ADVERTISEMENT

BCMoore Prediction Game: Week 3

bcmoore87

All District
Oct 30, 2001
8,004
29
48
1. Guess the game margin for the top 10 matchups in Class 4A. Beat the computer - earn iapreps fame - prove to everyone that your team is underrated.
<p>
2. Notation: +2 --> Home wins by 2 points. -7 --> Visitor wins by 7 points.
<p>
3. Game is open to all, and all are encouraged to play. The greater the number of participants, the more accurate the overall predictions should be (central limit theorem).
<p>
3.1 Reply to this post in this form:
<p>
username, -2, -4, +3, +5, -3, +10, -20, +3, +3, -1
<p>
where: username --> your rivals.com username,<br>
-2, -4, +3, +5, -3, +10, -20, +3, +3, -1 --> your predictions for the 10 games this week
<p>
3.2 Notes:
<p>
3.2.1 Use one line only.<br>
3.2.2 Use a comma and space as separators
<p>
3.2 The following are example of bad replies:
<p>
name
<p>
-13, +3, +14, +12, +4, +26, +10, -8, +21, +43 (entry uses two lines, comma after name missing)
<p>
name,-13,+3,+14,+12,+4,+26,+10,-8,+21,+43 (spaces missing)
<p>
name: -13,+3,+14,+12,+4,+26,+10,-8,+21,+43 (don't use colon, spaces missing)
<p>
name -13, +3, +14, +12, +4, +26, +10, -8, +21, +43 (commas after name missing)
<p>
3.3 Motivation of format rules is to minimize mistakes in reading your entry and minimize the amount of my time in working on this contest.
<p>
4. Rankings: Official rankings of the contest are the BCMoore Rankings, <a href="http://bcmoorerankings.com">http://bcmoorerankings.com</a>.
<p>
5. Scoring: Entry with most correct answers wins. Ties will be broken by sum of (absolute values of predictions - actual scores).
<p>
6. Deadline: All entries must be submitted by <b>Friday, 5PM</b>. Entries can be changed up to deadline. To change your entry, submit a new entry. Do not edit a previous entry; I may not see the change you want to make.
<p>
7. Duration: Contest will run for the length of the regular season. Updates on season long standings will be given weekly in addition to each individual weeks winner.
<p>
8. Game selection: Games are generally selected based on two criteria:<br>
8.1 Each district has at least one team in the weekly game.<br>
8.2 Remaining games are selected based on state-wide merit.<br>


<p>
This week's games:


Code:
 N Conf Rk (W-L) Visitor              Conf Rk (W-L) Home                Matchup
21 4A-2  2 (2-0) WDM Dowling          4A-4  1 (2-0) WDM Valley           198.36
22 4A-5  3 (2-0) CR Washington        4A-7  8 (1-1) Bettendorf           197.83
23 4A-5  9 (2-0) Cedar Falls          4A-8  7 (2-0) Iowa City West       187.25
24 4A-1 14 (1-1) Johnston             4A-3 15 (0-2) Urbandale            183.50
25 4A-6 13 (1-1) Pleasant Valley      4A-7 10 (1-1) Muscatine            182.04
26 4A-3 16 (0-2) Ankeny               4A-1  6 (1-1) Waukee               177.66
27 4A-8 12 (1-1) CR Prairie           4A-7 18 (2-0) Clinton              171.25
28 4A-5 22 (2-0) Waterloo West        4A-6 19 (2-0) Western Dubuque      166.65
29 4A-8 25 (0-2) Marion Linn-Mar      4A-5 30 (0-2) CR Jefferson         165.47
30 4A-2 23 (1-1) Ames                 4A-7 11 (2-0) Iowa City High       161.69

Composite Computer, 17, -13, -4, -25, 4, 28, 19, 12, -7, 26
Modified Computer, 28, -5, 5, -1, 2, 17, -8, 7, -3, 20
Standard Computer, 16, 3, 2, -2, -3, 12, -7, 2, 2, 15
Momentum Computer, 16, 3, 2, -2, -3, 12, -7, 2, 2, 15
 
