New Formula for predicting point spreads

[arkfanchip]

Ok I'm not a heavy better but I have been betting small amounts on games for about 5 years now.  For the past 5 years I've just went with my gut feeling and I'm probably close to 50/50.  So the past few weeks I've been coming up with a formula to try to predict the point spreads and I've came up with one that I think will work.  So far this bowl season (if Wake wins by more than 2) I'm 4-0.  Picking Texas, Mich St., O. State, and WF.  The formula is very confusing but I'll post it to see if you think it's a good one or not. 

Team A & Team B

I take the last 4 non-overtime games against BCS teams and total up the total rushing and passing yards for the offense and defense.

[arkfanchip]

Just did the Mizzou/Ark game and my numbers predict Mizzou by 10 points. 

[hogs25]

What about MSU and Pats game tonight?  If it works you want have to work again. 

[arkfanchip]

UCF is not a team in a BCS conf. this is way more accurate when comparing two teams with similar strengths of schedule.

[arkfanchip]

I did the Alamo Bowl today and it predicts Penn State to win by 5 points and since the point spread is 6 that's to close for me to bet on. 

[MuskogeeHogFan]

Just did the Mizzou/Ark game and my numbers predict Mizzou by 10 points. 

My calculations have them by 4, 35-31.

[arkfanchip]

Ok my predictions for tonight's game and tom. game

Penn State   24.5
Texas A&M   19

Alabama       32
Colorado       16

So I'm going to bet on Alabama since they are a 4 point favorite.  I'm staying away from tonight's game. 

Clemson        24
Auburn          10

So I'm going to bet on Clemson being - 2.5

[arkfanchip]

Penn State 24
Texas A&M 17

My prediciton is close...I hope it ends this way!

[arkfanchip]

Ok my picks according to my formula

Alabama -4         32
Colo.                  16


Clemson -2.5       24
Auburn                10


Ok State -6          49
Indiana                31


Wisconsin +2        30
Tenn                    14

Mizzou -3.5          41
Arkansas              31

[dffhogs]

How does the LSU/Ohio st game come out?

[MuskogeeHogFan]

How does the LSU/Ohio st game come out?

I've got it 31-24, LSU over Ohio State.

[arkfanchip]

I got it 27-22 LSU over OSU.  To close for me to bet on. 

[arkfanchip]

Blowout of the bowl season

Florida -11.5     41
Mich.                10


I didn't realize mich was so bad against the run....

[arkfanchip]

Texas Tech -6     42
Virginia              24

I just went back and redid my calculations for this game and 42-24 isn't right

It is
Texas Tech-6    37
Virginia            20

[upperdeck_hawg]

the final scores are derived how? adding the total offensive and defensive yards and then what?

[MuskogeeHogFan]

I got it 27-22 LSU over OSU.  To close for me to bet on. 

Are you taking home field advantage into consideration? For LSU, this is the next thing to playing in Baton Rouge. That alone, is worth 3 points. It doesn't apply in many bowl games, but it does in this one.

[Feralhog]

Are you taking home field advantage into consideration? For LSU, this is the next thing to playing in Baton Rouge. That alone, is worth 3 points. It doesn't apply in many bowl games, but it does in this one.

I have a feeling LSU will lose this game and one day soon, tiger fans will be pissed the mad hatter didn't leave for meeeeshigan.

[MuskogeeHogFan]

Texas Tech -6     42
Virginia              24

I've got it TT 30  Va 24   Might seem crazy, but Virginia and the points might be a good bet.

[MuskogeeHogFan]

Alabama 27  Colorado  21  Take the under.

[Immensehog]

Putting it to the test tonight.

Gotta get back square.


Rooolllll   Tide!!!

[MuskogeeHogFan]

Putting it to the test tonight.

Gotta get back square.


Rooolllll   Tide!!!

Hey, don't go by what I say. This my pick.....but there ain't no guarantees.

[Hoofer]

Ok I'm not a heavy better but I have been betting small amounts on games for about 5 years now.  For the past 5 years I've just went with my gut feeling and I'm probably close to 50/50.  So the past few weeks I've been coming up with a formula to try to predict the point spreads and I've came up with one that I think will work.  So far this bowl season (if Wake wins by more than 2) I'm 4-0.  Picking Texas, Mich St., O. State, and WF.  The formula is very confusing but I'll post it to see if you think it's a good one or not. 

Team A & Team B

I take the last 4 non-overtime games against BCS teams and total up the total rushing and passing yards for the offense and defense.

that's cool that you figured that out but would you mind showing the numbers for one of the games you worked up so we can see how the numbers work? Not sure I understand what you do after adding up the total yards. Great insight though and it looks like it's working.

