Trying Our Luck

Trying Our Luck

Trying Our Luck

If you read my previous article you will know that my previous attempt at data driven betting excellence, was nothing short of absolutely woeful. However, undeterred by this or perhaps a glutton for punishment we are going to try our luck once again.
Luck that was most certainly out last week, as my main model ‘BelaBot’ returned just two correct data driven football predictions out of eight – a 25% hit rate and not something we can necessarily hang are hat on.
To make matters worse, this was a lesser return when using gut-instinct and no data, in the week prior – something that has made an early mockery of proceedings and now I need to somehow get back in the groove.
WHAT IS ON THE MENU
As before, there are eight Belarus Premier League fixtures taking place this weekend and once again I have created a form infograph to highlight which teams are in better shape going into these next few days, an infograph that looks as follows:
Trying Our Luck
HOME IS WHERE THE HEART IS
It does seem to be there is more strength in the home teams and hopefully that ties into the build of the model. A build that even after such disater last week, I’ve decided to keep the fundamental principles the same.
The reason being, I need a control measure and with the control measure bar being set so low, there is thankfully plenty of room for improvement and it is improvement, I hope that happens over the course of this weekend.
With that in mind, here are the next set of picks from ‘BelaBet’ – remember these are not ‘tips’ these are data driven football predictions, it’s more test and learn, rather “bashing the bookies’ for use of a better term:
Trying Our Luck
Only one draw has been predicted, which although risky territory, is not necessarily bad news either – ultimately there’s not a lot else that can be done and the analysis will take place on the other side of the weekend. Although with that said, we’re not done yet.
THREE WAS NOT THE MAGIC NUMBER
That’s because if you did read my previous article, you will know that I introudced a model that looked at form over the last three matches, than opposed to the last six and it meant there was a variant that could be compared against, to see if one train of thought was better than the other.
However, if the standard edition was bad, the smaller one was absolutely rotten – just one correct pick (Thanks BATE Borisov) from eight and although an effort like that is good for the bin, I shall peservere regardless.
Using the same approach in that the control bar has been set as low as it can possibly be, I’m going to leave the functionality unchanged this week, to see if was just an early disaster or fundamental changes need to be made going forward.
With that in mind, here are the picks using the last three matches as an indicator of form:
Trying Our Luck
Firstly, it has thrown up some different results again – which is good (although not when they both end up wrong) and aslo, a model with no draws whatsoever. I think this is due to the weighting of predictions from home vs away – I have a feeling this may need to recalibrated rather quickly.
Just to make things interesting, I’ve put a very small amount on each of the 8-folds, the odds are rather substantial but the outlay is miniscule. Therefore, even if I do win the lot, I’ll still be playing again next week.
Anyway, that’s enough of the psuedo analysis, now its time to watch the action (if you are that way inclined). Bit by bit, football is returning and next week, I will also focus on the return of Die Bundesliga.
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 [email protected]. 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.

Follow me on Facebook at Dan The Stat Man
You can also check out my Premier League Podcast on Soundcloud

 


Leave a Reply

Your email address will not be published. Required fields are marked *