Add Strike%, F-Str%, PAR%, and Behind%

pitch_leaderboard.py

@@ -10,9 +10,9 @@ from stats import compute_pitch_stats, filter_data_by_date_and_game_kind
 from convert import ball_kind, ball_kind_to_color, get_text_color_from_color, team_names_short_to_color, get_text_color_from_team
 from plotting import stat_cmap
 
-STATS = ['Count', 'Usage', 'Avg Velo', 'Max Velo', 'Swing%', 'Z-Swing%', 'Chase%', 'Contact%', 'Z-Contact%', 'O-Contact%', 'SwStr%', 'Whiff%', 'CSW%', 'Ball%', 'GB%', 'FB%', 'LD%', 'Zone%', 'Arm%', 'Glove%', 'High%', 'Low%', 'MM%']
-PCT_STATS = ['Usage', 'Swing%', 'Z-Swing%', 'Chase%', 'Contact%', 'Z-Contact%', 'O-Contact%', 'SwStr%', 'Whiff%', 'CSW%', 'Ball%', 'GB%', 'FB%', 'LD%', 'Zone%', 'Arm%', 'Glove%', 'High%', 'Low%', 'MM%']
-STATS_WITH_PCTLS = ['Swing%', 'Z-Swing%', 'Chase%', 'Contact%', 'Z-Contact%', 'O-Contact%', 'SwStr%', 'Whiff%', 'CSW%', 'Ball%', 'GB%', 'FB%', 'LD%', 'Zone%']
+STATS = ['Count', 'Usage', 'Avg Velo', 'Max Velo', 'Swing%', 'Z-Swing%', 'Chase%', 'Contact%', 'Z-Contact%', 'O-Contact%', 'SwStr%', 'Whiff%', 'CSW%', 'Strike%', 'Ball%', 'F-Str%', 'PAR%', 'GB%', 'FB%', 'LD%', 'Zone%', 'Arm%', 'Glove%', 'High%', 'Low%', 'MM%', 'Behind%']
+PCT_STATS = ['Usage', 'Swing%', 'Z-Swing%', 'Chase%', 'Contact%', 'Z-Contact%', 'O-Contact%', 'SwStr%', 'Whiff%', 'CSW%', 'Strike%', 'Ball%', 'F-Str%', 'PAR%', 'GB%', 'FB%', 'LD%', 'Zone%', 'Arm%', 'Glove%', 'High%', 'Low%', 'MM%', 'Behind%']
+STATS_WITH_PCTLS = ['Swing%', 'Z-Swing%', 'Chase%', 'Contact%', 'Z-Contact%', 'O-Contact%', 'SwStr%', 'Whiff%', 'CSW%', 'Strike%', 'Ball%', 'F-Str%', 'PAR%', 'GB%', 'FB%', 'LD%', 'Zone%']
 COLUMNS = ['Pitcher', 'Team', 'Throws', 'Pitch', 'Pitch (General)'] + STATS
 
 PITCH_TYPES = [pitch_type for pitch_type in ball_kind.values() if pitch_type != '-']

player_team_leaderboard.py

@@ -57,9 +57,9 @@ def create_player_team_leaderboard_app(player_team_type):
 
