College Football – BettorIQ – Analysis of Qual Signal

Published

June 10, 2024

Import dependencies

pathrs = "G:/Runshop/"
setwd(pathrs)

pacman::p_load(rmarkdown, data.table, tidyverse, magrittr, janitor, lubridate, scales, matrixStats, gt, patchwork)
source("G:/Colossus/codes/R/Runshop/Runshop_functions.R")
`%notin%` = Negate(`%in%`)

Import historical BettorIQ data

Rows: 38,792
Columns: 233
$ week                   <dbl> 9, 1, 3, 1, 1, 1, 2, 1, 1, 13, 3, 5, 12, 1, 3, …
$ season                 <dbl> 2013, 2014, 2014, 2015, 2016, 2017, 2017, 2018,…
$ week_qualify           <chr> "Yes", "Yes", "Yes", "Yes", "Yes", "Yes", "Yes"…
$ date                   <dttm> 2013-10-26, 2014-08-27, 2014-09-13, 2015-09-05…
$ time                   <dttm> 1899-12-31 20:00:00, 1899-12-31 19:00:00, 1899…
$ team                   <chr> "ABILENE CHR", "ABILENE CHR", "ABILENE CHR", "A…
$ opponent               <chr> "NEW MEXICO ST", "GEORGIA ST", "TROY", "FRESNO …
$ away                   <chr> "ABILENE CHR", "ABILENE CHR", "ABILENE CHR", "A…
$ home                   <chr> "NEW MEXICO ST", "GEORGIA ST", "TROY", "FRESNO …
$ team_concat            <chr> "ABILENE CHR41573", "ABILENE CHR41878", "ABILEN…
$ opp_concat             <chr> "NEW MEXICO ST41573", "GEORGIA ST41878", "TROY4…
$ team_fbs               <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ opp_fbs                <chr> "YES", "YES", "YES", "YES", "YES", "YES", "YES"…
$ fbs                    <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ team_qual_rating       <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ opp_qual_rating        <dbl> NA, NA, NA, NA, NA, NA, NA, NA, 0.2, -1.9, NA, …
$ qual_edge              <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ qual_edge_cat          <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ side_qualify_basic     <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ side_qualify_combined  <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ team_total_rating      <dbl> NA, NA, NA, NA, NA, NA, NA, NA, 0, 0, NA, NA, N…
$ opp_total_rating       <dbl> NA, NA, NA, NA, NA, NA, NA, NA, 1.5, 3.0, NA, N…
$ combined_total_edge    <dbl> NA, NA, NA, NA, NA, NA, NA, NA, 1.5, 3.0, NA, N…
$ basic_total_qualify    <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ combined_total_qualify <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ original_h_a           <chr> "Away", "Away", "Away", "Away", "Away", "Away",…
$ h_a                    <chr> "Away", "Away", "Away", "Away", "Away", "Away",…
$ q1awayscore            <dbl> 3, 3, 7, 7, 0, 7, 0, 7, 0, 0, 7, 0, 0, 0, 3, 0,…
$ q2awayscore            <dbl> 19, 13, 0, 0, 7, 0, 0, 13, 10, 7, 0, 3, 7, 3, 0…
$ q3awayscore            <dbl> 0, 14, 21, 0, 7, 0, 10, 7, 7, 0, 3, 7, 0, 6, 7,…
$ q4awayscore            <dbl> 7, 7, 10, 6, 7, 7, 0, 0, 14, 0, 3, 13, 8, 0, 7,…
$ otawaysore             <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
$ q1homescore            <dbl> 7, 7, 7, 14, 7, 7, 10, 17, 17, 7, 0, 7, 14, 7, …
$ q2homescore            <dbl> 14, 14, 14, 7, 21, 7, 7, 21, 21, 14, 10, 16, 14…
$ q3homescore            <dbl> 0, 0, 7, 10, 0, 10, 14, 10, 10, 7, 0, 15, 14, 1…
$ q4homescore            <dbl> 13, 17, 7, 3, 9, 14, 7, 7, 3, 17, 7, 17, 13, 7,…
$ othomescore            <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
$ awascore               <dbl> 29, 37, 38, 13, 21, 14, 10, 27, 31, 7, 13, 23, …
$ homscore               <dbl> 34, 38, 35, 34, 37, 38, 38, 55, 51, 45, 17, 55,…
$ ot                     <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA…
$ neutral                <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
$ openspr                <dbl> 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.…
$ openfav                <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ opentot                <dbl> 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.…
$ openhml                <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ spread                 <dbl> 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.…
$ favored                <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ team_score             <dbl> 29, 37, 38, 13, 21, 14, 10, 27, 31, 7, 13, 23, …
$ opp_score              <dbl> 34, 38, 35, 34, 37, 38, 38, 55, 51, 45, 17, 55,…
$ margin                 <dbl> -5, -1, 3, -21, -16, -24, -28, -28, -20, -38, -…
$ spread_op              <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ fave_dog_op            <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ cover_margin_op        <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ ats_w_op               <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ ats_l_op               <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ ats_p_op               <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ spread_cl              <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ spread_cl_qual         <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ fave_dog_cl            <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ cover_margin_cl        <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ ats_w_cl               <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ ats_l_cl               <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ ats_p_cl               <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ total                  <dbl> 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.…
$ total_points           <dbl> 63, 75, 73, 47, 58, 52, 48, 82, 82, 52, 30, 78,…
$ totals_week_qualify    <chr> NA, "Yes", "Yes", "Yes", "Yes", "Yes", "Yes", "…
$ total_cover_margin_op  <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ over_op                <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ under_op               <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ push_op                <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ total_cover_margin_cl  <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ over_cl                <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ under_cl               <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ push_cl                <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ homeml                 <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ team_ml_op             <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ team_ml_cl             <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ opp_ml_op              <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ opp_ml_cl              <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ openh1spr              <dbl> 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.…
$ openh1fav              <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ openh1sprodds          <dbl> 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00,…
$ openh1tot              <dbl> 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.…
$ openh1totou            <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ openh1totodds          <dbl> 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00,…
$ openh1ml               <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ half1spr               <dbl> 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.…
$ half1fav               <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ half1sprodds           <dbl> 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00,…
$ half1tot               <dbl> 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.…
$ half1totou             <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ half1totodds           <dbl> 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00,…
$ half1ml                <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, …
$ x1h_team_score         <dbl> 22, 16, 7, 7, 7, 7, 0, 20, 10, 7, 7, 3, 7, 3, 3…
$ x1h_opp_score          <dbl> 21, 21, 21, 21, 28, 14, 17, 38, 38, 21, 10, 23,…
$ x1h_margin             <dbl> 1, -5, -14, -14, -21, -7, -17, -18, -28, -14, -…
$ x1h_margin_qual        <chr> NA, "2H TRAIL", "2H TRAIL", "2H TRAIL", "2H TRA…
$ x1h_spread_op          <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ x1h_cover_margin_op    <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ x1h_ats_w_op           <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ x1h_ats_l_op           <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ x1h_ats_p_op           <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ x1h_spread_cl          <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ x1h_cover_margin_cl    <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ x1h_ats_w_cl           <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ x1h_ats_l_cl           <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ x1h_ats_p_cl           <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ x1h_total_points       <dbl> 43, 37, 28, 28, 35, 21, 17, 58, 48, 28, 17, 26,…
$ x1h_total_margin_op    <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ x1h_over_op            <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ x1h_under_op           <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ x1h_push_op            <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ x1h_total_margin_cl    <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ x1h_over_cl            <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ x1h_under_cl           <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ x1h_push_cl            <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ half2spr               <dbl> 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.…
