Scottish Data Theory
Scottish Data Theory
In my previous article I discussed how assuming that certain clubs will matches, will then enable you to get closer to picking a winning combination and by using a data driven betting approach it allows you to get closer to turning a profit.
The league I tested the theory on was the Scottish Premiership and although I was unfortunate enough to get five out of six picks correct, it did at least lend itself to an idea that is not a million miles away from success.
In the aforementioned previous article, I looked at it more from a theory point of view and now I thought I would highlight more on the actual background work that was required in order to place the bets.
As mentioned before the aim here is to try and pick a winning six-fold and if were to do that in the Scottish Premiership there are 729 different combinations which you could choose. However, if you were to only pick Rangers and Celtic as both winners, that dramatically drops to 81.
An 81 in question that looked something like this in the midweek set of fixtures:
|Game 1||Game 2||Game 3||Game 4||Payout||Status|
|Ross County||St Johnstone||Hibernian||Aberdeen||£55.49||IGNORE|
|Ross County||St Johnstone||Hibernian||Draw||£94.33||BET|
|Ross County||St Johnstone||Hibernian||Motherwell||£108.21||BET|
|Ross County||St Johnstone||Draw||Aberdeen||£169.80||BET|
|Ross County||St Johnstone||Draw||Draw||£288.66||BET|
|Ross County||St Johnstone||Draw||Motherwell||£331.11||BET|
|Ross County||St Johnstone||Hamilton||Aberdeen||£264.13||BET|
|Ross County||St Johnstone||Hamilton||Draw||£449.02||BET|
|Ross County||St Johnstone||Hamilton||Motherwell||£515.06||BET|
The above table has all of the individual and unique midweek permutations, the cash return after placing £1 each bet and whether the winnings return the necessary threshold to return a profit. If they do it returns ‘BET’ if they don’t it returns ‘IGNORE’
The winning combination was:
Draw – Livingston – Hibernian – Motherwell and this was at odds of 51/1
Of course I could have bet on every individual line to get a winner but that would not guarantee a profit, therefore I was only left 56 combinations that cleared the threshold, meaning that there were 25 traps so to speak.
If you were to list the above table in cash order, the winning combination would have ranked 67th from longest to shortest odds which means it was the 15th most likely winning combination in midweek.
Unfortunaely for me in this case, it seemed as if the results went to form and this meant that there was no shock of sorts. Although with this being a league in which outside the top two, anyone can take points off of anyone, it doesn’t mean the idea is a total bust either.
Maybe there needs to be an extra layer of filtering – will 4 draws or 4 away wins ever come good, do we really need to bet on them. Perhaps it makes more sense to bet from shortest to longest and ignore the hail mary bets?
Ultimately some of the 81 lines need to be dropped each week, the key will be which ones would make the most sense to leave behind and that is something that I am going to try and map out in a future article, so keep your eyes on the main page for that.
Happy punting and thanks for reading. Dan
If this has grabbed your interest and you would like to discuss/feedback then please feel free to drop me a message at firstname.lastname@example.org. While I am always looking for new football/data projects to work on and if you feel that my skills would be of use, I can be contacted at the same address.