Questions linear regression

Just messing around with some linear regression but I'm no data scientist and can't figure it all out with Google. My questions:

1. Do you always check box 'Constant is zero'? Viewing video's I think the answer is "yes" but when I do check the R-squared is 0.042 and when I don't it's 0.721.

2. If I add a second stat the R-squared rises to 0.735 but the P-value from the first stat also goes from 8.33E-06 to 0.090? What does that tell me?

(@mods: If this is the wrong subforum please adjust)

I have to laugh.  This thread has so much advanced vocabulary that it has effectively shut down the forum for a month.  This forum is more likely to answer questions like : Should I chase home dogs on Monday Night Football? than why is the P-value jumping when I add I add second stat?  By the way the R-squared seems in line.  I actually took a course at the University of San Diego called econometrics that was all about linear regression.  R-squared, P-value, and constant = 0 all came up regularly.  Unfortunately, I can't recall much from the class, partly because I prefer to work outdoors and not at a computer.  Anyhow Charon, good luck finding an answer to your question.  Linear regression likely can be used to predict final scores, rushing yards, and a lot of other sports stats.  It's just that on Covers, the syatems tend to be a lot more simple.

I teach regression; not certain about what you found on line, I’ll break it down for you simply: 

1. basically your taking the formula of a straight line  y= a + bx and trying to use it to predict things. 

2. in order to have a line you need 2 points… thus in order to have a line you have 2 variables. 

3. the variables break down to x and a y .   Remember x drives y .   Or another way of looking at it y/x = dependent/independent . 

4.  There are two test that you need to do in order to use this equation to ensure it’s worthy of using .    We call that r squared which measures the strength of the relationship and most stat books say it needs to be 80% or .8 or higher . 

5. the 2nd test is called probability not zero or big T over little t ….. they measure the same thing which is is there simply a relationship between x and y.    You need both tests because it acts like a check and balance as one could be good while the other is not. 

6.  There are tools out there that you can use but more importantly if you can think in terms of cause and effect  aka x drives why it will help you think about sports betting from a different perspective .    I think from this perspective and it helps me tremendously .