My NBA Power Rankings - Ongoing (merged)

blitzology

blitzology

G-League
#1
#1
I have put together a ranking system which is based on a Performance Rating which is based mostly Wins and Losses, but takes into account strength of schedule. There are other details, but too many to list (ask me about something specific if you want to know). These ratings/rankings are not my or anyone's opinion :). They are purely Mathematical. Ratings will be updated daily, but posted here Weekly.
  • At the beginning of the season, each team started out with a 1600 rating.
  • The maximum rating points a team an gain/lose in one game is 50. Minimum is 0.
  • The columns are, {Rank#}. {Team} {(Rating)-(GamesPlayed)}
Ratings current of 11/17/06

1. Utah 1736-9

2. Cleveland 1695-9

3. G State 1687-9

4. S Antonio 1684-9

5. Houston 1670-9

6. Orlando 1665-9

7. LA Clippers 1663-7

8. N Orleans 1645-9

9. N Jersey 1645-8

10.Sacramento 1642-8

11.LA Lakers 1639-9

12.Dallas 1614-8

13.Atlanta 1609-7

14.Detriot 1586-9

15.Denver 1585-7

16.Seattle 1582-10

17.Philly 1581-8

18.New York 1577-10

19.Phoenix 1568-8

20.Boston 1562-9

21.Indiana 1562-9

22.Minnesota 1557-8

23.Miami 1556-8

24.Portland 1544-10

25.Washington 1542-8

26.Milwaukee 1540-9

27.Chicago 1538-9

28.Charlotte 1525-8

29.Toronto 1525-8

30.Memphis 1470-8

Some Notes:

  • First team to reach 1700+: S Antonio achieved 1706-7 after beating Houston on 11/14/06. Utah also achieved 1719-8 on 11/14/06, but it took more games(8) compares to S Antonio(7).
  • First team to fall to 1400+: Charlotte fell to 1486-7 after losing to N Orleans on 11/14/06. Memphis also fell to 1487-7 in the same amount of games, but on the later date of 11/15/06
Let the comments roll....
 
jacobdrj

jacobdrj

Bench
#4
#4
This is interesting, and it seems to be perdy accurate right now, however, I don't see how this system can work over time. It will give a skewed average to teams who do better early vs. those who peak late.

It think more important in basketball is to get the change over time of these rankings/raitings (dr/dt).
 
blitzology

blitzology

G-League
#8
#8
I forgot to add some notes....

  • NBA AVG Rating: 1599.8
  • Atlantic Division AVG Rating: 1578
  • Central Division AVG Rating: 1584
  • Southeast Division AVG Rating: 1579.4
  • EAST Conference AVG RAting: 1580.5
  • Southwest Division AVG Rating: 1616.6
  • Northwest Division AVG Rating: 1600.8
  • Pacific Division AVG Rating: 1639.8(wow) - and PHX is holding it back
  • WEST Conference AVG Rating: 1619.1
 
blitzology

blitzology

G-League
#9
#9
Bad Boy Bo said:
So basically, this is a ladder system.
Click to expand...

No, a true ladder system doesn't take Performance Rating into consideration.

In a ladder, if you win a game vs. someone ahead of you, you move up half way from your rung to their rung. Sometimes the players you beats moves down 1, or just stays where they're at.

So if team on run 10 beats team on rung 4, players on rung 10 would move up to rung 7 (middle of 4-10) and team on rung 4 either moves down 1, or stays where they're at, given the rules.


This rating system doesn't guarantee you can move up or down after 1 game.
 
blitzology

blitzology

G-League
#10
#10
jacobdrj said:
This is interesting, and it seems to be perdy accurate right now, however, I don't see how this system can work over time. It will give a skewed average to teams who do better early vs. those who peak late.

It think more important in basketball is to get the change over time of these rankings/raitings (dr/dt).
Click to expand...

Actually, this rating system is the opposite, it gives a better average over a longer period of time. It takes more games to give the most accurate Performance Rating.
 
VF21

VF21

Super Moderator Emeritus
SME
#12
#12
blitzology said:
I have put together a ranking system which is based on a Performance Rating which is based mostly Wins and Losses, but takes into account strength of schedule. There are other details, but too many to list (ask me about something specific if you want to know). These ratings/rankings are not my or anyone's opinion :). They are purely Mathematical. Ratings will be updated daily, but posted here Weekly.
Click to expand...
How can they be strictly mathematical if you use "strength of schedule" - which is subjective - as part of the equation?

I, too, would like to see the formula you've decided to use.
 
D

Dave McNulla

Bench
#13
#13
VF21 said:
How can they be strictly mathematical if you use "strength of schedule" - which is subjective - as part of the equation?

