Game Results

calc_spread.jl

#!/usr/bin/env julia

function get_team_games(games::DataFrame, team_names::Vector, and=true)
    df = ""
    for i = 1:nrow(games)
            if (and && contains(team_names, games[i,"AwayTeam"]) && contains(team_names, games[i, "HomeTeam"])) || (!and && (contains(team_names, games[i,"AwayTeam"]) || contains(team_names, games[i, "HomeTeam"])))
            if df == "" 
                df = games[i,:]
            else 
                df = rbind(df, games[i,:])
            end
        end
    end
    df
end


function get_teams(games)
    teams = Set()
    teams = union!(union!(teams, vector(games["AwayTeam"])), vector(games["HomeTeam"]))
    teams = sort!(collect(teams))
end

function calculate_ratings(games, teams=get_teams(games))
    games = get_team_games(games, teams)
    X = zeros(Float64, nrow(games), length(teams))
    Y = zeros(Float64, nrow(games))
    for i in 1:nrow(games)
        team1 = findfirst(teams, games[i, "AwayTeam"])
        team2 = findfirst(teams, games[i, "HomeTeam"])

        spread = games[i, "WinMargin"]
        if spread != 0
            X[i, team1] = sign(games[i, "WinMargin"])
            X[i, team2] = -sign(games[i, "WinMargin"])
        else
            X[i, team1] = 1.0
            X[i, team2] = -1.0
        end
        Y[i] = spread
    end
    Ratings = X \ Y
end
function calculate_standings(games, teams=get_teams(games))
    N = length(teams)
    gp = zeros(Int64, N)
    wins = zeros(Int64, N)
    loss = zeros(Int64, N)
    ties = zeros(Int64, N)
    pts = zeros(Int64, N)
    gf = zeros(Int64, N)
    ga = zeros(Int64, N)
    diff = zeros(Int64, N)
    winpct = zeros(Float64, N)
    ratings = calculate_ratings(games, teams)

    for i in 1:nrow(games)
        team1 = findfirst(teams, games[i, "AwayTeam"])
        team2 = findfirst(teams, games[i, "HomeTeam"])
        spread = games[i, "WinMargin"]
        gp[team1] += 1
        gp[team2] += 1
        if spread > 0
            wins[team1] += 1
            loss[team2] += 1
        elseif spread < 0
            wins[team2] += 1
            loss[team1] += 1
        else
            ties[team1] += 1
            ties[team2] += 1
        end
        gf[team1] += games[i, "AwayScore"]
        gf[team2] += games[i, "HomeScore"]
        ga[team1] += games[i, "HomeScore"]
        ga[team2] += games[i, "AwayScore"]
    end 
    for team in 1:N
        pts[team] = wins[team] * 2 + ties[team]
        diff[team] = gf[team] - ga[team]
        winpct[team] = (wins[team] + ties[team]/2) / gp[team]
    end
    standings = DataFrame(Team=teams, GP=gp, W=wins, L=loss, T=ties, PTS=pts, GF=gf, GA=ga, DIFF=diff, PCT=winpct, Rating=ratings)
end


games = readtable("games.csv")
ratings = calculate_ratings(games)
standings_by_points = sortby(calculate_standings(games), "PTS")
standings_by_rating = sortby(A_standings, "Rating")

games.csv

results.jl

julia> reload("calc_spread.jl")

julia> standings_by_points
6x11 DataFrame:
        Team GP  W  L T PTS GF GA DIFF       PCT    Rating
[1,]     "F" 15  0 14 1   1 27 77  -50 0.0333333  -1.48166
[2,]     "E" 14  5  9 0  10 42 58  -16  0.357143  0.456697
[3,]     "D" 16  6  9 1  13 44 69  -25   0.40625 -0.239425
[4,]     "C" 15  9  5 1  19 52 34   18  0.633333  0.377248
[5,]     "B" 17 11  5 1  23 73 45   28  0.676471  0.309839
[6,]     "A" 17 14  3 0  28 85 40   45  0.823529    0.5773

julia> standings_by_rating
6x12 DataFrame:
        Team GP  W  L T OT PTS GF GA DIFF       PCT    Rating
[1,]     "F" 15  0 14 1  0   1 27 77  -50 0.0333333  -1.48166
[2,]     "D" 16  6  9 1  0  13 44 69  -25   0.40625 -0.239425
[3,]     "B" 17 11  5 1  0  23 73 45   28  0.676471  0.309839
[4,]     "C" 15  9  5 1  0  19 52 34   18  0.633333  0.377248
[5,]     "E" 14  5  9 0  0  10 42 58  -16  0.357143  0.456697  # This rating seems wrong
[6,]     "A" 17 14  3 0  0  28 85 40   45  0.823529    0.5773

simple_spread.jl

X = readdlm("X.tsv", '\t')
Y = readdlm("Y.tsv", '\t')
X\Y

X.tsv

Y.tsv

-4
-3
1
1
-4
11
-2
-4
-2
5
-6
4
-1
-1
-1
6
4
6
6
-1
-2
0
-4
7
-1
8
4
3
1
-1
-1
-1
-5
2
3
-2
-2
-4
2
0
-6
2
4
5
4
3
3