$ half2fav               <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ half2sprodds           <dbl> 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00,…
$ half2tot               <dbl> 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.…
$ half2totou             <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ half2totodds           <dbl> 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.…
$ half2ml                <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, …
$ x2h_team_score         <dbl> 7, 21, 31, 6, 14, 7, 10, 7, 21, 0, 6, 20, 8, 6,…
$ x2h_opp_score          <dbl> 13, 17, 14, 13, 9, 24, 21, 17, 13, 24, 7, 32, 2…
$ x2h_margin             <dbl> -6, 4, 17, -7, 5, -17, -11, -10, 8, -24, -1, -1…
$ x2h_spread_cl          <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ x2h_fave_dog           <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ x2h_cover_margin_cl    <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ x2h_ats_w_cl           <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ x2h_ats_l_cl           <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ x2h_ats_p_cl           <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ x2h_total_points       <dbl> 20, 38, 45, 19, 23, 31, 31, 24, 34, 24, 13, 52,…
$ x2h_total_margin       <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ x2h_over               <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ x2h_under              <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ x2h_push               <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ conf                   <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 0, 0, 0, 0,…
$ desc                   <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ confchamp              <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
$ bowl                   <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
$ hfdowns                <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ afdowns                <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hratt                  <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ aratt                  <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hrush                  <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ arush                  <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hpcomp                 <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ apcomp                 <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hpatt                  <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ apatt                  <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hpass                  <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ apass                  <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hint                   <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ aint                   <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hpen                   <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hpenyds                <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ apen                   <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ apenyds                <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hfum                   <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hfumlost               <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ afum                   <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ afumlost               <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hsack                  <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hsackyds               <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ asack                  <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ asackyds               <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ htime                  <dbl> 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00,…
$ atime                  <dbl> 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00,…
$ hpuntret               <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hpuntretyds            <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ apuntret               <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ apuntretyds            <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hintret                <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hintretyds             <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ aintret                <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ aintretyds             <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hkoret                 <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hkoretyds              <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ akoret                 <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ akoretyds              <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ h3downc                <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ h3downa                <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ a3downc                <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ a3downa                <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ h4downc                <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ h4downa                <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ a4downc                <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ a4downa                <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hpunt                  <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hpuntyds               <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hpuntin20              <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hpunttb                <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ apunt                  <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ apuntyds               <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ apuntin20              <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ apunttb                <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hkickoff               <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hkickoffyds            <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hkickofftb             <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ akickoff               <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ akickoffyds            <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ akickofftb             <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hfg20a                 <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hfg20m                 <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hfg30a                 <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hfg30m                 <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hfg40a                 <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hfg40m                 <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hfg50a                 <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hfg50m                 <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ afg20a                 <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ afg20m                 <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ afg30a                 <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ afg30m                 <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ afg40a                 <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ afg40m                 <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ afg50a                 <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ afg50m                 <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hepm                   <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hepa                   <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ aepm                   <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ aepa                   <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ hwin                   <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ hloss                  <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ awin                   <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ aloss                  <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ temp                   <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ cond                   <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ wind                   <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, NA, NA, NA,…
$ dir                    <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ outdoor                <dbl> 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0,…
$ surface                <chr> "G", "FT", "FT", "FT", "FT", "FT", "FT", "FT", …

