We Go Again
We Go Again
If you read my previous article you will be fully aware that this season’s pursuit of data driven betting excellence has finally begun and after an average showing last week, it is time to double our output in terms of second guessing the Premier League
DOUBLE TROUBLE
Because this week, I am going to be working with not one but two models and in doing so, it will hopefully make the task of making data driven football predictions a far sight easier than it was before.
To take stock of where, we are currently at in terms of performance, here is the current lie of the land:
Pure Gut Instinct: 43/78 (55.1%)
PremBot: 5/10 (50%)
Hybrid Model (Combining First 68 matches and then PremBot only): 43/78 (55.1%)
So as you can see, it is mid-table showing across the board and once again, I am first going to trust my gut instinct and this week, this is what I think will happen without looking anywhere near a spreadsheet:
Chelsea
Aston Villa
Tottenham/Manchester City Draw
Manchester United
Everton
West Ham
Leeds/Arsenal Draw
Leicester
Crystal Palace
Wolves/Southampton Draw
Three draws and an away win for Leicester has meant, this could well be the biggest odds I’ve ever been up against – 8,332/1. An eye-watering amount, I am sure you will agree but of course it all counts for nothing if all 10 don’t get over the line.
PREMBOT AND HIS YOUNGER BROTHER
Ok so with the gut instinct out of the way, it is now time to look at what “PremBot” has served up in terms of his data driven football predictions and even though it only returned 5 correct picks out of 10 before the international break, it is a case of same again this time
But before we look at the picks, lets first take a look at the form guide that drives them:
The best form in the division is Orange
The worst form is in Blue
The best per weekend meeting is Green
So armed with that information and the piece of code that has gone behind it, lets see what “PremBot” has given us for his data driven football predictions this week:
Well if you thought the previous odds were big, this set is even bigger – 12,077/1 which means there must be some against the grain picks in all of this. Then again, knowing what we know about the bookmakers this season, they are far from on the money all the time themselves.
Therefore, we could try and second guess them but at the same time it is something of a fruitless pursuit in itself and with those odds and betting slip, now locked in ready for the weekend – it is time to double up.
AWAY FAVOURITES
Taking the existing model that is in place and using it as above, I’ve decided to make a variation and run the two side by side. With the difference in this one being that there is more weight applied to teams away from home.
In doing so, “PremBot 2.0” has served up this raft of data driven football predictions instead:
Whereas the was one solitary draw in the offering above, there are none in this one and this does offer something of a risky strategy. Then again, without any shared points boosting the odds, the coupon is not as weighty.
With that said, it is 7664/1 and I don’t think you or I would turn our collective noses at a win of that size. Now usually, the odds that are generated from such a raft of data driven predictions is never that big, which means we must be betting against the grain this weekend.
What is also interesting is that the second model has only provided one different result to the first and that is Leicester to win at Liverpool. So even with more away weight added, there is a lot of similarity between the two.
While my gut instinct picks have all picked the same winners, but have opted for three draws instead – so again we are certainly thinking along the same path. But how does that compare to the bookmakers, lets now have a look:
| 21 Nov 2020 | 1 | X | 2 | B’s | ||
|---|---|---|---|---|---|---|
| 12:30 | Newcastle – Chelsea | 553/100 | 71/20 | 51/100 | 13 | |
| 15:00 | Aston Villa – Brighton | 31/25 | 131/50 | 221/100 | 13 | |
| 17:30 | Tottenham – Manchester City | 159/50 | 153/50 | 41/50 | 13 | |
| 20:00 | Manchester Utd – West Brom | 8/25 | 114/25 | 907/100 | 13 | |
| 22 Nov 2020 | 1 | X | 2 | B’s | ||
| 12:00 | Fulham – Everton | 59/20 | 141/50 | 23/25 | 13 | |
| 14:00 | Sheffield Utd – West Ham | 41/20 | 231/100 | 73/50 | 13 | |
| 16:30 | Leeds – Arsenal | 223/100 | 269/100 | 6/5 | 13 | |
| 19:15 | Liverpool – Leicester | 19/20 | 72/25 | 277/100 | 13 | |
| 23 Nov 2020 | 1 | X | 2 | B’s | ||
| 17:30 | Burnley – Crystal Palace | 89/50 | 21/10 | 46/25 | 13 | |
| 20:00 | Wolves – Southampton | 133/100 | 229/100 | 231/100 | 13 | |
Here there is quite a difference, as the bookmakers have opted for the following differences:
Manchester City Win
Arsenal Win
Liverpool Win
Burnley Win
Wolves Win
So no fewer than five differences of opinion and it is for that reason, that all three sets of odds that have been generated are as big as they are. Which means, ultimately we want another week where things go against the grain.
Another way we can look at this, is the result split of the season so far:
| HW | 29 | 37.18% |
| AW | 34 | 43.59% |
| DRAW | 15 | 19.23% |
| TOTAL | 78 | 100.00% |
As you can see, there are still more away wins than there are home and my models have opted for 5 and 6 respectively and using these numbers, there may be scope to generate a further model in terms of the split in results thus far.
Anyway that’s your lot for this article, as we look ahead to another raft of Premier League fixtures, will there be cause for celebration come Monday evening. I don’t know about you, but I sincerely hope there is.
Happy punting and thanks for reading.
(THESE ARE NOT TIPS PER SE, THESE ARE JUST DATA DRIVEN FOOTBALL PREDICTIONS IN A PURE TEST AND LEARN MODELLING CAPACITY)