In this article we will be describing the well known Win Draw Loss method.
Here are the basic rules…
The Win, Draw, Loss method is very simple and takes into account the proportion of past wins, draws, and losses over a specified period of time. By calculating the number of wins, draws and losses for both the home and away sides and then combining them it is possible to produce a statistical probability of the possible outcome of the game.
Here is how a typical calculation is made…
1. For a specified period of time the number of wins, losses, and draws are counted for each side. This could be for a whole season or you may wish to have a rolling time period, say the last six months for example.
2. The percentage probabilities for home wins, draws, and away wins are then made. For example, the home win percentage is calculated as follows:
(( HW + AL) * 100) / total number of matches
Where,
HW = number of home wins by home team
AL = number of away losses by away team
Once you have a percentage probability for each outcome a decision can be made on the most likely outcome.
Let’s look at a simple example of how you could apply this…
First of all you would collect the data for a specified period of time and for each of the three possible outcomes, i.e.
HW = number of home wins for home team
HD = number of draws for home team
HL = number of home losses for home team
AW = number of away wins for away team
AD = number of draws for away team
AL = number of away losses for away team
TOTALGAMES = HW + HD + HL + AW + AD + AL
HOMEWIN = ((HW + AL)*100)/TOTALGAMES
DRAW = ((HD+AD)*100)/TOTALGAMES
AWAYWIN = ((HL + AW)*100)/TOTALGAMES
This gives you a percentage probability for each of the three possible outcomes.
Now let’s extend this to provide 1, X, 2, 1X, and X2 predictions…
To do this we can employ a simple voting system, here is how it works:
First we set three thresholds;
DRAWTHRESHOLD = 35%
AWAYTHRESHOLD = 45%
HOMETHRESHOLD = 55%
Then we initialise our voting counts for each method;
DRAWP = 0
HOMEP = 0
AWAYP = 0
Now we compare each of the calculated probabilities against our thresholds to arrive at a vote count for each outcome;
IF DRAW > DRAWTHRESHOLD THEN
DRAWP = DRAWP + 3
ELSE
HOMEP = HOMEP + 1
AWAYP = AWAYP + 1
END IF
IF AWAYWIN > AWAYTHRESHOLD THEN
AWAYP = AWAYP + 3
ELSE
HOMEP = HOMEP + 1
DRAWP = DRAWP + 1
END IF
IF HOMEWIN > HOMETHRESHOLD THEN
HOMEP = HOMEP + 3
ELSE
DRAWP = DRAWP + 1
AWAYP = AWAYP + 1
END IF
This means we now have a total vote count for each of the three possible outcomes.
Next we can deduce our final 1, X, 2, 1X, and X2 predictions based on the voting as follows;
IF HOMEP > AWAYP AND HOMEP > DRAWP THEN
PREDICTION = “1”
ELSEIF AWAYP > HOMEP AND AWAYP > DRAWP THEN
PREDICTION = “2”
ELSEIF DRAWP > HOMEP AND DRAWP > AWAYP THEN
PREDICTION = “3”
ELSEIF HOMEP = DRAWP AND HOMEP > AWAYP THEN
PREDICTION = “1X”
ELSEIF AWAYP = DRAWP AND AWAYP > HOMEP THEN
PREDICTION = “X2”
ELSEIF HOMEP = DRAWP AND HOMEP = AWAYP THEN
PREDICTION = “X”
Now it’s your turn…
Of course you may choose to use different values to those shown above and by experimenting you may come up with better values to use. So, you could create better threshold values and decide to collect data over any period of time you like. Experimenting with threshold values and durations will let you home in on the best settings for you.