Code:
Conf Rk (W-L) Visitor              Conf Rk (W-L) Home                  Open  Curr
4A-2  2 (2-0) WDM Dowling          4A-4  1 (2-0) WDM Valley            16.5  16.0
4A-5  3 (2-0) CR Washington        4A-7  8 (1-1) Bettendorf            -1.0   0.5
4A-5  9 (2-0) Cedar Falls          4A-8  7 (2-0) Iowa City West         2.0   3.5
4A-1 14 (1-1) Johnston             4A-3 15 (0-2) Urbandale             -2.0  -2.0
4A-6 13 (1-1) Pleasant Valley      4A-7 10 (1-1) Muscatine             -0.5   3.0
4A-3 16 (0-2) Ankeny               4A-1  6 (1-1) Waukee                14.5  12.0
4A-8 12 (1-1) CR Prairie           4A-7 18 (2-0) Clinton               -7.0  -3.0
4A-5 22 (2-0) Waterloo West        4A-6 19 (2-0) Western Dubuque        4.5   2.0
4A-8 25 (0-2) Marion Linn-Mar      4A-5 30 (0-2) CR Jefferson          -0.5  -5.0
4A-2 23 (1-1) Ames                 4A-7 11 (2-0) Iowa City High        17.5  15.0
 
Code:
Conf Rk (W-L) Visitor              Conf Rk (W-L) Home                  Open  Curr
4A-2  2 (2-0) WDM Dowling          4A-4  1 (2-0) WDM Valley            16.5  16.0
4A-5  3 (2-0) CR Washington        4A-7  8 (1-1) Bettendorf            -1.0   3.0
4A-5  9 (2-0) Cedar Falls          4A-8  7 (2-0) Iowa City West         2.0   4.0
4A-1 14 (1-1) Johnston             4A-3 15 (0-2) Urbandale             -2.0  -2.0
4A-6 13 (1-1) Pleasant Valley      4A-7 10 (1-1) Muscatine             -0.5   2.0
4A-3 16 (0-2) Ankeny               4A-1  6 (1-1) Waukee                14.5  12.0
4A-8 12 (1-1) CR Prairie           4A-7 18 (2-0) Clinton               -7.0   1.0
4A-5 22 (2-0) Waterloo West        4A-6 19 (2-0) Western Dubuque        4.5   2.0
4A-8 25 (0-2) Marion Linn-Mar      4A-5 30 (0-2) CR Jefferson          -0.5  -6.0
4A-2 23 (1-1) Ames                 4A-7 11 (2-0) Iowa City High        17.5  15.0
 
Code:
Conf Rk (W-L) Visitor              Conf Rk (W-L) Home                  Open  Curr
4A-2  2 (2-0) WDM Dowling          4A-4  1 (2-0) WDM Valley            16.5  15.0
4A-5  3 (2-0) CR Washington        4A-7  8 (1-1) Bettendorf            -1.0   3.0
4A-5  9 (2-0) Cedar Falls          4A-8  7 (2-0) Iowa City West         2.0   4.5
4A-1 14 (1-1) Johnston             4A-3 15 (0-2) Urbandale             -2.0  -4.5
4A-6 13 (1-1) Pleasant Valley      4A-7 10 (1-1) Muscatine             -0.5   1.5
4A-3 16 (0-2) Ankeny               4A-1  6 (1-1) Waukee                14.5  11.5
4A-8 12 (1-1) CR Prairie           4A-7 18 (2-0) Clinton               -7.0  -1.0
4A-5 22 (2-0) Waterloo West        4A-6 19 (2-0) Western Dubuque        4.5   2.0
4A-8 25 (0-2) Marion Linn-Mar      4A-5 30 (0-2) CR Jefferson          -0.5  -6.0
4A-2 23 (1-1) Ames                 4A-7 11 (2-0) Iowa City High        17.5  15.0
 
