008
Posts: 12061

Banned
#57
Posted: 3/16/2012 2:30:54 PM
Record if you bet all games with 1 or more points of value 63-33

Record if you bet all games with 2 or more points of value 47-24

Record if you bet all games with 3 or more points of value 31-20

Record if you bet all games with 4 or more points of value 25-16

Record if you bet all games with 5 or more points of value 15-10

Record if you bet all games with 6 or more points of value 5-6

si1ly PM si1ly
Joined: Oct 2008
Posts: 12061

Banned
#58
Posted: 3/16/2012 2:34:41 PM
QUOTE Originally Posted by mmeiselman731:

I have a similar system, and I found that using full season numbers is starting to producer worse results. I tweaked my formula to weight more recent performances at a higher rate. Have you considered this?

Yes, I've considered this.  My models seem to perform better with more data.  I understand that weighing recent performance more than the season long averages might have some advantages.  But at the end of the day, I'm just trying to see how two teams match up if they both played average games.  I can adjust my wagers accordingly if one team has a lot more motivation, or is on a hot streak, or whatever...

Gogic PM Gogic
Joined: Feb 2012
Posts: 125

Prospect
#59
Posted: 3/16/2012 3:08:14 PM
QUOTE Originally Posted by si1ly:

Record if you bet all games with 1 or more points of value 63-33

Record if you bet all games with 2 or more points of value 47-24

Record if you bet all games with 3 or more points of value 31-20

Record if you bet all games with 4 or more points of value 25-16

Record if you bet all games with 5 or more points of value 15-10

Record if you bet all games with 6 or more points of value 5-6

Are you using opening or closing lines for determining what is a play and what isn't?? Also, I haven't been keeping records, but I'm pretty sure there were more than 4 games with a 1 or better point advantage over the last 3 days.

Anyway, great job, one of the best threads here, hope you keep updating the weekly numbers.
JRyap_08 PM JRyap_08
Joined: Feb 2012
Posts: 13

Prospect
#60
Posted: 3/16/2012 3:15:04 PM
Hey! how often you change your statistical analysis in your system? Like every week?
si1ly PM si1ly
Joined: Oct 2008
Posts: 12061

Banned
#61
Posted: 3/16/2012 4:59:07 PM
@JRyap: Once every 7-10 days. Takes about that long for a team to play enough home/away games to significantly affect the output.
Posted using a mobile device.
fresh4lyph PM fresh4lyph
Joined: Sep 2010
Posts: 2668

Veteran
#62
Posted: 3/16/2012 5:48:22 PM
what are plays today?
Posted using a mobile device.
JRyap_08 PM JRyap_08
Joined: Feb 2012
Posts: 13

Prospect
#63
Posted: 3/17/2012 12:31:55 AM
I see.. so you mean for Monday's game next week there should be a new statistical analysis? I did your calculation..
I got 4-1.. Nice!
oddsbuster PM oddsbuster
Joined: Oct 2006
Posts: 20924

Banned
#64
Posted: 3/17/2012 2:01:41 AM
Excellent thread!!!!
si1ly PM si1ly
Joined: Oct 2008
Posts: 12061

Banned
#65
Posted: 3/18/2012 2:31:49 PM
Record if you bet all games with 1 or more points of value 77-45

Record if you bet all games with 2 or more points of value 56-35

Record if you bet all games with 3 or more points of value 35-29

Record if you bet all games with 4 or more points of value 29-23

Record if you bet all games with 5 or more points of value 19-15

Record if you bet all games with 6 or more points of value 8-6
si1ly PM si1ly
Joined: Oct 2008
Posts: 12061

Banned
#66
Posted: 3/18/2012 4:21:40 PM
Today's only game with more than 3.5 points of value is Orlando +8.5.  The true spread on this game is Miami -4 thus Orlando has 4.5 points of value.  Games with 4 or more points of value against my model are 29-23 (55.8%).
SweetDickWormy PM SweetDickWormy
Joined: Apr 2011
Posts: 275

Prospect
#67
Posted: 3/20/2012 12:13:35 AM

Si1ly:

Hope all is well buddy.  Just wanted to ask are you planning on updating your rankings this week?  I haven't been capping a game without it!  As always, much appreciated, and best of luck to you!

si1ly PM si1ly
Joined: Oct 2008
Posts: 12061

Banned
#68
Posted: 3/20/2012 1:10:28 PM
QUOTE Originally Posted by SweetDickWormy:

Si1ly:

Hope all is well buddy.  Just wanted to ask are you planning on updating your rankings this week?  I haven't been capping a game without it!  As always, much appreciated, and best of luck to you!

Been pretty busy.. but I'll be working on the updated rankings later tonight.  They'll be ready to use for tomorrow's games going forward.
JBender PM JBender
Joined: Jul 2007
Posts: 314

Prospect
#69
Posted: 3/20/2012 5:19:33 PM
Excellent. Thank you
Posted using a mobile device.
 Page 3 of 3  1 2 3
You have entered the forum as a GUEST.
You must login/register to post or reply.
Desktop View: Switch to Mobile View