Data pre-processing

setDT(biq_data)
biq_data = biq_data[!is.na(time)]
biq_data[, ko_time := strftime(time, format = "%H:%M:%S")]
biq_data[, date_time := as.POSIXct(paste(date, ko_time))]
biq_data[, year := year(date_time)]
setorder(biq_data, "date_time", "original_h_a")

# check corresponding team and opponent names are spelled and formatted consistently
biq_data[team %notin% opponent, unique(team)]

character(0)

biq_data[opponent %notin% team, unique(opponent)]

[1] "UMASS"

# only discrepancy is U MASS, so coerce for both team and opponent fields
biq_data[team=="U MASS", team := "UMASS"]
biq_data[opponent=="U MASS", opponent := "UMASS"]

# create unique match-up identifier based on kick-off time and team names
biq_data[, match_id := paste0(date_time, ifelse(original_h_a=="Home", paste0(team, opponent), paste0(opponent, team)))]
biq_data[, duplicate_check := uniqueN(team), by = .(match_id)]
biq_data[duplicate_check!=2] # 0 rows

Empty data.table (0 rows and 238 cols): week,season,week_qualify,date,time,team...

# only keep home team row per match-up (home ordered first)
biq_data = biq_data[!duplicated(match_id, fromLast = T)]
cat("Number matches remaining: ", comma(biq_data[, .N]))

Number matches remaining:  17,965

# From Matt: "Just a heads up to filter out those rows in which the data looks funky / off. Specifically, I'd filter the "OPENTOT" column and remove, or tag, those rows that are either "0", "blank" or any outlier number that doesn't make sense - for example, any "OPENTOT" value below 20 or above 200 are likely errors, as no football total would ever fall outside those boundaries."
biq_data[, summary(opentot)]

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max.    NA's 
   0.00   47.00   53.00   49.32   58.50   90.00     359

biq_data = biq_data[!is.na(opentot) & opentot != 0 & opentot %between% c(20, 200)]

