Go Back   Sports Handicapping Forum > Welcome Forums > Main Street > Capping

Capping All handicapping, betting systems, spreadsheets, mathematics & quantitative technicapping.

Reply
 
LinkBack Thread Tools Display Modes
  #1  
Old 12-22-2010, 10:16 PM
Registered User
 
Join Date: Apr 2007
Posts: 10,494
Rewards: 1,017
Your Uses For the Correlation Function in Excel

How have you used the Correlation Function in Excel to help you study what works with your predictions for sports gambling?

I've been experimenting with it, finding the correlation between columns of prediction and result data, and would like to hear of any one else's experiences and uses for it.
Reply With Quote
  #2  
Old 12-22-2010, 11:45 PM
Hall of Fame
 
Join Date: Oct 2005
Location: Salem, VA
Posts: 22,450
Rewards: 1,800
Correlations are a very valuable tool. Some variables relate highly with others and it can help you with your capping. Its a delicate process however, and you need to assess all the data and the content of each variable, and understand the nature of sample sizes. Over time most variables will hover around a central tendency with an occasional outlier, so I would suggest graphing the two variables you are relating in respect to time, and see if the line in the graph stabilizes. If it oscillates wildly, then correlations are meaningless, for its probably random, and you would need a higher sample size.

A lot of things correlate without much regard to logic or reason other than an obvious empty relationship. For instance, anything that increases steadily will correlate highly with the distance of Haley's comet from the earth.

Do you mind sharing what kind of data you are working with?
__________________
"Nobody goes there anymore, its too crowded." --Yogi Berra

"Always tell the truth, that way you won't have to remember what you said." --Mark Twain


*=$50,000
Reply With Quote
  #3  
Old 12-23-2010, 10:03 AM
Registered User
 
Join Date: Apr 2007
Posts: 10,494
Rewards: 1,017
Quote:
Originally Posted by uva3021 View Post
Do you mind sharing what kind of data you are working with?
I've tried different things...

For example finding the correlation factor between various predictive stats like for hockey:

[team total goals/g avg] or
[goalie save % avg's] or
[team penalty min/g avg]

with:
[actual goals/g] from the games on a particular night

By doing this over time, one can see if there is a positive or negative correlation between the factors, and see which ones are more able to accurately predict the final score... for over/under betting purposes...

Another thing i`ve been looking at is how accurate that certain fantasy sites are in predicting the outcomes for NFL player props, like number of QB completions, RB total rush yardage etc.

If one looks at columns of predictions from various sites, then finds the correlation factor between each column and the actual results, one can find out which fantasy site is the best at those predictions.
Reply With Quote
  #4  
Old 12-23-2010, 04:47 PM
Hall of Fame
 
Join Date: Oct 2005
Location: Salem, VA
Posts: 22,450
Rewards: 1,800
Quote:
Originally Posted by Grizzly Joe View Post
If one looks at columns of predictions from various sites, then finds the correlation factor between each column and the actual results, one can find out which fantasy site is the best at those predictions.
great idea
__________________
"Nobody goes there anymore, its too crowded." --Yogi Berra

"Always tell the truth, that way you won't have to remember what you said." --Mark Twain


*=$50,000
Reply With Quote
  #5  
Old 12-23-2010, 10:13 PM
Registered User
 
Join Date: Oct 2010
Posts: 39
Rewards: 190
I will say what uva said in different words..and he is correct.

Spread sheets like Excel are great tools. There are some challenges when using correlation and other statistical tools.

Get a college level statistics textbook at a used book store. How you set up the data, what you compare to what is important. The sample size is important.

But it is important to understand is that CORRELATION DOES NOT MEAN CAUSATION.

Be careful using averages...using averages to averages can yield r's that are too high.

The following site will calculate "p" from "r" and degrees of freedom for free P value calculator

The next logical step up from correlation is regression. Good luck.
Reply With Quote
Reply

Bookmarks

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are Off
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are On



All times are GMT -5. The time now is 06:00 PM.


Powered by vBulletin® Version 3.8.7
Copyright ©2000 - 2012, vBulletin Solutions, Inc.