A League Of Gentlemen
A League Of Gentlemen
If you read my previous article on trying to predict results in the Russian Premier League, you would have been aware of the fact that ‘StatBot’ provided another round of predictions last weekend and one that I must say I wasn’t all to confident about.
That’s because of the eight games that were on offer, ‘StatBot’ predicted that as many as five were going to end up in draws. Which to be honest, even to the untrained eye, is arguably too many matches ending all square.
So before, we review the weekend’s action, let’s remind ourselves of the picks from last weekend:
As you can see five draws, 2 home wins and 1 away win were what we were hoping for – did we get anywhere close to that, unfortunately we were very far from the case and this has not any led to a second loss in five weeks but also some serious head scratching.
First lets take a look at the results that did at least come in:
FC Ufa vs Lokomotiv Moscow – Draw
FC Krasnodar vs FC Rostov – Draw
Just two out of eight and they returned £3.55, which means from the £4 I spent last week, I actually lost 45p. Overall this has also eaten into my profit and after five weeks of activity, I am down to 32p or 1.6% ROI.
Once again though, these were successful outcomes that did at least find an edge over the bookmakers, as the odds of each draw were more than 2.0 (greater than evens) so as before picking what are perceived as unlikely draws has got him out of a hole.
Although we can ‘celebrate’ that profit is profit, something is seriously amiss here as I ‘StatBot’ only returned two correct picks but how much of this is his fault and how much is down to the activity of the Russian Premier League?
Let Me Get My Excuses In
If we take a quick snapshot of the Russian Premier League table (places and points will do for this exercise)
|1||Zenit St. Petersburg||33|
|7||Krylya Sovetov Samara||18|
You will see that there is almost a clear division within the top five clubs and the other eleven and when you consider that just six points separate fifth to bottom in the Russian Premier League, it means that each of those clubs are capable of taking points off of the other.
Now this may be great for entertainment purposes, but it is becoming a bit of a nightmare from a predictive modelling point of view, for the simple reason that form as an indicator is becoming a pretty useless guide.
But without form what can we do, well I cannot necessarily abandon the concept of using form as a guide but what I can do is compare the home team’s last six home matches and the away teams last six away matches from this coming weekend.
At present the form index is looking at the last six games of each team regardless of home and away and I have a feeling that this is creating more difficulties than solution and after such a paltry return, it is now time to change tact.
Venting My Frustration
However, the most frustrating thing in all of this, is that there where actually five draws last weekend, it’s just my picks and the actual outcomes did not correlate. Had one more got over the line, then it would have been another week of profit.
Not only but, out of the eight matches there was not one home win, which means the bias applied to teams playing at home was absolutely worthless. However, do we put that down to a freak set of results or something more fundamental?
In summary: It has not been a particularly good week at the office and I now have to wonder, whether this modelling project has been based on one of the most difficult leagues in Europe. It certainly seems that way!
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.