Introduction

dates <- c("2015-01-30","2015-02-06","2015-02-13","2015-02-20","2015-02-27","2015-03-06","2015-03-13","2015-03-20")

This is the number 8 in a series of forecasts of football match outcomes, following on from my efforts last week and the previous weeks. The method of forecasts is unchanged from previous week.

Bookmaker prices are compared with forecasts, potentially offering opportunities, should you sufficiently believe my forecasts to be better than those of bookmakers. Naturally, I make no such claim, and indeed since the bookmaker prices were collected, they may well have changed.

Loading the Data

The dataset used is all English matches recorded on http://www.soccerbase.com, which goes back to 1877 and the very first football matches. Experimentation will take place in time (i.e., not this week) with adjusting the estimation sample size, since it is not necessarily useful to have all matches back to 1877 when forecasting matches in 2015. The Elo ranks have been calculated since the very first matches, and hence historical information is retained, to the extent that it is useful in determining a team’s current strength, back throughout footballing history.

Forecast Model

A linear regression model is estimated here and reported:

library(knitr)
library(MASS)
date.1 <- tail(dates,1)
forecast.matches <- read.csv(paste("forecasts_",date.1,".csv",sep=""),stringsAsFactors=F)
forecast.matches <- forecast.matches[is.na(forecast.matches$outcome)==F,]

res.eng <- read.csv(paste("historical_",date.1,".csv",sep=""))
model <- lm(outcome ~ E.1 + pts1 + pts.D + pts.D.2 + pld1 + pld.D + pld.D.2 + gs1 + gs.D + gs.D.2 
            + gd1 + gd.D + gd.D.2 
            + pos1 + pos.D + pos.D.2 + form1 + form.D + form.D.2 + tier1 + tier.D + tier.D.2 + season.d,
            data=res.eng)
summary(model)

## 
## Call:
## lm(formula = outcome ~ E.1 + pts1 + pts.D + pts.D.2 + pld1 + 
##     pld.D + pld.D.2 + gs1 + gs.D + gs.D.2 + gd1 + gd.D + gd.D.2 + 
##     pos1 + pos.D + pos.D.2 + form1 + form.D + form.D.2 + tier1 + 
##     tier.D + tier.D.2 + season.d, data = res.eng)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -1.0273 -0.3115  0.1149  0.3370  0.8955 
## 
## Coefficients:
##               Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  3.311e-01  5.417e-03  61.116  < 2e-16 ***
## E.1          5.449e-01  6.050e-03  90.065  < 2e-16 ***
## pts1         9.878e-04  4.237e-04   2.331  0.01974 *  
## pts.D       -2.381e-03  3.076e-04  -7.738 1.01e-14 ***
## pts.D.2     -1.374e-05  6.467e-06  -2.125  0.03360 *  
## pld1        -1.660e-03  6.092e-04  -2.724  0.00644 ** 
## pld.D        2.273e-03  7.110e-04   3.197  0.00139 ** 
## pld.D.2     -4.053e-05  3.083e-05  -1.314  0.18870    
## gs1          4.755e-04  1.709e-04   2.782  0.00541 ** 
## gs.D        -2.457e-04  1.525e-04  -1.611  0.10725    
## gs.D.2      -5.899e-07  4.687e-06  -0.126  0.89984    
## gd1         -6.014e-04  2.406e-04  -2.499  0.01245 *  
## gd.D         2.456e-03  1.759e-04  13.964  < 2e-16 ***
## gd.D.2      -5.564e-06  2.342e-06  -2.376  0.01751 *  
## pos1         7.665e-04  3.005e-04   2.551  0.01075 *  
## pos.D       -1.841e-03  2.514e-04  -7.324 2.42e-13 ***
## pos.D.2      3.743e-05  1.170e-05   3.198  0.00138 ** 
## form1        6.187e-04  3.518e-04   1.759  0.07858 .  
## form.D       1.355e-03  2.959e-04   4.579 4.67e-06 ***
## form.D.2    -7.278e-05  2.995e-05  -2.430  0.01509 *  
## tier1        1.894e-03  7.650e-04   2.476  0.01328 *  
## tier.D      -3.171e-02  2.862e-03 -11.081  < 2e-16 ***
## tier.D.2    -5.244e-03  1.258e-03  -4.170 3.05e-05 ***
## season.d    -1.084e-03  3.080e-05 -35.183  < 2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.3948 on 216186 degrees of freedom
##   (38368 observations deleted due to missingness)
## Multiple R-squared:  0.0852, Adjusted R-squared:  0.0851 
## F-statistic: 875.4 on 23 and 216186 DF,  p-value: < 2.2e-16