## also remove rows with other missing lines/spreads
opening_prices = c("spread_op", "cover_margin_op", "opentot", "total_cover_margin_op")
closing_prices = c("spread_cl", "cover_margin_cl", "total", "total_cover_margin_cl")
biq_data[, lapply(.SD, function(x) sum(is.na(x))), .SDcols = opening_prices]

   spread_op cover_margin_op opentot total_cover_margin_op
1:         0               0       0                     1

biq_data[, lapply(.SD, function(x) sum(is.na(x))), .SDcols = closing_prices]

   spread_cl cover_margin_cl total total_cover_margin_cl
1:         0               0     0                     1

biq_data = biq_data[!is.na(spread_op) & !is.na(opentot) & !is.na(cover_margin_op) & !is.na(total_cover_margin_op)]
cat("Number matches remaining: ", comma(biq_data[, .N]))

Number matches remaining:  15,945

# Keep game-weeks 0 through 16
biq_data = biq_data[week %in% 0:16]
cat("Number matches remaining: ", comma(biq_data[, .N]))

Number matches remaining:  15,175

Simple descriptive analysis

Spreads:

Sense-checking error distribution on lines: overall symmetric and mean-zero
Spread errors decreasing across seasons (markets becoming more efficient); but errors not significantly decreasing over gameweeks within a season


Call:
lm(formula = op_spread_error ~ year + week, data = biq_data)

Residuals:
   Min     1Q Median     3Q    Max 
-64.14 -10.17  -0.01  10.39  62.65 

Coefficients:
              Estimate Std. Error t value Pr(>|t|)   
(Intercept) 104.644109  40.560552   2.580  0.00989 **
year         -0.052063   0.020156  -2.583  0.00981 **
week          0.004166   0.032781   0.127  0.89886   
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 15.69 on 15172 degrees of freedom
Multiple R-squared:  0.0004396, Adjusted R-squared:  0.0003079 
F-statistic: 3.336 on 2 and 15172 DF,  p-value: 0.03559


Call:
lm(formula = spread_error ~ year + week, data = biq_data)

Residuals:
    Min      1Q  Median      3Q     Max 
-71.225 -10.142  -0.009  10.376  63.175 

Coefficients:
             Estimate Std. Error t value Pr(>|t|)  
(Intercept) 70.196764  40.269292   1.743   0.0813 .
year        -0.034976   0.020012  -1.748   0.0805 .
week         0.004657   0.032545   0.143   0.8862  
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 15.57 on 15172 degrees of freedom
Multiple R-squared:  0.0002017, Adjusted R-squared:  6.994e-05 
F-statistic: 1.531 on 2 and 15172 DF,  p-value: 0.2164

Totals:

Totals lines seem quite well calibrated on average (quite symmetric, with mean-zero error)
Generally, actual totals are increasing both across and within seasons
Opening totals errors are higher in later game-weeks, on average (more line-up uncertainty/fixture congestion/injuries?)
Closing totals errors are higher in later game-weeks, on average. No significant association with year


Call:
lm(formula = actual_total ~ year + week, data = biq_data)

Residuals:
    Min      1Q  Median      3Q     Max 
-51.821 -12.585  -1.053  11.380  89.498 

Coefficients:
              Estimate Std. Error t value Pr(>|t|)    
(Intercept) -341.94665   46.99842  -7.276 3.61e-13 ***
year           0.19644    0.02336   8.411  < 2e-16 ***
week           0.15589    0.03798   4.104 4.08e-05 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 18.18 on 15172 degrees of freedom
Multiple R-squared:  0.005906,  Adjusted R-squared:  0.005775 
F-statistic: 45.07 on 2 and 15172 DF,  p-value: < 2.2e-16


Call:
lm(formula = op_totals_error ~ year + week, data = biq_data)

Residuals:
    Min      1Q  Median      3Q     Max 
-54.467 -11.606  -0.779  10.665  97.828 

Coefficients:
            Estimate Std. Error t value Pr(>|t|)   
(Intercept) 42.52393   43.28158   0.982  0.32587   
year        -0.02143    0.02151  -0.996  0.31915   
week         0.10687    0.03498   3.055  0.00225 **
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 16.74 on 15172 degrees of freedom
Multiple R-squared:  0.0006668, Adjusted R-squared:  0.0005351 
F-statistic: 5.062 on 2 and 15172 DF,  p-value: 0.006345


