[nikossf]

Id like to hear some new ideas here.....
Please....
Only people who really sit down and do the math firgure all the stats for each game and come up with their own lines. Not based on what vegas has but what their stats and number crunching really comes up with.



Id like to hear some ideas and opinions on what you do to come up with your own personalized lines?!!
Thanks everyone.

[nikossf]

Hopes this helps some people understand a little more on what im talking about.
Thanks everyone..


For each team, you’d take the total points scored and subtract the total points given up, then divide that result by the number of games played. That will give you the average points differential, which can be plus or minus.

Example, take the Buffalo at Miami game.

Bills calculation: (149-118)/6 == +5.2
Dophins calculation: (120-130)/6 == -1.7

So, the Bills would be favored by +5.2 –(1.7) == 6.9
If you round that off, you’d have a 7 point favorite.

You can adjust that by giving the home team 3 points as many odds makers do. Since the Dolphins are the home team, the point spread would be reduced to Bills by 4.

There are a couple of huge fallacies with this. The teams probably haven’t played equal oppositions and the fewer games played, the less meaningful the calculations.

[nikossf]

Here's a simple way.. Just thought I would share.


I am going to use the Washington Redskins and Dallas Cowboys as my example in which the books have listed the Redskins as 3 point favorite at Dallas. We are going to take the last 6 scores from both teams 2005 NFL campaign.

Washington
10-20
17-10
31-20
35-20
35-7
17-13



Because we are looking for the median we are going to throw out the high and the low scores for both offense and defense. This will give us the numbers 17, 31, 35, and 17. We threw out one of the 35s and the low score of 10. These 4 games total add up to 100 and give us a median score of 25 on offense. We will do the same for the defense and come up with 20, 10, 20, and 13. The defense power rating is 15.75, (63 divided by 4).



We will than do the same thing with the Cowboys, our hypothetical matchup.


Dallas
10-20
24-20
7-35
31-28
10-17
21-24
20-7



You will see I listed 7 final scores for Dallas and we are going to throw out the last performance of 10-20 against Seattle since that was the Cowboys last game of the season and it was not a meaningful game. Games like these will throw off the power rating, just as you don’t use preseason games in your ratings since it is not an accurate measure of the teams full capabilities.
With that said we come up with a median score for the Dallas offense of 24, 10, 21, and 20 which gives us an offensive rating of 18.75. On the defensive side of the ball we have scores of 20, 28, 17, and 24 giving us the defensive power rating of 22.25.
Now that we have our team ratings lets formulate a point spread. Washington’s scoring rating is 25 and Dallas’s defensive rating is 22.25. We add these 2 numbers together and get 47.25 than subtract 20.5 (which is the average team rating that will be used in every power rating for the season) and get 26.75 for predicted Washington points. We would do the same for Dallas, 18.75 + 15.75=34.5. Than subtract the 20.5 and we would come up with 14.
The predicted final score in this game using our power rating system would be Washington 26.75, Dallas 20.5. Because this hypothetical game is being played at Dallas we must add 3 points to the Cowboys score (adding 1.5 to their score and taking 1.5 away from Washington, since the Cowboys are playing at home. Giving us the new total of Washington 25.25, Dallas 22. With our hypothetical point spread of Washington -3 it looks like it is dead on for the most part. This game would be a pass in our book at the current number of Washington -3. Now if the books had made Washington a 2 point favorite, a Washington -2 bet would be a very good investment, just as a line of Dallas +4 would have made a very good investment on Dallas at that number. I recommend wagering on any number that is 1 point off of your power rankings. By staying disciplined and keeping to this 1 point model you will find about 3 solid bets per week against the nfl point spreads.
Just keep in mind to come up with a new median score after each game is played out. Once the first week of the season is over, you would than take that result and eliminate the first of the 6 games you previously used to derive a median score.

[jgilmartin]

What you are asking about is modeling. Sorry to say, but it's unlikely anyone is going to answer with anything valuable. Countless hours go into making and testing models; people would be shooting themselves in the foot by posting them on a public forum. Justin's book, "Conquering Risk," is available with your betpoints in the store - there are some models in there. I haven't tried any of them, but at the very least you would learn some of the techniques some people use, and work on coming up with something on your own. Also, IMO you might want to start out with a market other than NFL. Not easy.

[Pokerjoe]

What you are asking about is modeling. Sorry to say, but it's unlikely anyone is going to answer with anything valuable. Countless hours go into making and testing models; people would be shooting themselves in the foot by posting them on a public forum. Justin's book, "Conquering Risk," is available with your betpoints in the store - there are some models in there. I haven't tried any of them, but at the very least you would learn some of the techniques some people use, and work on coming up with something on your own. Also, IMO you might want to start out with a market other than NFL. Not easy.

Exactly.

[CHUBNUT]

I love the way you've copied and pasted someone else's view from another site but not realised their mistake

[sharpcat]

J.R. Miller wants his formula back.