Code:
Conf Rk (W-L) Visitor              Conf Rk (W-L) Home                  Open  Curr
4A-2  2 (2-0) WDM Dowling          4A-4  1 (2-0) WDM Valley            16.5  14.0
4A-5  3 (2-0) CR Washington        4A-7  8 (1-1) Bettendorf            -1.0   3.0
4A-5  9 (2-0) Cedar Falls          4A-8  7 (2-0) Iowa City West         2.0   4.5
4A-1 14 (1-1) Johnston             4A-3 15 (0-2) Urbandale             -2.0  -5.0
4A-6 13 (1-1) Pleasant Valley      4A-7 10 (1-1) Muscatine             -0.5  -1.0
4A-3 16 (0-2) Ankeny               4A-1  6 (1-1) Waukee                14.5  12.0
4A-8 12 (1-1) CR Prairie           4A-7 18 (2-0) Clinton               -7.0  -5.0
4A-5 22 (2-0) Waterloo West        4A-6 19 (2-0) Western Dubuque        4.5   2.5
4A-8 25 (0-2) Marion Linn-Mar      4A-5 30 (0-2) CR Jefferson          -0.5  -4.5
4A-2 23 (1-1) Ames                 4A-7 11 (2-0) Iowa City High        17.5  15.0
 
Code:
Conf Rk (W-L) Visitor              Conf Rk (W-L) Home                  Open  Curr
4A-2  2 (2-0) WDM Dowling          4A-4  1 (2-0) WDM Valley            16.5  14.0
4A-5  3 (2-0) CR Washington        4A-7  8 (1-1) Bettendorf            -1.0   3.0
4A-5  9 (2-0) Cedar Falls          4A-8  7 (2-0) Iowa City West         2.0   5.0
4A-1 14 (1-1) Johnston             4A-3 15 (0-2) Urbandale             -2.0  -6.5
4A-6 13 (1-1) Pleasant Valley      4A-7 10 (1-1) Muscatine             -0.5   1.5
4A-3 16 (0-2) Ankeny               4A-1  6 (1-1) Waukee                14.5  12.0
4A-8 12 (1-1) CR Prairie           4A-7 18 (2-0) Clinton               -7.0  -5.0
4A-5 22 (2-0) Waterloo West        4A-6 19 (2-0) Western Dubuque        4.5   2.0
4A-8 25 (0-2) Marion Linn-Mar      4A-5 30 (0-2) CR Jefferson          -0.5  -4.5
4A-2 23 (1-1) Ames                 4A-7 11 (2-0) Iowa City High        17.5  16.0
 
Sorry I missed this week as the overall leader. But, the FCS stuff I'm doing for college sports madness takes most of my free time.
 
no problem

I planned on doing some college rankings this weekend, hadn't done anything yet.

Hope your venture goes well.
 
Code:
Conf Rk (W-L) Visitor              Conf Rk (W-L) Home                  Open  Curr Act
4A-2  2 (2-0) WDM Dowling          4A-4  1 (2-0) WDM Valley            16.5  14.0   7
4A-5  3 (2-0) CR Washington        4A-7  8 (1-1) Bettendorf            -1.0   3.0  -1
4A-5  9 (2-0) Cedar Falls          4A-8  7 (2-0) Iowa City West         2.0   5.0   7
4A-1 14 (1-1) Johnston             4A-3 15 (0-2) Urbandale             -2.0  -6.5   4
4A-6 13 (1-1) Pleasant Valley      4A-7 10 (1-1) Muscatine             -0.5   1.5  14
4A-3 16 (0-2) Ankeny               4A-1  6 (1-1) Waukee                14.5  12.0   3
4A-8 12 (1-1) CR Prairie           4A-7 18 (2-0) Clinton               -7.0  -5.0  -4
4A-5 22 (2-0) Waterloo West        4A-6 19 (2-0) Western Dubuque        4.5   2.0 -24
4A-8 25 (0-2) Marion Linn-Mar      4A-5 30 (0-2) CR Jefferson          -0.5  -4.5  -7
4A-2 23 (1-1) Ames                 4A-7 11 (2-0) Iowa City High        17.5  16.0  -3
 