[arkfanchip]

My formula is hard to explain in detail.  I have tried in many messages and people are just more confused.  I do take home field into consideration and I did for the LSU game b/c my formula had them winning 24-22.  The numbers after the formula is done is usually between 200-300 so I have derived a point total for each number in that range.  So far it's worked but I'm not going to put big money on these bets until I do the entire bowl season and NFL playoffs.  I'm 5-0 with it so far so take it for what it's worth.  I wouldn't trust me just yet...wait til next year.  I promise I'll post them on here next year for everyone if they do work.  Somebody just start finding me a football coaching job in NWA...I'm stuck in NEA right now! 

[arkfanchip]

All the games I've posted I'm betting on and the rest are to close for comfort for me to bet on.

[PhillyHog]

To really see if this works, you should try backtesting it for a couple of years.  I mean, go back to the 2001-2006 seasons and use the methodology to predict each of the bowl games for the respective year.  Then see if your record is significantly different from 50%.

Successfully picking 4 games in one year is by no means an accurate indicator of a successful formula.  Imagine having 1000 people randomly picking a winner of any given game.  Assuming that each person has a 50% chance of getting the winner right (and thus a 50% chance of getting it wrong), there will be (.5)(.5)(.5)(.5)(1000)=6.125 people that have a 4-0 record on chance alone.

[arkfanchip]

13-0 Bama....ROLL TIDE!  Make me 6-0 this year

I know that this formula is not full-proof but it seems to give matchup differences between teams.  You can never predict turnovers, motivation, Stupid (HDN) Coaching moves but you can always match up a great running team against a weak defense. 

[arkfanchip]

It takes time to research each game so I might do that over the off-season if I do I'll let you know how it comes out.  Thanks for the idea though

[arkfanchip]

20-0!!!!!   Roll Tide Baby!!!!

[Erockster20]

Can you finish your thoughts on how you arrive at your numbers. At least PM me how you get your numbers if you don't mind.  Thanks.

[Hoofer]

same here, PM me the details if you have them detailed somewhere, i'm very familiar with complicated models for point spreads, would love to take a closer look cus your obviously on to something.

i understand totaling the offense and defensive number but not sure how you convert to a score.

[MuskogeeHogFan]

My formula is hard to explain in detail.  I have tried in many messages and people are just more confused.  I do take home field into consideration and I did for the LSU game b/c my formula had them winning 24-22.  The numbers after the formula is done is usually between 200-300 so I have derived a point total for each number in that range.  So far it's worked but I'm not going to put big money on these bets until I do the entire bowl season and NFL playoffs.  I'm 5-0 with it so far so take it for what it's worth.  I wouldn't trust me just yet...wait til next year.  I promise I'll post them on here next year for everyone if they do work.  Somebody just start finding me a football coaching job in NWA...I'm stuck in NEA right now! 

I understand what you are doing and I have something similar that I use. The first season I used it I was about 72% for the entire season. The next season, about 63%. It is darned near impossible to predict the odd bounces that oblong ball can take. Take this season, with all the odd and unpredicatable losses that occured, who could have predicted those? No formula, I can tell you that and I am a firm believer in results being somewhat predictable using stats.

[arkfanchip]

I've never used a formula before so I have never seen one.  This is my entire formula:

Team A and Team B

I take the previous 4 BCS non-overtime games that each team has previously played.  So this only works when BCS plays BCS

Here is the rushing formula: (You have to do this for both teams obviously)

1. Total Rushing Yards+Total rushing yards allowed by other team
2. Take that number and multply it by 1.5 (Only do this on the rushing b/c running is more important than passing
3. Take that number and mulitply it by the percentage of the time the offensive team ran the ball compared to passing
4. Take that number and multiply it by the percentage of the time opposing teams ran the ball against the defensive team

The passing formula is the same except take out step 2

You will get a Rushing Number and a Passing Number for each team
Add the two numbers together and divide by 4 (or however many games you research)

You will get a number between 175-300

My range is anything under 200 = 10 points
200-210 = 11-16
210-225=  17-23
225-240=  24-29
240-255=  30-34
255-270=  35-39
270-280=  40-44

[arkfanchip]

Getting a little nervous....Freaking Alabama....we couldn't beat this team???

[Erockster20]

On the Ark/Mo game, did you use the Oklahoma game (big 12 champ game) or did you go back one extra and use just reg season games?

[Erockster20]

Also, is the number you get, what the team will score? or give up?

[arkfanchip]

What the team will score.  And I didn't use the Champ game. 

[Hoofer]

anybody have their favorite sites for researching all the information?

[arkfanchip]

I use CBSSPORTSLINE.COM.   