An ordered logistic regression model is:

model.ord <- polr(as.factor(outcome) ~ E.1 + pts1 + pts.D + pts.D.2 + pld1 + pld.D + pld.D.2 + 
                    gs1 + gs.D + gs.D.2 + gd1 + gd.D + gd.D.2 + pos1 + pos.D + pos.D.2 + 
                    form1 + form.D + form.D.2 + tier1 + tier.D + tier.D.2 + season.d, 
                  data=res.eng, method = "logistic")
summary(model.ord)

## 
## Re-fitting to get Hessian

## Call:
## polr(formula = as.factor(outcome) ~ E.1 + pts1 + pts.D + pts.D.2 + 
##     pld1 + pld.D + pld.D.2 + gs1 + gs.D + gs.D.2 + gd1 + gd.D + 
##     gd.D.2 + pos1 + pos.D + pos.D.2 + form1 + form.D + form.D.2 + 
##     tier1 + tier.D + tier.D.2 + season.d, data = res.eng, method = "logistic")
## 
## Coefficients:
##               Value Std. Error   t value
## E.1       2.636e+00  3.017e-02  87.37266
## pts1      4.240e-03  2.083e-03   2.03574
## pts.D    -1.226e-02  1.521e-03  -8.06218
## pts.D.2  -7.683e-05  3.608e-05  -2.12950
## pld1     -9.243e-03  2.996e-03  -3.08461
## pld.D     1.289e-02  3.567e-03   3.61445
## pld.D.2  -2.331e-04  1.637e-04  -1.42353
## gs1       3.426e-03  8.509e-04   4.02692
## gs.D     -1.565e-03  7.579e-04  -2.06455
## gs.D.2   -4.351e-07  2.555e-05  -0.01703
## gd1      -3.355e-03  1.193e-03  -2.81213
## gd.D      1.333e-02  8.793e-04  15.16266
## gd.D.2    7.987e-06  1.637e-05   0.48789
## pos1      3.054e-03  1.466e-03   2.08409
## pos.D    -7.976e-03  1.240e-03  -6.43396
## pos.D.2   2.025e-04  5.944e-05   3.40642
## form1     2.817e-03  1.719e-03   1.63933
## form.D    7.045e-03  1.456e-03   4.83698
## form.D.2 -2.289e-04  1.494e-04  -1.53202
## tier1     9.204e-03  3.726e-03   2.47018
## tier.D   -1.746e-01  1.505e-02 -11.59582
## tier.D.2 -6.568e-03  6.882e-03  -0.95439
## season.d -5.416e-03  1.526e-04 -35.49424
## 
## Intercepts:
##       Value    Std. Error t value 
## 0|0.5   0.2396   0.0264     9.0811
## 0.5|1   1.4285   0.0266    53.7811
## 
## Residual Deviance: 428424.66 
## AIC: 428474.66 
## (38368 observations deleted due to missingness)

The Forecasts

simpleplot <- function(div) {
  matches <- forecast.matches[forecast.matches$division==div,]
  matches <- matches[order(matches$date),]
  matches$id <- 1:NROW(matches)
  par(mar=c(9,4,4,5)+.1)
  plot(matches$id,matches$outcome,xaxt="n",xlab="",ylim=range(0,1),
       main=paste("Forecasts of Weekend ",div," Matches",sep=""),
       ylab="Probability of Outcome")
  lines(matches$id,matches$Ph,col=2,pch=15,type="p")
  lines(matches$id,matches$Pd,col=3,pch=16,type="p")
  lines(matches$id,matches$Pa,col=4,pch=17,type="p")
  legend("topleft",ncol=4,pch=c(1,15,16,17),col=c(1:4),
         legend=c("OLS","OL (home)","OL (draw)","OL (away)"),bty="n")
  abline(h=0.5,lty=2)
  abline(h=0.6,lty=3)
  abline(h=0.7,lty=2)
  axis(1,at=matches$id,labels=paste(matches$team1,matches$team2,sep=" v "),las=2,cex.axis=0.65)
  for(i in 2:NROW(matches)){
    if(matches$date[i]!=matches$date[i-1]) {
      lines(rep(c(i-0.5),2),c(0,1),lty=2)
    }
  }
}