Code:
Modified Computer wins

Name                 Cor Att   %    Dev
Modified Computer      7  10 0.700  120
MarionHawk             6  10 0.600   68
GoBigD6788             6  10 0.600   99
iowaqb                 6  10 0.600  105
Dragon7678             6  10 0.600  108
makelovenotwar         5  10 0.500   94
Straw1397              5  10 0.500   98
bcmoore87              5  10 0.500   98
Composite Computer     5  10 0.500  185
Standard Computer      4  10 0.400  106
Momentum Computer      4  10 0.400  106
Iafootballfan          4  10 0.400  129


MarionHawk leads

Name                 Cor Att   %    Dev
MarionHawk            21  30 0.700  392
Dragon7678            21  30 0.700  437
GoBigD6788            21  30 0.700  471
Straw1397             20  30 0.667  441
bcmoore87             19  30 0.633  459
Iafootballfan         19  30 0.633  494
Modified Computer     19  30 0.633  510
Composite Computer    18  30 0.600  632
screwloose            17  20 0.850  359
Standard Computer     15  30 0.500  504
Momentum Computer     15  30 0.500  504
makelovenotwar        11  20 0.550  507
iowaqb                10  20 0.500  507
zmac328                9  10 0.900  441
DSMan                  8  10 0.800  473
Simpson Storm          6  10 0.600  487
makelovenorwar         4  10 0.400  514


Marionhawk leads

Name                 ATS Att   %    p-val Fame
MarionHawk            20  26 0.769  0.002   14
iowaqb                14  19 0.737  0.017    9
Dragon7678            18  28 0.643  0.066    8
screwloose            12  17 0.706  0.045    7
Straw1397             14  23 0.609  0.154    5
zmac328                5   7 0.714  0.145    3
bcmoore87             13  25 0.520  0.423    1
DSMan                  5   9 0.556  0.380    1
makelovenotwar         9  18 0.500  0.500    0
GoBigD6788            11  23 0.478  0.420   -1
makelovenorwar         3   7 0.429  0.368   -1
Composite Computer    13  28 0.464  0.356   -2
Simpson Storm          2   7 0.286  0.145   -3
Iafootballfan          9  23 0.391  0.154   -5
Modified Computer      9  24 0.375  0.114   -6
Standard Computer      5  20 0.250  0.010  -10
Momentum Computer      5  20 0.250  0.010  -10
 
I'm still tracking FCS for my ratings, but they won't make their first appearance on that website until after Week 4. The Top 25 and Bracketology, I have to do every week.
 
screw and/or BC, would one (or both) of you post a link or recommend a site which outlines how these models work? I know there are "tricks to the trade" and "proprietary" things that you use to make your models uniquely your own. I am just curious to see how this is done. Is there a reasonable starting point? Yes, I am a reasonably good mathematician/statistician.

Thanks.....
 
Moore's system is a fancy system that repeatedly compares teams and their standing with everybody else until the numbers don't move any more. Look up "Simple Rating System" for a generic explanation, but it will show give you the general idea.

Mine is much more primitive. Although, I'm constantly making tweaks to see where I can improve it. But I'm not hiding anything, so I'll tell you how mine works...

First, I create a RAW Points total for each team, which is 125 team in FCS. Points are awarded or taken for wins, losses, and points scored and allowed.

RAW = Wins(x5) - Losses(x3) + Points Scored - Points Allowed(x 0.5)

Last year, I also gave a per game bonus for each of the schools. This was weighted based on the conference a team was in. Teams from FCS Power Conferences(Big Sky / Big South / Colonial / Missouri Valley / Ohio Valley / Patriot / Southern / Southland), those with automatic bids and scholarships, were given 21 points per game. Teams from leagues that get automatic bids but don't give scholarships(or greatly reduced scholarships - Northeast / Pioneer), got a 14 points per game bonus. The league without automatic bids(Ivy / Mid-Eastern / Southwestern), got 7 point bonuses. This was done as a way of shuffling the stronger conferences towards the top, and to ensure that teams didn't finish with a negative rating.

This year, my main tweak is coming in this area. The per game rewards will start at 30-20-10 for those same levels. I'm moving the Ivy League from the last group to the middle group, as they receive an automatic playoff bid, but always decline the invitation. The big difference is each conference will receive a different reward, depending on how each conference performs against their fellow FCS counterparts. For example, in games ONLY against other FCS opponents, if the Missouri Valley finishes with a winning percentage of .750, then each team will get a bonus of 22.5(30 x .750). If the Big Sky finishes with a .500 record, they will get 15(30 x .500).