I, too, would like to see the formula you've decided to use.
Click to expand...
more specifically, the record when you play the opponent or the record when the calculation is done? some opponents start strong or start slow, or temporary injuries occurred. couldn't that skew the numbers?

also, if opponent's record is important, is the opponents' strength of schedule taken into consideration? wouldn't you end up with a circular formula if you did take the opponents' opponents record, or the opponents' opponents' opponents schedule into account?
 
blitzology

blitzology

G-League
#15
#15
VF21 said:
How can they be strictly mathematical if you use "strength of schedule" - which is subjective - as part of the equation?

I, too, would like to see the formula you've decided to use.
Click to expand...

Strength of schedule is not based on opinion, a strength of schedule is based on the team's opponent's Ratings. Not each win is given the same amount of rating points. If two team go on a 3-0 winning streak, it does not mean that they achieve the same amount of rating points. Team A could be low rated, and go on a 3-0 winning streak beating very high rated teams, gaining the maximum 50 rating points each game - while Team B could have a 3-0 winning streak playing lower rated players, maybe only gaining an average of 23 points per game.

"strength of schedule'' is a given when you use a Performance Rating. if you were to give the same amount of points for each win, then the Rankings will be the same as the Win/Loss standings used in the NBA.

I can take San Antonio and Cleveland for example...Both were 7-2 in the latest Rating update, but San Antonio hit a huge blow when their 1700+ rating was ripped by Charlottles 1400+ rating on 11/15. This bumped Charlottle back to 1500+ while it took S.Antonio to around 1660. Cleveland's early losses were not to 1400+ teams, so they are higher rated than S.A even though their W/L is the same.

This is all the strength of schedule is...that is main difference between a Performance rating and a regular W/L performance. Lots of wins is a given in getting a high rating, but it's about who you beat to get you higher than the rest.
 
Last edited:
blitzology

blitzology

G-League
#16
#16
Here is 11/18's games results. I showed each team's rating before the result, then showed both team's new ratings. I also give how many points the winner gained - I didn't bother give how many points the loser lost, because it is the opposite of what the winner gained.

I am still pondering if I should make the formula public on this forum.

MIA 1556 wins over ATL 1609
MIA New Rating: 1585(+29)
ATL New Rating: 1580

WAS 1542 wins over CLE 1695
WAS New Rating: 1577(+34)
CLE New Rating: 1660

BOS 1562 wins over NYK 1577
BOS New Rating: 1588(+26)
NYK New Rating: 1551

NOK 1645 wins over MIN 1557
NOK New Rating: 1664(+29)
MIN New Rating: 1538

ORL 1665 wins over CHA 1525
ORL New Rating: 1680(+15)
CHA New Rating: 1510

DET 1586 wins over HOU 1670
DET New Rating: 1617(+31)
HOU New Rating: 1639

POR 1544 wins over NJN 1645
POR New Rating: 1576(+32)
NJN New Rating: 1613

DAL 1614 wins over MEM 1470
DAL New Rating: 1629(+15)
MEM New Rating: 1455

IND 1562 wins over MIL 1540
IND New Rating: 1585(+23)
MIL New Rating: 1517

UTA 1736 wins over PHX 1568
UTA New Rating: 1750(+14)
PHX New Rating: 1554

LAC 1663 wins over PHI 1581
LAC New Rating: 1682(+19)
PHI New Rating: 1562

DEN 1585 wins over TOR 1525
DEN New Rating: 1606(+21)
TOR New Rating: 1504

GST 1687 wins over SEA 1582
GST New Rating: 1705(+18)
SEA New Rating: 1564
 
uolj

uolj

Starter
#21
#21
Can the ratings be applied to previous years? I'm curious how they compare to final results (both objective and subjective).

I'd be curious to see the formula as well, but if you'd prefer not to share I can deal. :)
 
blitzology

blitzology

G-League
#22
#22
uolj said:
Can the ratings be applied to previous years? I'm curious how they compare to final results (both objective and subjective).
Click to expand...
For each game I have to get the result, plug in the ratings for the winner and loser into the formula and do the math, then update the ratings for the respective teams. This is very easy, but takes sooo much time.

I actually decided to do this ratings list about 2 weeks into the season, so I had to make up the 2 weeks I missed. This was when each team was about 5-9 games into the season and this took about 2 hours just plugging and doing the math, etc, to get caught up. In other words, I would love to experiment on another season, but this could take weeks, if I do it off and on.

Each night takes about 10-30 minutes to update the ratings, depending on how many games were played that night.