bkplot <- function(div,div.short) {
  matches <- forecast.matches[forecast.matches$division==div,]
  matches <- matches[order(matches$date),]
  matches$id <- 1:NROW(matches)
  bk <- data.frame()
  loc <- paste("/home/readejj/Dropbox/Research/Data for Ideas/Betting/football/",Sys.Date(),"/",div.short,sep="")
  bk.matches <- dir(loc,pattern="*.csv")
  for(i in bk.matches) {
    temp <- read.csv(paste(loc,i,sep="/"),stringsAsFactors=F)
    temp <- clean.data(temp)
    temp$event <- gsub("[.]csv","",i)
    bk <- rbind(bk,temp)
  }
  bk <- bk[order(bk$event,bk$Date.Time),]
  bk$mean <- rowMeans(bk[colnames(bk)[nchar(colnames(bk))==2]],na.rm=T)
  bk$event <- gsub("sheffield-wednesday","sheff wed",bk$event)
  bk$event <- gsub("nottingham-forest","nottm forest",bk$event)
  bk$event <- gsub("middlesbrough","middlesbro",bk$event)
  bk$event <- gsub("man-utd","man utd",bk$event)
  bk$event <- gsub("man-city","man city",bk$event)
  bk$event <- gsub("aston-villa","aston villa",bk$event)
  bk$event <- gsub("west-brom","west brom",bk$event)
  bk$event <- gsub("west-ham","west ham",bk$event)
  bk$event <- gsub("crystal-palace","c palace",bk$event)
  bk$team1 <- gsub("^(.*?)-v-(.*?)-(.*?)$","\\1",bk$event)
  bk$team2 <- gsub("^(.*?)-v-(.*?)-(.*?)$","\\2",bk$event)
  bk$event <- gsub("^(.*?)-v-(.*?)-(.*?)$","\\3",bk$event)
  bk$event.1 <- c(bk$event[-1],NA)
  bk.h <- bk[(bk$event!=bk$event.1 | is.na(bk$event.1)==T) & bk$event==bk$team1,]
  bk.d <- bk[bk$event!=bk$event.1 & regexpr("draw",bk$event)>-1,]
  bk.a <- bk[bk$event!=bk$event.1 & bk$event==bk$team2,]
  matches$team1 <- tolower(matches$team1)
  matches$team2 <- tolower(matches$team2)
  
  matches <- merge(matches,bk.h[c(colnames(bk.h[nchar(colnames(bk.h))==2]),"team1","team2")],by=c("team1","team2"),all.x=T)
  colnames(matches)[colnames(matches) %in% colnames(bk.h) & regexpr("team",colnames(matches))==-1] <- paste(colnames(matches)[colnames(matches) %in% colnames(bk.h) & regexpr("team",colnames(matches))==-1],".h",sep="")
  matches <- merge(matches,bk.d[c(colnames(bk.d[nchar(colnames(bk.d))==2]),"team1","team2")],by=c("team1","team2"),all.x=T)
  colnames(matches)[colnames(matches) %in% colnames(bk.d) & regexpr("team",colnames(matches))==-1] <- paste(colnames(matches)[colnames(matches) %in% colnames(bk.d) & regexpr("team",colnames(matches))==-1],".d",sep="")
  matches <- merge(matches,bk.a[c(colnames(bk.a[nchar(colnames(bk.a))==2]),"team1","team2")],by=c("team1","team2"),all.x=T)
  colnames(matches)[colnames(matches) %in% colnames(bk.a) & regexpr("team",colnames(matches))==-1] <- paste(colnames(matches)[colnames(matches) %in% colnames(bk.a) & regexpr("team",colnames(matches))==-1],".a",sep="")
  
  par(mar=c(9,4,4,5)+.1)
  plot(matches$id,matches$Ph,xaxt="n",xlab="",ylim=range(0,1),col=2,pch=15,type="p",cex=1.5,
       main=paste("Forecasts of Weekend ",div," Matches",sep=""),
       ylab="Probability of Outcome")
  for(i in colnames(matches)[grep("[.]h",colnames(matches))]) {
    for(j in matches$id) {
      lines(j,1/matches[matches$id==j,i],col="darkred",pch=0,type="p",cex=0.5)
    }  
  }
  