Call:
lm(formula = totals_error ~ year + week, data = biq_data)

Residuals:
    Min      1Q  Median      3Q     Max 
-75.558 -11.377  -0.819  10.468  98.681 

Coefficients:
             Estimate Std. Error t value Pr(>|t|)  
(Intercept)  5.328234  42.656165   0.125   0.9006  
year        -0.002796   0.021198  -0.132   0.8951  
week         0.087095   0.034474   2.526   0.0115 *
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 16.5 on 15172 degrees of freedom
Multiple R-squared:  0.0004206, Adjusted R-squared:  0.0002889 
F-statistic: 3.192 on 2 and 15172 DF,  p-value: 0.04111

Line movement analysis: opening vs closing

Seems to be less spread movement (from open to close) for home favourites vs. home dogs, on average
Spread movement somewhat increases over gameweeks and across seasons for home dogs
Totals movements are higher earlier in the season, seem to stabilise by GW10.
Totals movements have been higher past few years vs pre-2013.


Call:
lm(formula = spread_change_pct ~ year + week, data = biq_data[favored == 
    "H"])

Residuals:
    Min      1Q  Median      3Q     Max 
-6.9793 -0.1017  0.0122  0.1226  8.0448 

Coefficients:
             Estimate Std. Error t value Pr(>|t|)   
(Intercept) -5.385869   1.643821  -3.276  0.00106 **
year         0.002672   0.000817   3.271  0.00107 **
week        -0.002304   0.001313  -1.755  0.07931 . 
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.5029 on 9506 degrees of freedom
Multiple R-squared:  0.001406,  Adjusted R-squared:  0.001196 
F-statistic: 6.692 on 2 and 9506 DF,  p-value: 0.001247


Call:
lm(formula = spread_change_pct ~ year + week, data = biq_data[favored == 
    "A"])

Residuals:
    Min      1Q  Median      3Q     Max 
-7.3559 -0.0935  0.0521  0.2041 10.0262 

Coefficients:
             Estimate Std. Error t value Pr(>|t|)    
(Intercept) 13.721593   2.899046   4.733 2.27e-06 ***
year        -0.006861   0.001440  -4.763 1.95e-06 ***
week         0.001043   0.002406   0.434    0.665    
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.6798 on 5581 degrees of freedom
Multiple R-squared:  0.004066,  Adjusted R-squared:  0.003709 
F-statistic: 11.39 on 2 and 5581 DF,  p-value: 1.154e-05


Call:
lm(formula = total_change_pct ~ year + week, data = biq_data[favored == 
    "H"])

Residuals:
     Min       1Q   Median       3Q      Max 
-0.29439 -0.03176 -0.00070  0.02877  0.34809 

Coefficients:
              Estimate Std. Error t value Pr(>|t|)    
(Intercept) -1.211e+00  1.579e-01  -7.666 1.95e-14 ***
year         6.047e-04  7.849e-05   7.704 1.45e-14 ***
week        -4.071e-04  1.261e-04  -3.227  0.00125 ** 
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.04832 on 9506 degrees of freedom
Multiple R-squared:  0.007108,  Adjusted R-squared:  0.0069 
F-statistic: 34.03 on 2 and 9506 DF,  p-value: 1.88e-15


Call:
lm(formula = total_change_pct ~ year + week, data = biq_data[favored == 
    "A"])

Residuals:
    Min      1Q  Median      3Q     Max 
-58.693  -0.015   0.015   0.050   0.577 

Coefficients:
             Estimate Std. Error t value Pr(>|t|)
(Intercept)  2.736247   4.327599   0.632    0.527
year        -0.001381   0.002150  -0.642    0.521
week         0.003850   0.003591   1.072    0.284

Residual standard error: 1.015 on 5581 degrees of freedom
Multiple R-squared:  0.0002735, Adjusted R-squared:  -8.478e-05 
F-statistic: 0.7634 on 2 and 5581 DF,  p-value: 0.4662

Qual ratings data

Qualitative ratings are recorded from 2019-2023. Higher numbers represent stronger qualitative edges and the ratings are compared or combined for each match-up.
Generally ratings seem quite well calibrated (spreads/totals monotonic in edge score)
Only simple correlational analysis for now; need to decompose individual edge ratings more and blend with other datasets provided

