Ups And Downs
Ups And Downs
If you read my previous article in regard to trying to predict results in the Russian Premier League, you will know that I had a sneaky feeling that the weekend just passed was not to be a particularly good one.
It turns out that gut feeling was correct and not only that with three weeks of predictive attempts now in the history books, it means I’ve had to suffer my first loss and one that has nearly wiped out all the hard work from the first two weeks.
Week 3 saw me outlay the usual 8 units (£4), however I only managed to return £2.83 – a loss of £1.17 for the week (nearly 2.50 units in total) and the upshot of that, is that I am only on 23p profit overall (just under 0.5 units)
If we try and put a positive spin on this,I am always going to have up and downs each week, but on a long-term basis, I cannot really absorb such busts as this, especially when only getting two out of eight match picks correct.
Therefore we can say that ‘StatBot’ has not has a good week and changes will be afoot for the following week. However, there is always room for a caveat or two and this week’s ‘excuse’ is that the Russian Premier League results continue to be so volatile.
First lets recap what last week’s picks were:
Firstly a huge thanks to both Lokomotiv Moscow for winning away from home and the pairing of FC Ufa and Spartak Moscow for sharing the points otherwise last week would have been a complete bust and more importantly ‘StatBot’ would have been launched into the theoretical bin.
But more importantly if we go back to my point regarding volatility, the week before last there were no draws in any of the eight games and this week there was only one solitary home win, so it’s hard to get an idea of where we should be pitching our logic in any given week.
Not only but had FC Krasnodar or PFC Sochi won on the Sunday, I would have turned the week around into a small profit and ultimately added to the profit base from the two previous efforts – in this early stage this is the fine margins in which we are dealing with.
In addition to that FC Krasnodar were held by a much more inferior side, in a game that by all accounts they should have won, even though they were away from home. Then again, should home bias have been more of a factor here?
Still I can’t play the ‘if only” card as ultimately ‘StatBot’ needs to be a lean predictive betting machine and now that I’ve suffered a loss, I will be making some development tweaks for the weekend’s round of matches.
Although a losing week is a difficult pill to swallow, it does allow to highlight the deficiencies quicker and as long as we can steady the ship and return to calmer waters this weekend, then we will be able to but this small blip behind us.
With that in mind, here are the action points for this week:
- Introduce home bias to the model
- Change the coding for the form index and the basis of how results are picked
- Look at home form (last six) vs away form (last six) and not just overall last six form
I don’t want to make all these changes at once, because then we’ll lose track of what works best and what doesn’t. That said, I will tinker away over the next couple of days and come Friday, I will not only have my weekend picks ready but explain the changes that I’ve made.
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.