Another Lucky Seven
Another Lucky Seven
If you read my previous article, you will know that my pursuit of data driven excellence was not all that far away and with ‘BundesBot’ getting so close to the finishing line, the anticipation for another round of Bundesliga bets is increasing.
Last weekend my trusted (at times) German supermodel returned seven correct match predictions from a possible nine and this is the second time in four gameweeks that this occurance has taken place.
Which means one of two things:
1) We’re onto something special here
2) We’re just incredibly lucky
With 36 matches predicted and 22 correct in total, it is perhaps a little too early to start proclaiming that this is an overwhelming success. However, a 61.1% hit rate, is certainly not to be sniffed at and with that in mind, it is time to go again.
THE LIE OF THE LAND
As you can see, the most appealing clash will be Bayer Leverkusen playing host to Bayern Munich and with the top five all winning last week (something that played a huge part in my overall success), that obviously cannot happen this tirme around.
Then again, its not as if I can be relying on the top five to win every weekend, as that is a risky strategy in terms of reliance and more importantly it only gets just over half the job done (should the top quintent all get over the line)
Another interesting clash will be Hertha Berlin away at Borussia Dortmund, they really are undergoing something of a new manager bounce right now and after picking up 10 points from the last 12 under new manager Bruno Labbadia, they are one of the form teams at present.
However, they go up against a Borussia Dortmumd side that made incredibly light work of Paderborn last Sunday and with Champions League qualification far from guaranteed, Lucien Favre’s men cannot slip up at Signal-Iduna Park.
Ok that’s enough of the preview, let’s see what ‘BundesBot’ has served up this weekend:
A SHOCK TO THE SYSTEM
Somewhat surprisingly, Hertha Berlin have been selected to beat Borussia Dortmund, something that I personally would not have opted for, but the model has spoken and in its own humble opinion, it is the team from the capital that will come out on top.
Another interesting pick is Union Berlin to draw with Schalke. David Wagner’s men have lost all four games since the restart and have not won in any of the last 11, this means some reversal of fortune is going to be required this weekend.
Apart form that game being the only draw selected out of the nine data driven football predictions on offer, it does seem as if the picks do seem rather logical. Although to get a better idea, lets look at the accumulative odds from Betfair.
WHAT ARE THE ODDS?
This week, it seems we are swimming against the tide as the odds of the nine-fold coming is a hefty 3,496/1. To give you some additional context, last week’s offering was a more palatable 585/1 instead.
Which means, yes should I somehow return nine correct data driven football predictions, I will win very handsomely – although with that said, trying to win will now be a lot harder than I had an anticipated.
For example, Hertha Berlin are 7/1 on their own to win against Dortmund, stick that into the acca and it acts a huge multiplier for any final odds. Has this unncessarily skewed proceedings, time will tell I guess.
In terms of the ‘BundesBot’ model itself, once again I’ve left it unchanged – there’s very much a sense of if keeps falling just short, why bother about changing in it. A logic that can go in two directions:
1) We stay on the correct path, and return another impressive set of results
2) Failure to change, gives us another horrific slump (where we got just 2 out of 9 correct)
TIGHTENING THE BELT
For me personally, I always attempt to get a clean sweep of results no matter what league I’m in and although I haven’t got there yet. Part of the fun of this project, is attempting to chase the somewhat near impossible dream.
It seems that in terms of trying to find the perfect accumulative value – somewhere in the region of 500/1 seems like the most realistic chance of winning. Of course, just because you may hit those odds, doesn’t guarantee you overwhelming success.
At the same time, with this weekend’s example of nearly 4,000/1 – it’s as if there is too much work against the grain required and although placing the bet at least gives you a chance, there is a sense of not a lot of hope in this one.
But where does this all lead to? Well quite simply, what if I could drop three ‘troublesome’ fixtures a week and slim a six-fold down to say 100/1. Something that would give me a far greater ROI potential and could easily be repeated against multiple weeks.
That way, there would be less boom and bust and more constant wins. Sounds easy doesn’t it. Here, the trick is, trying to work out a ranking that would scale the probability of each of the picks and therefore knock out the problem predictions.
That seemingly is the evolution of all of this, the sticking point is that I don’t know quite how to work out the probability of an occurance (I can with goals/clean sheets) but I’m not sure if it would be the same logic – still I guess it cannot hurt to try (might attempt more in the Premier League)
TAKE A DOUBLE CHANCE
While once again, it also time to have a look at the double chance market – of the last 27 data driven football predictions in this market, ‘BundesBot’ has predicted 22 correctly and last week returned eight from a possible nine.
With that in mind, here is this weeks list:
Borussia Monchengladbach and Draw
Hoffenheim and Draw
RB Leipzig and Draw
Bayern Munich and Draw
Mainz 05 and Draw
Hertha Berlin and Draw
Wolfsburg and Draw
Draw or Union Berlin
FC Koln or Draw
Much more insurance and only odds of 11/1 , but the hit rate has been very good in the small sample (81.48%) so no reason why this cannot reach the final step over the course of another Bundesliga weekend. That’s your lot for this week, enjoy another healthy dose of football.
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.