[arkfanchip]

Alabama is TERRIBLE!!!!!  They are going to give me an ulcer.....they had this game in the bag

[MuskogeeHogFan]

anybody have their favorite sites for researching all the information?

ncaasports.com
stassen.com
college web sites

[MuskogeeHogFan]

Alabama is TERRIBLE!!!!!  They are going to give me an ulcer.....they had this game in the bag

Chill, all they have to do is run out the clock, but I sure would not have used a time out.

[arkfanchip]

They pulled it out!  Barely!  I didn't realize how bad Alabama was.

[razorbacker231]

sooo, by using this formula, i would say you are not going to bet the 1st 4 games of the season? 

also, how does a cupcake non conferance game affect your numbers? or do yo just keep with the system?

example, ar plays a non conf team like utah state or whatever, that is terrible against the run, and runs up 470 yards on the ground in game 3.  game 4 is luisianna monroe, and about the same or still in 400s anyway.  then game 5 is against, oh say auburn or georgia, florida .  you know that they will prob hold the run and we may not hit 300s.

that is purely based on any bcs school in weaker non conferance games.  not just hogs, or whatever.

[Hoofer]

One Question.

4. Take that number and multiply it by the percentage of the time opposing teams ran the ball against the defensive team

Say your looking at TT vs VA.

If your analyzing TT's offense, what does multiplying the total of #3 by the % VA runs the ball have to do with TT?

Just wondering the thought/methodology behind that. Maybe I am not interpreting that right. BTW, I think this is a great way to analyze games, just want to make sure I'm looking at the information correctly. Look forward to seeing how it plays out rest of bowl season.

[arkfanchip]

sooo, by using this formula, i would say you are not going to bet the 1st 4 games of the season? 

I won't use it until each team has played at least 3 quality teams.  No cupcakes it would make the numbers way to high.  I try to make sure the quality of the 4 teams I calculate for is about even.  For example I think it was Texas A&M who played Mizzou, Kansas, Ok, and Texas their last  4 games so I replaced one of them with KSTATE I think. 

also, how does a cupcake non conferance game affect your numbers? or do yo just keep with the system?

example, ar plays a non conf team like utah state or whatever, that is terrible against the run, and runs up 470 yards on the ground in game 3.  game 4 is luisianna monroe, and about the same or still in 400s anyway.  then game 5 is against, oh say auburn or georgia, florida .  you know that they will prob hold the run and we may not hit 300s.

that is purely based on any bcs school in weaker non conferance games.  not just hogs, or whatever.

[arkfanchip]

One Question.

4. Take that number and multiply it by the percentage of the time opposing teams ran the ball against the defensive team

Say your looking at TT vs VA.

If your analyzing TT's offense, what does multiplying the total of #3 by the % VA runs the ball have to do with TT?

Just wondering the thought/methodology behind that. Maybe I am not interpreting that right. BTW, I think this is a great way to analyze games, just want to make sure I'm looking at the information correctly. Look forward to seeing how it plays out rest of bowl season.

I think you're misunderstanding it.  For the Texas Tech/VIR. Game here is the formula

Texas Tech Rushing vs. Virginia Rushing Defense

175 (Yrds rush by tt) + 450 (Yrds allowed by VA) = 625
625 * 1.5 (The number I multiply rushing yrds by)= 937
937 * .18% (The percent of the time TT runs the ball)= 169
169 * .58% (The Percent of the time Opponents run against VA)=98

[arkfanchip]

sooo, by using this formula, i would say you are not going to bet the 1st 4 games of the season? 

I won't use it until each team has played at least 3 quality teams.  No cupcakes it would make the numbers way to high.  I try to make sure the quality of the 4 teams I calculate for is about even.  For example I think it was Texas A&M who played Mizzou, Kansas, Ok, and Texas their last  4 games so I replaced one of them with KSTATE I think. 

also, how does a cupcake non conferance game affect your numbers? or do yo just keep with the system?

example, ar plays a non conf team like utah state or whatever, that is terrible against the run, and runs up 470 yards on the ground in game 3.  game 4 is luisianna monroe, and about the same or still in 400s anyway.  then game 5 is against, oh say auburn or georgia, florida .  you know that they will prob hold the run and we may not hit 300s.

that is purely based on any bcs school in weaker non conferance games.  not just hogs, or whatever.



I wouldn't use this formula unless each team has played at least 3 quality teams. 

[razorbacker231]

thx arkfanchip. 

and +1 for takin time answer some questions on this.  looks great on paper.  i will check it out, maybe just do it on paper for the playoffs or something.  good luck with the rest of your games.

[arkfanchip]

No Problem If you have any more questions or want me to figure out a game for you let me know. 

[razorsox]

Wouldn't it be a hell of a lot easier just to take the the time machine to get next years Gray's Sports Almanac?