  lines(matches$id,matches$Pd,col=3,pch=16,type="p",cex=1.5)
  for(i in colnames(matches)[grep("[.]d",colnames(matches))]) {
    for(j in matches$id) {
      lines(j,1/matches[matches$id==j,i],col="darkgreen",pch=0,type="p",cex=0.5)
    }  
  }
  lines(matches$id,matches$Pa,col=4,pch=17,type="p",cex=1.5)
  for(i in colnames(matches)[grep("[.]a",colnames(matches))]) {
    for(j in matches$id) {
      lines(j,1/matches[matches$id==j,i],col="darkblue",pch=0,type="p",cex=0.5)
    }  
  }
  legend("topleft",ncol=3,pch=c(15,16,17),col=c(2:4),
         legend=c("OL (home)","OL (draw)","OL (away)"),bty="n")
  abline(h=0.5,lty=2)
  abline(h=0.6,lty=3)
  abline(h=0.7,lty=2)
  abline(h=0.4,lty=3)
  axis(1,at=matches$id,labels=paste(matches$team1,matches$team2,sep=" v "),las=2,cex.axis=0.65)
}

clean.data <- function(data) {
  require(zoo)
  colnames(data)[1] <- "Date.Time"
  colnames(data)[-1] <- gsub("^.*?\\d+_(\\w+)[.]href.*?$","\\1",colnames(data)[-1])
  data <- data[,-NCOL(data)]
  data$Date.Time <- as.Date(substring(data$Date.Time,2),"%Y-%m-%d")
  data <- data[is.na(data$Date.Time)==F,]
  data <- data[order(data$Date.Time),]
  for(i in colnames(data)[nchar(colnames(data))==2]) {
    data[,i] <- gsub("<div class=.*?>(.*?)</div>","\\1",data[,i])
    data[,i] <- gsub("<div class=.*?>(.*?)\\s+$","\\1",data[,i])
    data[,i] <- gsub(">","",data[,i])
    data[,i] <- gsub("^\\s+|\\s+$","",data[,i])
    #need to turn fractional odds into decimal odds
    numerator <- gsub("(\\d+)/(\\d+)$","\\1",data[,i])
    denominator <- gsub("(\\d+)/(\\d+)$","\\2",data[,i])
    data[,i] <- as.numeric(numerator)+1
    data[numerator!=denominator,i] <- as.numeric(numerator[numerator!=denominator])/as.numeric(denominator[numerator!=denominator])+1
    data[,i] <- na.locf(data[,i],na.rm=F)
  }
  return(data)
}

Premier League

simpleplot("English Premier")

The coloured dots are forecasts from the ordered logistic regression model; the black circles are the forecasts from a simple OLS linear probability model. Hence the black circles are essentially a probability of a home win occurring (given the ordinal variable defined to capture all three outcomes), whereas the red squares are the probability of a home win, the green solid circles are the probability of a draw, and the blue triangles the probability of an away win. The home bias in football is notable in that the majority of red squares lie above blue triangles.

bkplot("English Premier","premier-league")

## Loading required package: zoo
## 
## Attaching package: 'zoo'
## 
## The following objects are masked from 'package:base':
## 
##     as.Date, as.Date.numeric

The darker coloured and smaller symbols are the range of bookmaker implied probabilities for outcomes. Unlike in previous weeks where I’ve simply plotted the mean bookmaker implied probability, here all bookmaker prices (usually 20) are plotted. By and large, they are not particularly widely spread, but then this is only a small sample of the total bookmakers pricing these matches.

The only really notable difference is in the Man City West Brom match: my model fancies Man City much less than the bookmakers do. A West Brom win seems highly unlikely, but City have already had a pretty awful week.

In terms of making stronger picks than simply stating probabilities, as requested by some, here goes (nothing spectacular here):

Southampton to beat Burnley.
Tottenham to beat Leicester.
Chelsea to win at Hull.

Championship

Next, our Championship forecasts:

simpleplot("English Championship")

This week I’m adding in Championship bookmaker prices:

bkplot("English Championship","championship")

The big divergences here are:

Wolves vs Derby: My model fancies Wolves much more than the bookies.
Bournemouth vs Middlesbrough: My model fancies Bournemouth more than the bookies.
Blackburn vs Brighton: My model fancies Blackburn much more than the bookies.
Charlton vs Reading: My model fancies Charlton much more than the bookies.
Watford vs Ipswich: My model fancies Watford much more than the bookies.

