clear	
	clear matrix
	set matsize 800
	set mem 500m
	cd [your drive partition]
	use lanham_pitching.dta






	collapse (sum)  HR BB SO  IBB WP HBP  R  IPouts , by(playerID yearID teamID lgID)


	
	gen IP=IPouts/3



	merge m:m playerID using  lahman_names
	drop _merge




	


merge m:m bbrefID yearID using bbr 2024 pitcher patch,update 
drop _merge
drop if yearID==.











*** determine performance values as rates


gen hr_pip=HR/IP
gen k_pip=SO/IP
gen bb_pip=BB/IP
gen hbp_pip=HBP/IP
gen rapg=R/IP*9

*** create fipr scores for pitchers

gen fipr=.

foreach y of numlist 1900/2024 {
		di "year `y'"
		regress rapg k_pip bb_pip hr_pip hbp_pip [iweight=IP] if year==`y'
		predict yhat 
	replace fipr = yhat if year==`y'
	drop yhat 
	 }
	 
	 
	 
collapse fipr  k_pip bb_pip hr_pip hbp_pip rapg (sum) IP , by(bbrefID playerID yearID)

*** merge conventional FIP

merge 1:1 bbrefID yearID using fg fip data, update 
drop _merge


*** season-by-season regression

gen R2_fip=.
gen R2_fipr=.


foreach y of numlist 1900/2024 {
    di "Processing year `y'" 

   

    quietly regress rapg FIP  [iweight=IP] if year == `y'
    replace R2_fip = e(r2) if year == `y'
	
	quietly regress rapg fipr  [iweight=IP]   if year == `y'
	
	replace R2_fipr = e(r2) if year == `y'
	
}



	
	gen R2_fipr_scaled=R2_fipr*100
	gen R2_fip_scaled=R2_fip*100 

	
	
	
			


	twoway ///  
	   (lpoly R2_fip_scaled year  , bw(5) fcolor(blue) lpattern(dash)) ///
       (lpoly  R2_fipr_scaled year  ,  bw(5)lcolor(red) lpattern(dash)) ///
      , ///
       ylabel(20(10)70, nogrid) xlabel(1912 1920(10)2024, angle(45) nogrid) ///
       title("") xtitle("") ///
       legend(off)
	   
	   
	   
	   
	   
*** store fipr weights *** 
	   
	   
	   
	  * Create a temporary file to store the coefficients
tempfile coef_results
postfile handle int year double(b_k_pip b_bb_pip b_hr_pip b_hbp_pip) using `coef_results'

* Loop over each year from 1900 to 2024
foreach y of numlist 1900/2024 {
    di "Processing year `y'..."
    quietly {
        regress rapg k_pip bb_pip hr_pip hbp_pip [iweight=IP] if year == `y'
        * Store coefficients for each variable
        local b_k    = _b[k_pip]
        local b_bb   = _b[bb_pip]
        local b_hr   = _b[hr_pip]
        local b_hbp  = _b[hbp_pip]
    }
    post handle (`y') (`b_k') (`b_bb') (`b_hr') (`b_hbp')
}

postclose handle

use `coef_results', clear
list


*** plot them

* Graph for b_k_pip
twoway (lpoly b_k_pip year, lpattern(dash) bw(2) lcolor(black)), ///
    xlabel(1900(10)2020, nogrid format(%02.0f) angle(45)) ///
    ylabel(, nogrid) ///
    ytitle("") xtitle("Season") ///
    graphregion(color(white)) ///
    plotregion(style(none)) ///
    xscale(r(1912 2020)) ///
    legend(off)
graph save graph1.gph, replace

* Graph for b_bb_pip
twoway (lpoly b_bb_pip year, lpattern(dash) bw(2) lcolor(red)), ///
    xlabel(1900(10)2020, nogrid format(%02.0f) angle(45)) ///
    ylabel(, nogrid) ///
    ytitle("") xtitle("Season") ///
    graphregion(color(white)) ///
    plotregion(style(none)) ///
    xscale(r(1912 2020)) ///
    legend(off)
graph save graph2.gph, replace

* Graph for b_hbp_pip
twoway (lpoly b_hbp_pip year, lpattern(dash) bw(2) lcolor(green)), ///
    xlabel(1900(10)2020, nogrid format(%02.0f) angle(45)) ///
    ylabel(, nogrid) ///
    ytitle("") xtitle("Season") ///
    graphregion(color(white)) ///
    plotregion(style(none)) ///
    xscale(r(1912 2020)) ///
    legend(off)
graph save graph3.gph, replace

* Graph for b_hr_pip
twoway (lpoly b_hr_pip year, lpattern(dash) bw(2) lcolor(blue)), ///
    xlabel(1900(10)2020, nogrid format(%02.0f) angle(45)) ///
    ylabel(, nogrid) ///
    ytitle("") xtitle("Season") ///
    graphregion(color(white)) ///
    plotregion(style(none)) ///
    xscale(r(1912 2020)) ///
    legend(off)
graph save graph4.gph, replace

graph combine graph1.gph graph2.gph graph3.gph graph4.gph, cols(2)


*** save weights **

export excel using "[your drive]\fipr_weights.xls", firstrow(variables) replace