   # stats
   if pitching:
-    pct_stats = ['K%', 'BB%', 'Swing%', 'Z-Swing%', 'Chase%', 'Contact%', 'Z-Contact%', 'O-Contact%', 'SwStr%', 'Whiff%', 'CSW%', 'Ball%', 'GB%', 'FB%', 'LD%', 'Zone%', 'Arm%', 'Glove%', 'High%', 'Low%', 'MM%']
-    stats_with_pctls = ['FB Velo', 'K%', 'BB%', 'Swing%', 'Z-Swing%', 'Chase%', 'Contact%', 'Z-Contact%', 'O-Contact%', 'SwStr%', 'Whiff%', 'CSW%', 'Ball%', 'GB%', 'FB%', 'LD%', 'Zone%']
-    cols = ['Pitcher', 'Team', 'Throws', 'IP', 'TBF', 'FB Velo', 'K%', 'BB%', 'Swing%', 'Z-Swing%', 'Chase%', 'Contact%', 'Z-Contact%', 'O-Contact%', 'SwStr%', 'Whiff%', 'CSW%', 'Ball%', 'GB%', 'FB%', 'LD%', 'Zone%', 'Arm%', 'Glove%', 'High%', 'Low%', 'MM%']
+    pct_stats = ['K%', 'BB%', 'Swing%', 'Z-Swing%', 'Chase%', 'Contact%', 'Z-Contact%', 'O-Contact%', 'SwStr%', 'Whiff%', 'CSW%', 'Strike%', 'Ball%', 'F-Str%', 'PAR%', 'GB%', 'FB%', 'LD%', 'Zone%', 'Arm%', 'Glove%', 'High%', 'Low%', 'MM%', 'Behind%']
+    stats_with_pctls = ['FB Velo', 'K%', 'BB%', 'Swing%', 'Z-Swing%', 'Chase%', 'Contact%', 'Z-Contact%', 'O-Contact%', 'SwStr%', 'Whiff%', 'CSW%', 'Strike%', 'F-Str%', 'PAR%', 'Ball%', 'GB%', 'FB%', 'LD%', 'Zone%', 'Behind%']
+    cols = ['Pitcher', 'Team', 'Throws', 'IP', 'TBF', 'FB Velo', 'K%', 'BB%', 'Swing%', 'Z-Swing%', 'Chase%', 'Contact%', 'Z-Contact%', 'O-Contact%', 'SwStr%', 'Whiff%', 'CSW%', 'Strike%', 'Ball%', 'F-Str%', 'PAR%', 'GB%', 'FB%', 'LD%', 'Zone%', 'Arm%', 'Glove%', 'High%', 'Low%', 'MM%', 'Behind%']
     if team:
       cols = [col for col in cols if col not in ('Pitcher', 'Throws')]
   else:

stats.py

@@ -18,6 +18,12 @@ whiff = (pl.col('whiff').sum() / pl.col('swing').sum()).alias('Whiff%')
 swstr = (pl.col('whiff').sum() / pl.col('pitch').sum()).alias('SwStr%')
 csw = (pl.col('csw').sum() / pl.col('pitch').sum()).alias('CSW%')
 ball = ((pl.col('presult') == 'Ball').sum() / pl.col('pitch').sum()).alias('Ball%')
+strike = ((pl.col('pitch') & (pl.col('presult') != 'Ball')).sum() / pl.col('pitch').sum()).alias('Strike%')
+is_two_str = pl.col('before_s') == 2 # named this way in case I use two_str for 2-Str%
+first_count = (pl.col('before_s') == 0) & (pl.col('before_b') == 0)
+f_strike = ((pl.col('csw') & first_count).sum() / first_count.sum()).alias('F-Str%')
+par = (((is_two_str & pl.col('presult').str.contains('strikeout')).sum()) / is_two_str.sum()).alias('PAR%')
+behind = (((pl.col('before_b') > pl.col('before_s')) & (pl.col('before_s') < 2) & (pl.col('before_b') > 1)).sum() / pl.len()).alias('Behind%')
 zone = (pl.col('zone').sum() / pl.col('pitch').sum()).alias('Zone%')
 glove = (pl.when(pl.col('pitLR') == 'r').then(pl.col('x') < 0).otherwise(pl.col('x') > 0)).mean().alias('Glove%')
 arm = (pl.when(pl.col('pitLR') == 'r').then(pl.col('x') >= 0).otherwise(pl.col('x') <= 0)).mean().alias('Arm%')
@@ -139,13 +145,17 @@ def compute_pitch_stats(data, player_type, pitch_class_type, min_pitches=1, pitc
           whiff,
           swstr,
           csw,
+          strike,
           ball,
+          f_strike,
+          par,
           zone,
           glove,
           arm,
           high,
           low,
           mm,
+          behind
       )
       .with_columns(
           (pl.col('count')/pl.sum('count').over('pitId')).alias('usage'),
@@ -162,8 +172,8 @@ def compute_pitch_stats(data, player_type, pitch_class_type, min_pitches=1, pitc
       )
       .drop('G', 'F', 'B', 'P', 'L', 'null')
       .with_columns(
-          (pl.when(pl.col('qualified')).then(pl.col(stat)).rank(descending=((stat in ['FB%', 'LD%', 'Ball%'] or 'Contact%' in stat)))/pl.when(pl.col('qualified')).then(pl.col(stat)).count()).alias(f'{stat}_pctl')
-          for stat in ['Avg KPH', 'Max KPH', 'Avg MPH', 'Max MPH', 'Swing%', 'Z-Swing%', 'Chase%', 'Contact%', 'Z-Contact%', 'O-Contact%', 'SwStr%', 'Whiff%', 'CSW%', 'Ball%', 'GB%', 'FB%', 'LD%', 'Zone%']
+          (pl.when(pl.col('qualified')).then(pl.col(stat)).rank(descending=((stat in ['FB%', 'LD%', 'Ball%', 'Behind%'] or 'Contact%' in stat)))/pl.when(pl.col('qualified')).then(pl.col(stat)).count()).alias(f'{stat}_pctl')
+          for stat in ['Avg KPH', 'Max KPH', 'Avg MPH', 'Max MPH', 'Swing%', 'Z-Swing%', 'Chase%', 'Contact%', 'Z-Contact%', 'O-Contact%', 'SwStr%', 'Whiff%', 'CSW%', 'Strike%', 'Ball%', 'F-Str%', 'PAR%', 'GB%', 'FB%', 'LD%', 'Zone%', 'Behind%']
       )
       .rename({pitch_col: 'ballKind_code', pitch_name_col: 'ballKind'} if pitch_class_type == 'general' else {})
       .sort(id_cols[0], 'count', descending=[False, True])
@@ -224,9 +234,9 @@ def compute_player_stats(data, player_type, qual='qualified', pitcher_lr='both',
 