Each team also has a three-part, RPI-style Strength of Schedule multiplier. The first part is their own winning percentage.

Win Percentage = Wins / (Wins + Losses)

Next, is a team's opponent's winning percentage.

Opponent's Win Percentage = Opponent's Wins / (Opponent's Wins + Opponent's Losses)

Thirdly, is a team's adjust opponent's winning percentage. This takes away games that a tem played against their opponents. There's also some level bonuses and penalties for opponents that are played. I won't get into how that's done exactly, because it's hard to explain, but games against FBS teams(Iowa or Iowa State, for example) help this adjustment. Playing against teams from Division II, III or NAIA hurt this adjustment. The main reason for this is to aid or cripple your schedule strength according to how bad some teams beat or lose to these kinds of opponents.

Adjusted Opponent's Win Percentage = (Opponent's Win - Your Losses) / (Opponent's Win + Opponent's Losses - Your Wins - Your losses) +/- adjustments

The Strength of Schedule takes these three percentages and creates an average.

Strength of Schedule = (Win Percentage + Opponent's Win Percentage+ [Adjusted Opponent's Win Percentage x 2]) / 4

After all of this is done. You create the team's final Power Rating by multiplying the RAW Points by SoS.

Power Rating = RAW Points x Strength of Schedule

Hope that helps you out.
 
> Moore's system is a fancy system that repeatedly compares teams and their standing with everybody else until the numbers don't move any more. Look up "Simple Rating System" for a generic explanation, but it will show give you the general idea.

Nice compliment, but mine is simply a different way to look at data.

The following is part of a presentation I did to a class at Iowa State in Spring 2011. My rankings are very similar to this, with an added adjustment to attempt to remove the effect of non-competitive games:




=================================================
Multiple regression model, with home field advantage:
=================================================




Teams:

1 Colorado
2 Iowa State
3 Kansas
4 Kansas State
5 Missouri
6 Nebraska

Number of Games:

octave:15> nchoosek(6,2)
ans = 15


Scores: (15)

R 10/7
Nebraska 48 Kansas State 13

S 10/9
Colorado 0 Missouri 26

R 10/14
Kansas State 59 Kansas 7

S 10/18
Iowa State 20 Kansas State 27

S 10/30
Kansas 16 Iowa State 28
Missouri 17 Nebraska 31

S 11/06
Colorado 45 Kansas 52
Nebraska 31 Iowa State 30

S 11/13

Iowa State 14 Colorado 34
Kansas 3 Nebraska 20
Kansas State 28 Missouri 38

S 11/20
Kansas State 36 Colorado 44
Missouri 14 Iowa State 0

F 11/26
Colorado 17 Nebraska 45

S 11/27
Kansas 7 Missouri 35

Question: In looking at these game, which games produce an emotion reaction? Which games do you remember? Are there any games that seem unusual (score is not indicate of the strength of the teams)?



Code:
Standings:

Nebraska     (5-0)
Missouri     (4-1)

Kansas State (2-3)
Colorado     (2-3)

Iowa State   (1-4)
Kansas       (1-4)




Variables: (7)

Code:
1 Colorado
2 Iowa State
3 Kansas
4 Kansas State
5 Missouri
6 Nebraska
7 Home-Field Advantage

Equations: (15)

Code:
-1*Nebraska + 1*Kansas State + Home-Field Advantage = -35
-1*Colorado + 1*Missouri + Home-Field Advantage = 26
-1*Kansas State +1*Kansas + Home-Field Advantage = -52

-1*Iowa State +1*Kansas State + Home-Field Advantage = 7
-1*Kansas +1*Iowa State + Home-Field Advantage = 12
-1*Missouri +1*Nebraska + Home-Field Advantage = 14

-1*Colorado +1*Kansas + Home-Field Advantage = 7
-1*Nebraska +1*Iowa State + Home-Field Advantage = -1
-1*Iowa State +1*Colorado + Home-Field Advantage = 20