If I were to do a whole previous season, I would probably choose the season when the Kings lost in the WCF to the lakers in 7.

uolj said:
I'd be curious to see the formula as well, but if you'd prefer not to share I can deal. :)
Click to expand...
I prefer not to, because this would cause controversy and constant questioning and bickering over why it's done this way.
 
blitzology

blitzology

G-League
#23
#23
Current top 5 and bottom 5 as of games through 11/21/06

1. Utah 1760 10-1

2. S Antonio 1722 9-2

3. LA Lakers 1684 8-3

4. N Orleans 1683 8-3

5. Cleveland 1674 8-3


26. Chicago 1501 3-8

27. Milwaukee 1497 3-8

28. Toronto 1494 2-8

29. Charlotte 1493 2-8

30. Memphis 1480 2-9

Notice, the Kings have faced 4/5 teams in the bottom 5, and have a 3-1 record. 0-1 record against the top 5. Kings current rating is 1620 at 5-4 - there is one other team at 5-4, Denver, which is rated 1625.
 
uolj

uolj

Starter
#24
#24
If you PM me the formula I could write you a program that updates all the data automatically if you just input the winners and losers.

Of course, there's no guarantee I'd finish it before you finish calculating the full season yourself. ;)
 
blitzology

blitzology

G-League
#26
#26
Power Rankings, 11/24/06 !!!

Refer to http://www.kingsfans.com/forums/showthread.php?t=16452 For last week's Power Rankings.

Many teams made HUGE swings this week with lots of upsets and some nice winning streaks. Utah still holds the #1 spot with obvious domination - it would take a nice losing streak from UTA + a winning streak from DAL or SA to lose this spot by next week. I predict their 89 point cushion should hold this spot for another month to come.

Sacramento stays at 10th with just a small drop off of -11 points off their rating. Their losses this week were against top rated teams so their rating didn't take a huge hit, and they were able to pull off a nice win on the road against Seattle to stay above average.

The most improved team this week was Denver as they were able to achieve the most increased rating of +92 points and most increased rank change from 15th to 5th in the Power Rankings.

The most decreased rating was given to New Jersey, losing a whopping -108 points and falling from 9th to 21st. Though this rank drop off wasn't as big as Cleveland, who fell from 2nd(!) to 15th while putting together a group of losses against weak teams.

Power Rankings as of 11/24/06

(Rank) (Team) (Rating) (Rating Change) (W/L Record) (Last Week Rank)


1. UTA 1794 (+58) 12-1 (1)

2. SA 1706 (+22) 10-3 (4)

3. DAL 1696 (+85) 8-4 (12)

4. ORL 1680 (+15) 9-4 (6)

5. DEN 1677(+92) 7-4 (15)

6. HOU 1676 (+6) 8-4 (5)

7. DET 1675 (+89) 8-5 (14)

8. LAL 1665 (+26) 8-4 (11)

9. PHX 1641 (+73) 6-6 (19)

10.SAC 1631 (-11) 6-5 (10)

11.LAC 1626 (-37) 6-4 (7)

12.NOK 1620 (-25) 8-5 (8)

13.GST 1618 (-69) 7-6 (3)

14.IND 1612 (+50) 7-6 (21)

15.CLE 1608 (-87) 8-5 (2)

16.SEA 1596 (+14) 6-8 (16)

17.MIN 1594 (+37) 5-6 (22)

18.POR 1587 (+43) 6-7 (24)

19.ATL 1582 (-27) 5-5 (13)

20.PHI 1537 (-44) 5-7 (17)

21.NJ 1537 (-108) 5-7 (9)

22.NY 1536 (-41) 5-9 (18)

23.MIA 1535 (-21) 4-8 (23)

24.BOS 1527 (-35) 4-8 (20)

25.MIL 1525 (-15) 4-8 (26)

26.WAS 1510 (-32) 4-8 (25)

27.MEM 1509 (+39) 3-9 (30)

28.CHA 1509 (-14) 3-9 (28)

29.TOR 1508 (-17) 3-9 (29)

30.CHI 1477 (-61) 3-9 (27)

Some Notes...

NBA AVG Rating: 1599.6

Atlantic Divison AVG: 1529

Central Division AVG: 1579.4

Southeast Division AVG: 1563.2

EAST Conference AVG: 1557.2

Southwest Division AVG: 1641.4

Northwest Division AVG: 1649.6

Pacific Division AVG: 1636.2

WEST Conference AVG: 1642

 
6th

6th

Homer Fan Since 1985
#28
#28
You know, blitzology, as far as ranking go, yours is quite good. Thanks for your time and effort (even if Mr. Slim uses them for other purposes).
 
You must log in or register to reply here.