Rows: 3,461
Columns: 11
$ team_qual_rating       <dbl> 3.8, 8.8, 3.9, 2.0, 3.6, 5.9, 3.0, 5.3, -4.3, 5…
$ opp_qual_rating        <dbl> -4.3, 6.0, 7.6, -10.9, 4.0, -2.1, 6.5, 3.8, 1.2…
$ qual_edge              <dbl> 8.1, 2.8, -3.7, 12.9, -0.4, 8.0, -3.5, 1.5, -5.…
$ qual_edge_cat          <dbl> 8, 3, -4, 13, 0, 8, -4, 2, -6, 2, 9, 6, 7, 0, 7…
$ side_qualify_basic     <chr> "Yes", "Yes", "Yes", NA, "Yes", "Yes", "Yes", "…
$ side_qualify_combined  <chr> "Yes", NA, NA, "Yes", NA, "Yes", NA, NA, NA, NA…
$ team_total_rating      <dbl> -0.9, 1.0, 0.2, 3.9, 1.3, 1.4, -3.7, 4.7, 2.6, …
$ opp_total_rating       <dbl> -1.000000e+00, -2.700000e+00, 1.000000e+00, 2.4…
$ combined_total_edge    <dbl> -1.9, -1.7, 1.2, 6.3, 5.0, 3.6, -2.1, 2.5, 2.6,…
$ basic_total_qualify    <chr> NA, NA, NA, "OVER", NA, NA, "UNDER", "OVER", NA…
$ combined_total_qualify <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…

home_qual_edge_quintile	favored	number_selections	avg_home_win_rate	avg_spread_op	avg_spread_cl	avg_op_error	avg_cl_error	avg_spread_change_pct
(-18,-4.6]	H	229.0	67.2%	−10.4	10.4	−3.1	−3.2	−0.6%
(-18,-4.6]	A	211.0	21.8%	9.7	10.4	−3.1	−2.3	−13.6%
(-18,-4.6]	E	4.0	50.0%	1.0	0.0	2.0	1.0	0.0%
(-4.6,-1.4]	H	258.0	79.5%	−11.9	12.0	0.4	0.3	2.4%
(-4.6,-1.4]	A	187.0	28.9%	8.3	9.0	−1.6	−0.9	−16.5%
(-1.4,1.7]	H	276.0	74.6%	−12.5	12.5	0.3	0.3	2.6%
(-1.4,1.7]	A	163.0	29.4%	9.5	10.3	1.7	2.5	−9.9%
(-1.4,1.7]	E	2.0	0.0%	0.0	0.0	−2.0	−2.0	0.0%
(1.7,4.86]	A	162.0	38.3%	8.6	9.3	1.7	2.3	−18.6%
(1.7,4.86]	H	278.0	78.8%	−12.9	13.4	0.6	0.2	−6.1%
(1.7,4.86]	E	3.0	66.7%	0.5	0.0	11.2	10.7	0.0%
(4.86,17.3]	H	327.0	80.7%	−13.9	14.5	2.1	1.4	−4.5%
(4.86,17.3]	A	115.0	33.0%	8.0	8.3	−1.7	−1.4	−20.6%
(4.86,17.3]	E	2.0	100.0%	1.5	0.0	14.0	12.5	0.0%
NA	A	516.0	30.6%	8.6	9.6	−1.7	−0.7	−19.5%
NA	H	725.0	76.4%	−12.7	12.9	−0.3	−0.4	−0.4%
NA	E	3.0	100.0%	−1.3	0.0	1.7	3.0	0.0%

combined_total_edge_quintile	number_selections	avg_total_op	avg_total_cl	avg_total_score_actual	avg_totals_error_cl	avg_totals_change_pct
(-16.3,-3.2]	444.0	53.0	52.1	51.7	−0.3	2.2%
(-3.2,-0.5]	444.0	54.2	53.3	54.1	0.8	2.0%
(-0.5,1.76]	444.0	55.5	55.0	55.1	0.1	1.3%
(1.76,4.3]	445.0	55.1	54.7	56.1	1.4	1.2%
(4.3,15.6]	444.0	55.3	55.2	55.7	0.6	0.5%
NA	1,240.0	56.0	56.1	56.0	−0.2	0.1%