The Twelfth Night
The Twelfth Night
If you read my previous article, you will know that another weekly mop up has taken place and with another raft of Premier League action just around the corner, the pursuit for data driven betting excellence continues.
Before we dive into another week of data driven football predictions, lets first take a quick recap of how performance has been to date:
Pure Gut Instinct: 56/107 (52.33%)
PremBot 1: 19/39 (48.71%)
PremBot 2: 15/29 (51.72%)
Hybrid Model (Combining First 68 matches and then PremBot only): 57/107 (53.27%)
In a word – average. Not a shambles, but nothing to crow home about either, It’s as if we are lacking a little bit of edge. However, not to be deterred, it is time to go again and after a respectable performance last time out, it is time to trust my gut once more.
With 6 out of 9 get over the line last weekend, it is time to see if we can at least hit for six once more and to do so, here are first my random gut instinct picks:
Wolves/Aston Villa Draw
It’s fair to say, that this is a set of rather “safe” picks and this is reflected in the 10-fold odds, which are 535/1 and without looking just yet, I would imagine that bar the draw at Molineux, everything is just following the bookmakers own favourites.
So IF and that’s always a big IF, things go to plan, it will not be the most hefty of rewards but it would be welcomed all the same. While if this considered to be the bread and butter, now its time to delve into the main course.
THE DATA DISH
As always the crux of this project, is the data driven football predictions and before we see what “PremBot” and his younger brother “PremBot JR” have served up for this week, lets first take a look at the form guide which underpins it.
For those who have not got involved in the past:
Orange – The Hottest Form In The Division
Blue – The Coldest
Green – Best Per Game
Which means, with those stats being presented, here is what “PremBot” has to offer last week – 5 out of 9 was the performance in Week 11, what can be done this time around?
A little bit off the beaten curve and four different selections compared to my gut instinct, so its good that there’s the ability to compare and contrast. While if all of these 10 were to get over the line, the odds would be 3454/1. Not as outrageous as some previous offering, but chunky all the same.
Of course, with “PremBot Jr” also in the mix, it is time to see what his picks have served up.
SPOT THE DIFFERENCE
Now, here’s an interesting point and the fact, that all the outcomes are the same. Which does not offer us the ability to properly compare and contrast. Although we can ask as to why exactly this has happened.
The simple reason is that even with a lessening of league position comparison in the second model, it does nothing to change the overall outcomes and therefore, instead of fully backing the model again, I am going to place just have the usual stake but still run all 10 matches as normal.
WHERE’S THE MONEY
Which means, what to do with the other half a stake unit and with this going spare, I am going to use the original form guide as a data driven prediction tool, which means there is another 10-fold for fun which looks like this:
If this one comes in, it is odds of 2,590/1 – lying somewhere in between the two predictive models and the only difference here is Burnley getting the better of the Emirates on Sunday night.
A LOOK AT THE BOOK
I refefenced earlier that my gut instinct picks are likely to be in tune with what the bookmakers are saying and lets now see if that actually is the case or not:
|Tomorrow, 11 Dec||1||X||2||B’s|
|20:00||Leeds – West Ham||5/4||67/25||211/100||14|
|12 Dec 2020||1||X||2||B’s|
|12:30||Wolves – Aston Villa||33/25||229/100||23/10||14|
|15:00||Newcastle – West Brom||61/50||119/50||62/25||13|
|17:30||Manchester Utd – Manchester City||67/20||78/25||77/100||14|
|20:00||Everton – Chelsea||81/25||153/50||4/5||14|
|13 Dec 2020||1||X||2||B’s|
|12:00||Southampton – Sheffield Utd||3/4||279/100||197/50||14|
|14:15||Crystal Palace – Tottenham||93/25||14/5||39/50||14|
|16:30||Fulham – Liverpool||381/50||237/50||17/50||14|
|19:15||Arsenal – Burnley||14/25||321/100||269/50||14|
|19:15||Leicester – Brighton||26/25||63/25||139/50||14|
|15 Dec 2020||1||X||2||B’s|
|18:00||Wolves – Chelsea||96/25||139/50||3/4||12|
|20:00||Manchester City – West Brom||11/100||181/20||2333/100||12|
And as first thought, apart from the draw at Molineux which is different, I have picked the pre-match favourite in the other nine matches. However, we know that such a run of form following results never really happenes and therefore, I could be on something of a hiding to nothing.
To provide a bit more context, lets look at the 1×2 split in terms of Premier League results:
As you can see, wins on the road are still the dominant force in all of this and with the data driven predictive models opting for four more this week, that is prettu much in line with what the overall breakdown is saying.
A breakdown that give or take is like this:
4 Home Wins
4 Away Wins
Whereas my models are saying:
5 Home Wins
4 Away Wins
So we will see if the margin of error here is something that works in our favour or not and we’ll only know if it works by Monday as then its another round of fixtures to follow, so things are going to get very busy now.
Happy punting and thanks for reading – to discuss this or any other projects you have, feel free to message me at either @dantracey1983 on Twitter or email me: firstname.lastname@example.org
(THESE ARE NOT TIPS PER SE, THESE ARE JUST DATA DRIVEN FOOTBALL PREDICTIONS IN A PURE TEST AND LEARN MODELLING CAPACITY)