League One

Next, our League One forecasts:

simpleplot("English League One")

League Two

Next, our League Two forecasts:

simpleplot("English League Two")

Football Conference

Next, our Football Conference forecasts:

simpleplot("Football Conference")

List of all forecasts

For transparency, all forecasts are also listed as a table:

kable(forecast.matches[order(forecast.matches$date,forecast.matches$division),
                       c("date","division","team1","outcome","team2","Ph","Pd","Pa")])

	date	division	team1	outcome	team2	Ph	Pd	Pa
1	2015-03-20	English Championship	Wolves	0.6683253	Derby	0.5426365	0.2530953	0.2042683
2	2015-03-20	Football Conference	Bristol R	0.7654480	Aldershot	0.6591852	0.2047695	0.1360453
62	2015-03-21	Conference North	Tamworth	0.6436249	Hednesford	0.5138240	0.2624663	0.2237097
69	2015-03-21	Conference North	Boston Utd	0.7728899	Gainsborough	0.6648599	0.2020484	0.1330917
61	2015-03-21	Conference South	Farnborough	0.4457166	Bath City	0.2865024	0.2821696	0.4313281
10	2015-03-21	English Championship	Bournemouth	0.6726339	Middlesbro	0.5517542	0.2498941	0.1983518
11	2015-03-21	English Championship	Huddersfield	0.5062176	Fulham	0.3522602	0.2887492	0.3589906
12	2015-03-21	English Championship	Rotherham	0.4903325	Sheff Wed	0.3325754	0.2880744	0.3793502
13	2015-03-21	English Championship	Blackburn	0.7432269	Brighton	0.6276684	0.2193098	0.1530219
14	2015-03-21	English Championship	Charlton	0.7096900	Reading	0.5960485	0.2328576	0.1710939
15	2015-03-21	English Championship	Watford	0.8121298	Ipswich	0.7130300	0.1777772	0.1091928
16	2015-03-21	English Championship	Cardiff	0.5540088	Birmingham	0.4070745	0.2856302	0.3072953
17	2015-03-21	English Championship	Wigan	0.6063754	Bolton	0.4682199	0.2747715	0.2570086
18	2015-03-21	English Championship	Norwich	0.6272093	Nottm Forest	0.5017616	0.2660354	0.2322030
19	2015-03-21	English Championship	Brentford	0.7936501	Millwall	0.7039875	0.1824871	0.1135254
20	2015-03-21	English Championship	Blackpool	0.2772274	Leeds	0.1581541	0.2233521	0.6184938
21	2015-03-21	English League One	Barnsley	0.4835494	Preston	0.3286547	0.2878151	0.3835302
22	2015-03-21	English League One	Sheff Utd	0.6336539	Port Vale	0.5027016	0.2657651	0.2315333
23	2015-03-21	English League One	Crawley	0.5117721	Leyton Orient	0.3545966	0.2887622	0.3566412
24	2015-03-21	English League One	Gillingham	0.7781465	Colchester	0.6708656	0.1991356	0.1299988
25	2015-03-21	English League One	Bradford	0.6184173	Fleetwood	0.4836325	0.2709814	0.2453860
26	2015-03-21	English League One	Crewe	0.5805303	Oldham	0.4357724	0.2814103	0.2828173
27	2015-03-21	English League One	Peterborough	0.6494798	Chesterfield	0.5181546	0.2611326	0.2207128
28	2015-03-21	English League One	Coventry	0.5039417	Doncaster	0.3496906	0.2887189	0.3615905
29	2015-03-21	English League One	Rochdale	0.6700638	Scunthorpe	0.5554547	0.2485638	0.1959815
30	2015-03-21	English League One	MK Dons	0.4405967	Notts Co	0.3024233	0.2849415	0.4126352
31	2015-03-21	English League Two	Plymouth	0.5288782	Newport Co	0.3791693	0.2880847	0.3327459
32	2015-03-21	English League Two	AFC W’bledon	0.4807830	Portsmouth	0.3222693	0.2873006	0.3904302
33	2015-03-21	English League Two	Shrewsbury	0.7374737	Oxford	0.6302573	0.2181531	0.1515896
34	2015-03-21	English League Two	Tranmere	0.3497458	Burton	0.2085670	0.2553165	0.5361165