-1*Kansas +1*Nebraska + Home-Field Advantage = 17
-1*Kansas State +1*Missouri + Home-Field Advantage = 10
-1*Kansas State +1*Colorado + Home-Field Advantage = 8

-1*Missouri +1*Iowa State + Home-Field Advantage = -14
-1*Colorado +1*Nebraska + Home-Field Advantage = 28
-1*Kansas +1*Missouri + Home-Field Advantage = 28

Code:
X( 1,:) = [ 0  0  0  1  0 -1  1]
X( 2,:) = [-1  0  0  0  1  0  1]
X( 3,:) = [ 0  0  1 -1  0  0  1]
X( 4,:) = [ 0 -1  0  1  0  0  1]
X( 5,:) = [ 0  1 -1  0  0  0  1]
X( 6,:) = [ 0  0  0  0 -1  1  1]
X( 7,:) = [-1  0  1  0  0  0  1]
X( 8,:) = [ 0  1  0  0  0 -1  1]
X( 9,:) = [ 1 -1  0  0  0  0  1]
X(10,:) = [ 0  0 -1  0  0  1  1]
X(11,:) = [ 0  0  0 -1  1  0  1]
X(12,:) = [ 1  0  0 -1  0  0  1]
X(13,:) = [ 0  1  0  0 -1  0  1]
X(14,:) = [-1  0  0  0  0  1  1]
X(15,:) = [ 0  0 -1  0  1  0  1]

X =

Code:
   0   0   0   1   0  -1   1
  -1   0   0   0   1   0   1
   0   0   1  -1   0   0   1
   0  -1   0   1   0   0   1
   0   1  -1   0   0   0   1
   0   0   0   0  -1   1   1
  -1   0   1   0   0   0   1
   0   1   0   0   0  -1   1
   1  -1   0   0   0   0   1
   0   0  -1   0   0   1   1
   0   0   0  -1   1   0   1
   1   0   0  -1   0   0   1
   0   1   0   0  -1   0   1
  -1   0   0   0   0   1   1
   0   0  -1   0   1   0   1

Y = [-35; 26; -52; 7; 12; 14; 7; -1; 20; 17; 10; 8; -14; 28; 28]

Y =

Code:
  -35
   26
  -52
    7
   12
   14
    7
   -1
   20
   17
   10
    8
  -14
   28
   28

Equation:

XB = Y

Solve for B:

B = X\Y

Code:
   -5.1190
   -5.3810
  -16.6190
    1.3810
   10.2857
   15.4524
    2.2857

B(1:6) team estimates

B(7) = Home-field advantage

2.2857


Subtract mimimum score to make all numbers non-negative

NewB = B(1:6) - min(B(1:6))

Code:
   11.50000
   11.23810
    0.00000
   18.00000
   26.90476
   32.07143


How does that compare to the standings?

Standings:

Code:
1 Nebraska     (5-0)
2 Missouri     (4-1)

3 Kansas State (2-3)
3 Colorado     (2-3)

5 Iowa State   (1-4)
5 Kansas       (1-4)

----------------------

Code:
1 Nebraska     (5-0)  32.07
2 Missouri     (4-1)  26.90
3 Kansas State (2-3)  18.00
4 Colorado     (2-3)  11.50
5 Iowa State   (1-4)  11.24
6 Kansas       (1-4)   0.00
----------------------

Upsets:

Colorado 45 Kansas 52
Kansas State 36 Colorado 44

----------------------

How does the model predict Iowa State's results?

Home-field advantage = 2.29

Code:
Result                           Actual   Predict
Iowa State 20 Kansas State 27        +7     +9.05
Kansas 16 Iowa State 28             +12    +13.53
Nebraska 31 Iowa State 30            -1    -18.54
Iowa State 14 Colorado 34           +20     +2.55
Missouri 14 Iowa State 0            -14    -13.37

----------------------

Iowa State:

Best game of year:

Nebraska, 17.54 points better than expected


Worst game of year:

Colorado, 17.45 points worse than expected
 
ADVERTISEMENT

Latest posts

ADVERTISEMENT