   # percentile ascending/descending
   if pitching:
-    stat_descending_pctl = lambda stat: stat in ['BB%', 'Ball%', 'FB%', 'LD%', 'Z-Swing%', 'OBP'] or 'Contact%' in stat
+    stat_descending_pctl = lambda stat: stat in ['BB%', 'Ball%', 'FB%', 'LD%', 'Z-Swing%', 'Behind%', 'OBP'] or 'Contact%' in stat
   else:
-    stat_descending_pctl = lambda stat: not (stat in ['BB%', 'Ball%', 'FB%', 'LD%', 'Swing%', 'Z-Swing%', 'OBP'] or 'Contact%' in stat)
+    stat_descending_pctl = lambda stat: not (stat in ['BB%', 'Ball%', 'FB%', 'LD%', 'Swing%', 'Z-Swing%', 'Behind%', 'OBP'] or 'Contact%' in stat)
 
   # col names
   match player_type:
@@ -273,13 +283,17 @@ def compute_player_stats(data, player_type, qual='qualified', pitcher_lr='both',
         whiff,
         swstr,
         csw,
+        strike,
         ball,
+        f_strike,
+        par,
         zone,
         glove,
         arm,
         high,
         low,
         mm,
+        behind,
         pl.col('AB').first(),
         h,
         bb,
@@ -300,7 +314,7 @@ def compute_player_stats(data, player_type, qual='qualified', pitcher_lr='both',
     .drop('G', 'F', 'B', 'P', 'L')
     .with_columns(
         (pl.when(pl.col('qualified')).then(pl.col(stat)).rank(descending=stat_descending_pctl(stat))/pl.when(pl.col('qualified')).then(pl.col(stat)).count()).alias(f'{stat}_pctl')
-        for stat in ['FB Velo', 'K%', 'BB%', 'Swing%', 'Z-Swing%', 'Chase%', 'Contact%', 'Z-Contact%', 'O-Contact%', 'SwStr%', 'Whiff%', 'CSW%', 'Ball%', 'GB%', 'FB%', 'LD%', 'Zone%', 'OBP']
+        for stat in ['FB Velo', 'K%', 'BB%', 'Swing%', 'Z-Swing%', 'Chase%', 'Contact%', 'Z-Contact%', 'O-Contact%', 'SwStr%', 'Whiff%', 'CSW%', 'Strike%', 'Ball%', 'F-Str%', 'PAR%', 'GB%', 'FB%', 'LD%', 'Zone%', 'Behind%', 'OBP']
     )
     .sort(qual_col, descending=True)
   )