If you read my previous article on trying to predict the outcomes in the Russian Premier League, you will be fully aware that I made some fundamental to the predictive model which is otherwise known as ‘StatBot’
While if you are also a regular to the series, you will know that at this time of the week, is the one where we review just how ‘StatBot’ performed over the course of the weekend and that means, it is time to reveal whether I made a profit or a loss over the past week.
Before we do that though, lets have a look at the predictions from last weekend and a quick word from our sponsor: Join FanTeam and enter daily fantasy sports competitions.
And now I can reveal which picks were correct:
FK Akhmat vs Ural – Draw
Arsenal Tula vs Zenit St.Petersburg – Away Win
PFC Sochi vs CSKA Moscow – Away Win
Lokomotiv Moscow vs FC Krasnodar – Draw
As you can see I got off to the perfect start, as the first game ended in a draw and this means my work was almost done. However, the bad stretch across Saturday meant there was some real danger of it being washout. Thankfully though, the last three matches all came in and profit was ensured.
Also, I should highlight the fact that FC Krasnodar have now drawn four league matches in a row. A string of outcomes, that has been a real profit booster for me – especially as this has been more of an ‘edge’ over the bookmakers, so in this aspect alone the model has a degree of success.
Which means after ‘StatBot’ improved on his efforts on last week and returned four correct picks, I actually returned a profit. From the 8 units I outlayed (50p per unit = £4), I recouped just over 10 units, which also means this is the fourth week in six in which profit has returned.
More importantly, what does that do for the overall P/L?
From the £24 that has been spent thus far, ‘StatBot’ has returned £25.08 and that means a ROI of 5.83% and 66.6% of the weeks have been profitable, with only two out four busts (for use of a better of term)
Therefore, we can say that the changes in comparing pure home form versus pure away form looks a more successful path. However, one swallow does not make a Summer and this means that the next approach has to be to test the same method again after the international break.
There might be a couple of more tweaks under the bonnet, if only because I feel the model is still a little too rigid in places and therefore this overriding ‘one size fits all’ concept, is something that will need to evolve.
Any thing to consider, is to add more elements of information into the background modelling – factors such as league position comparison and historical head to head could (and probably should) also be added to the equation this side of Christmas,
On the one hand, the international break is a little bit annoying as there is no live betting to do. Although on the other, it will allow me the perfect amount of time, to go back and add some more elements to the model.
Which means, for the next week or so, I will be looking at adding extra elements to the model and moving a step closer to the pursuit of riches!
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.