What type of regression/predictive modelling should I do? ( more like should i learn to approach) I have a score based outcome column that is based on other variables. For example:




MatchStatus1
MatchStatus2
Score




50
50
35




Above is just a simplified version of the datasets where the score was obtained from this formula ( just an example) : Score = (MatchStatus1 * 0.5) + (MatchStatus2*0.2).
Would I be able to run a linear regression on this using MatchStatus1 and MatchStatus2 to predict the Score? cause since it is already a formula based I figured that is a model by it self?  I am probably confused with regression all together.
 A: Welcome to Cross Validated. Hope you learn a lot in this forum! :-)
With respect to your question, it doesn't make much sense to build a regression model if you have a formula that relates both parameters.
In other words, your equation is alredy a model! Let's demonstrate that:
MatchStatus1 <- round(runif(n = 50, min = 1, max = 20),0)
MatchStatus2 <- round(runif(n = 50, min = 1, max = 20),0)
Score <- MatchStatus1 * 0.5 + MatchStatus2 * 0.2 

If you fit a linear model containing the variables MatchStatus1 and MatchStatus2 the result is exactly the equation you used to determine the variable Score:
model <- lm(Score ~ MatchStatus1 + MatchStatus2)
summary(model)

This is the result of the summary():
Coefficients:
                    Estimate Std. Error   t value Pr(>|t|)    
(Intercept)         0.000e+00  2.469e-16 0.000e+00        1    
MatchStatus1        5.000e-01  1.598e-17 3.128e+16   <2e-16 ***
MatchStatus2        2.000e-01  1.476e-17 1.355e+16   <2e-16 ***
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 5.669e-16 on 47 degrees of freedom
Multiple R-squared:      1, Adjusted R-squared:      1 
F-statistic: 5.64e+32 on 2 and 47 DF,  p-value: < 2.2e-16

So, the adjusted equation for the obtained model is $Score = 0.5 \cdot MatchStatus1 + 0.2 \cdot MatchStatus2$
Regression models are used to determine the relationship between variables when a priori that relation is not known.
