The Bundes Bets
The Bundes Bets
If you read my previous article, you will know that my pursuit of data driven betting excellence was nothing short of a disaster in midweek, as the Bundesliga threw up some strange results – ok not strange as such, but not the results I wanted.
Whether that is down to home advantage being a thing of the past or just poor data modelling can be argued, whichever the answer turns out to be, we can only find out by once again trying our luck in Germany’s top flight.
Before I serve up another round of data driven football predictions, here is what the form guides look like for the past six matches (home last six and away last six)
It comes as no surrpise that Bayern Munich have the hottest form of all the home teams this weekend and after their important 1-0 win over Borussia Dortmund on Tuesday, they have gone a huge step clearer to an eight straight German league championship.
With the Bavarians being seven points clear at the top, they cannot afford to take their foot off of the pedal and with Fortuna Duesseldorf the visitors on Saturday, this has all the hallmarks of a home banker.
Then again, they could be famous last words, as the Bundesliga seems to have an aversion to home wins at the moment and with just 5 coming from the first 27 matches since the restart it equates to just 18.5% of all meetings.
Obviously the sample size is rather small, so we cannot confirm this as a new order in German football, but at the same time it does offer us a rather interesting trend to look at and one that may carry over to the Premier League when it restarts next month.
While although Borussia Dortmund’s title hopes have taken a huge dent, Lucien Favre’s men will know that its not over until its over and with bottom of the table Paderborn offering the opposition, this should be another win that is recorded.
Now that the picture has been set, its time to serve up the latest raft of data driven football predictions from ‘BundesBot’
A TEST OF STRENGTH
Admittedly in my previous article I said that I would make fundamental changes to the way the model is built, but at the same time the more I look at the way these predictions have landed, the more I think its worth at least one more week.
In theory, the predictive model should not perform badly two weeks in a row, there should be an element of it swinging back into life this time around (although whether this is wishful thinking remains to be seen)
Although with that said and even with the crazy results post-hiatus aside. I’d be confident of Bayer Leverkusen, Bayern Munich, Borussia Monchengladbach, Borussia Dortmund and RB Leipzig all getting over the line this weekend.
Should that be the case, that would be a respectable five out of nine. Which would at least salvage some pride from midweek, then again all that highlights is that someone with some Bundesliga knowledge can pick or winner five.
Something that you reading this article could do and therefore doesn’t really test the strength of a predictive model, because you or I are now just plucking results out of thin air and hoping for the best.
A WEIGHTY BIAS
I’ve often referenced that having pre-conceived bias can be a difficult thing to deal with, because it is those biases that have led me to believe, that I can get at least five outcomes correct – although I am making two huge assumptions here:
1) The top five all win
2) Teams actually win at home
Therefore, it seems like this pre-conceived bias of knowing that big teams will win regardless is a difficult path to walk down. However, it’s something that I am prepared to stick with for this week, in a bid to get a better idea of how successful this model is.
WHAT ARE THE ODDS
With that quintet of teams being named and them all being favourites, the nine-fold acca is not as ludicrous as it was in midweek. This time, it is a more managable and slightly more acheivable return of 585/1.
The split of backing five favourites and four returns that have gone against the grain is not untoward and when you consider that last week’s nine-fold was 38,000 odd to 1 – there does seem to be a correlation between the size of the acca and the eventual return.
Which means, we don’t necessarily want to shoot for the moon, because the evidence highlights that we’ll only crash down very hard. In theory, this means that the smaller odds, the better chance of hitting all nine correct data driven football predictions.
TAKE A DOUBLE CHANCE
With ‘BundesBot’ returning 13 out of 27 (48.1%) in these past three gamweeks, the success has not been great and the model is not really beating the man in the street. However, the double chance market has shown signs of life.
From two gameweeks (18 matches) 12 correct predictions have come good and I’m operating on a two-thirds success rate. Obviously the risk is far less, but it does at least mean we can get closer to the promised land.
Which means, my double chance picks are:
Bayer Leverkusen and Draw
Werder Bremen and Draw
Draw and Wolfsburg
Draw and Hertha Berlin
Hoffenheim and Draw
Bayern Munich and Fortuna Duesseldorf (wouldn’t let me go Bayern and Draw)
Borussia Monchengladbach and Draw
Borussia Dortmund and Draw
RB Leipzig and Draw
Can we get at least two thirds of these correct and keep the impressive streak going, there certainly does seem like there is a good level of insurance here and hopefully this will help me in due course over the weekend.
THE WRONG CHOICE
Also to make things even more interesting, I’ve done a separate double chance that picks the opposite two markets to what ‘BundesBot’ originally serves up. Just to look at things at a different angle, although to be honest we don’t really want success on this one, as it means the other has failed.
Anyway that’s your lot for this week, as the action once again comes round thick and fast. I’ll be back at the start of the next one, to dissect the performance of ‘BundesBot’ and my latest raft of data driven football predictions.
Happy punting and thanks for reading. Dan
(THESE ARE NOT TIPS PER SE, THESE ARE JUST DATA DRIVEN FOOTBALL PREDICTIONS IN A PURE TEST AND LEARN MODELLING CAPACITY)
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.