ratingslib.datasets.soccer module

This module includes all parameters and configuration for soccer. column names are based on football-data.co.uk data scheme.

Notes

The dictionary COLUMNS_DICT is based on football-data.co.uk data files for other data-source change dictionary values specify column names:

• key is the variable name in code

• value is the column_name

• ‘ITEM_I’: column name for home team

• ‘ITEM_J’: column name for away team

• ‘points_i’ : column name for home team goals

• ‘points_j’ : column name for away team goals

• ‘HS’ : column name for home team total shoots

• ‘AS’ : column name for away team total shoots

• ‘HST’ : column name of home team total shoots on target

• ‘AST’ : column name of away team total shoots on target

1. date column is written at the beginning of this module (DATE_COL var)

2. HS, AS, HST, AST columns below are used in: ratingslib.ratings.markov.Markov ratingslib.ratings.accurate.AccuRate

parser_date_func(x)

class championships

Bases: object

Class for championship names

PREMIERLEAGUE = 'EPL'

ENG_1 = 'E1'

ENG_2 = 'E2'

ENG_3 = 'E3'

ENG_CONF = 'ECONF'

SCOTLAND_PS = 'SPS'

SCOTLAND_1 = 'SC1'

SCOTLAND_2 = 'SC2'

SCOTLAND_3 = 'SC3'

BUNDESLIGAS_1 = 'GBL1'

BUNDESLIGAS_2 = 'GBL2'

ITALY_A = 'IA'

ITALY_B = 'IB'

SPAIN_PREMERA = 'SPP'

SPAIN_SEGUNDA = 'SPS'

FRANCE_CHAMP = 'FRCH'

FRANCE_2 = 'FR2'

NETHERLANDS = 'NKPN'

BELGIUM = 'BEL'

PORTUGAL = 'PLIGA'

TURKEY = 'TLIGI'

GREECE_SL = 'GREK'

class stats

Bases: object

Class for soccer statistics, In this class all statistics can represented here.

1. TW : TOTAL WINS/NUMBER OF GAMES

2. TG : TOTAL GOALS/NUMBER OF GAMES

3. TS : TOTAL SHOOTS/NUMBER OF GAMES (for soccer)

4. TST : TOTAL SHOOTS TARGET/NUMBER OF GAMES (for soccer)

• STATISTIC COLUMNS DICTIONARY

  ITEM_I: column name of statistic for home team e.g. in football-data.co.uk the column for goals is FTHG.

  ITEM_J: column name of statistic for away team e.g. in football-data.co.uk the column for goals is FTAG.

  TYPE: {WIN, POINTS}

  1. if type is WIN then compare if ITEM_I > ITEM_J or ITEM_J < ITEM_I

  2. if type is POINTS then count points

for more information about the TYPE see the implementation of ratingslib.ratings.methods.calc_items_stats()

• PARAMS DICTIONARY FOR MARKOV RATING SYSTEMS

  • key = statistic name e.g. TW

  • value = dictionary where

  VOTE : total votes (e.g. 10), then will be converted as weight.

  ITEM_I: H: column name of statistic for home team e.g. in football-data.co.uk the column for goals is FTHG.

  ITEM_J: column name of statistic for away team e.g. in football-data.co.uk the column for goals is FTAG.

METHOD:{‘VotingWithLosses’, ‘WinnersAndLosersVotePoint’, ‘LosersVotePointDiff’} for more information about the methods see ratingslib.ratings.markov module.

TW = 'TW'

TG = 'TG'

TS = 'TS'

TST = 'TST'

STATS_COLUMNS_DICT = {'TG': {'ITEM_I': 'FTHG', 'ITEM_J': 'FTAG', 'TYPE': 'POINTS'}, 'TS': {'ITEM_I': 'HS', 'ITEM_J': 'AS', 'TYPE': 'POINTS'}, 'TST': {'ITEM_I': 'HST', 'ITEM_J': 'AST', 'TYPE': 'POINTS'}, 'TW': {'ITEM_I': 'FTHG', 'ITEM_J': 'FTAG', 'TYPE': 'WIN'}}

STATS_MARKOV_EQUAL_DICT = {'TG': {'ITEM_I': 'FTHG', 'ITEM_J': 'FTAG', 'METHOD': 'WinnersAndLosersVotePoint', 'VOTE': 10.0}, 'TS': {'ITEM_I': 'HS', 'ITEM_J': 'AS', 'METHOD': 'WinnersAndLosersVotePoint', 'VOTE': 10.0}, 'TST': {'ITEM_I': 'HST', 'ITEM_J': 'AST', 'METHOD': 'WinnersAndLosersVotePoint', 'VOTE': 10.0}, 'TW': {'ITEM_I': 'FTHG', 'ITEM_J': 'FTAG', 'METHOD': 'VotingWithLosses', 'VOTE': 10.0}}

STATS_MARKOV_DICT = {'TG': {'ITEM_I': 'FTHG', 'ITEM_J': 'FTAG', 'METHOD': 'WinnersAndLosersVotePoint', 'VOTE': 8.0}, 'TS': {'ITEM_I': 'HS', 'ITEM_J': 'AS', 'METHOD': 'WinnersAndLosersVotePoint', 'VOTE': 4.0}, 'TST': {'ITEM_I': 'HST', 'ITEM_J': 'AST', 'METHOD': 'WinnersAndLosersVotePoint', 'VOTE': 6.0}, 'TW': {'ITEM_I': 'FTHG', 'ITEM_J': 'FTAG', 'METHOD': 'VotingWithLosses', 'VOTE': 10.0}}