35	2015-03-21	English League Two	Hartlepool	0.4465967	Mansfield	0.2869138	0.2822515	0.4308347
36	2015-03-21	English League Two	Cheltenham	0.4776438	Exeter	0.3214544	0.2872266	0.3913190
37	2015-03-21	English League Two	Southend	0.6899824	Cambridge U	0.5638601	0.2454773	0.1906627
38	2015-03-21	English League Two	Carlisle	0.6508966	Morecambe	0.5210340	0.2602309	0.2187351
39	2015-03-21	English League Two	Bury	0.6675887	Northampton	0.5383542	0.2545606	0.2070851
40	2015-03-21	English League Two	Accrington	0.4867336	York	0.3254142	0.2875684	0.3870174
41	2015-03-21	English League Two	Stevenage	0.6744996	Dag & Red	0.5518389	0.2498638	0.1982973
3	2015-03-21	English Premier	Stoke	0.6190870	C Palace	0.4816349	0.2714946	0.2468704
4	2015-03-21	English Premier	Newcastle	0.3005543	Arsenal	0.1726128	0.2339110	0.5934762
5	2015-03-21	English Premier	Southampton	0.7787170	Burnley	0.6850935	0.1921031	0.1228034
6	2015-03-21	English Premier	Tottenham	0.7821516	Leicester	0.6787133	0.1952794	0.1260073
7	2015-03-21	English Premier	Man City	0.6889980	West Brom	0.5866152	0.2366833	0.1767015
8	2015-03-21	English Premier	Aston Villa	0.5778841	Swansea	0.4336757	0.2817732	0.2845512
9	2015-03-21	English Premier	West Ham	0.6466131	Sunderland	0.5242069	0.2592235	0.2165696
42	2015-03-21	Football Conference	Wrexham	0.7011045	Lincoln	0.5837898	0.2378091	0.1784012
43	2015-03-21	Football Conference	Torquay	0.5958380	Kidderminster	0.4581215	0.2770380	0.2648405
44	2015-03-21	Football Conference	Barnet	0.8488533	Welling	0.7620009	0.1511356	0.0868635
45	2015-03-21	Football Conference	Macclesfield	0.7374843	Nuneaton	0.6427770	0.2124630	0.1447601
46	2015-03-21	Football Conference	Alfreton	0.5874616	Chester	0.4473761	0.2792534	0.2733704
47	2015-03-21	Football Conference	Dover	0.5408086	Gateshead	0.3928702	0.2870932	0.3200366
48	2015-03-21	Football Conference	Grimsby	0.6346850	Eastleigh	0.5028478	0.2657229	0.2314293
49	2015-03-21	Football Conference	Woking	0.5494215	Forest Green	0.4016516	0.2862396	0.3121088
50	2015-03-21	Football Conference	Southport	0.6849996	Dartford	0.5621175	0.2461245	0.1917580
51	2015-03-21	Football Conference	Altrincham	0.5151278	Halifax	0.3626124	0.2887024	0.3486852
106	2015-03-22	English Premier	QPR	0.4635192	Everton	0.3072164	0.2856196	0.4071640
107	2015-03-22	English Premier	Hull	0.2736013	Chelsea	0.1561006	0.2217537	0.6221458
108	2015-03-22	English Premier	Liverpool	0.6046988	Man Utd	0.4640892	0.2757198	0.2601910
109	2015-03-22	JP Trophy	Bristol C	0.8279363	Walsall	0.7368224	0.1650660	0.0981117
112	2015-03-24	English League One	Oldham	0.4423522	Rochdale	0.2853290	0.2819327	0.4327383
133	2015-03-24	English League One	Sheff Utd	0.6008226	Scunthorpe	0.4628322	0.2760026	0.2611652
113	2015-03-24	English League Two	Luton	0.4458798	Wycombe	0.2872629	0.2823206	0.4304164
114	2015-03-24	Football Conference	Braintree	0.7518244	Telford	0.6524652	0.2079521	0.1395827
115	2015-03-24	Football Conference	Alfreton	0.6881130	Dartford	0.5657038	0.2447883	0.1895079
116	2015-03-24	Football Conference	Dover	0.4362090	Grimsby	0.2798994	0.2807767	0.4393239
117	2015-03-24	Football Conference	Woking	0.7674044	Torquay	0.6620147	0.2034165	0.1345688
118	2015-03-24	Football Conference	Nuneaton	0.4497043	Wrexham	0.2938011	0.2835413	0.4226576
134	2015-03-24	Football Conference	Halifax	0.5234314	Gateshead	0.3704449	0.2884916	0.3410636

Forecasting Football Matches, March 20-25

J James Reade

20/03/2015