Reaching Double Figures

Reaching Double Figures

If you read my previous article, you will know that data driven betting excellence has once again been the order of the day and with week 10 of the new Premier League season (at what point do we stop calling it new?) just around the corner, it is time to test our mettle once again.

A QUICK RECAP

Avid readers of the series will now that I’ve also been testing my gut instinct across the season so far and with that and data driven football predictions now in the mix, it allows us to compare and contrast at such an early stage – while the table looks a little like this:

Pure Gut Instinct: 48/88 (54.54%)
PremBot 1: 11/20 (55%)
PremBot 2: 6/10 (60%)
Hybrid Model (Combining First 68 matches and then PremBot only): 49/88 (55.68%)

As you can see, its a reasonable return from yours truly and a decent enough start from “PremBot” and his younger brother “PremBot Jr” – however, now is not the time to rest on our laurels, as now its time to serve up another round of picks that are purely made on gut instinct alone.

A set of 10 predictions that looks as follows:

Crystal Palace
Liverpool
Manchester City
Everton
West Brom/Sheffield United Draw
Southampton/Manchester United Draw
Tottenham
Arsenal
Leicester
West Ham/Aston Villa Draw

Now with my batting average being 5 or 6 ten correct per week, I’ve not been pulling up the biggest of trees and when you think about it, some of these predictions almost pick themselves. However, I’ll take credit wherever I can get it.

While talking of credit, should all 10 of these get over the line – it will turn 1 currency unit into 3,570. So far more in tune with the bookmakers than it was last week.

GET THE MODELS OUT

Now it is time to move on to the data driven element and before I reveal what the two models are saying this week, lets first take a look at the form guides:

While for those who do not know the colour key for the central columns:

Orange – Hottest Form In The Division
Blue – Coldest Form In The Division
Green – Best Form In That Pairing

Also in another addition, if we used that colour key as our prompt for the bets – it would offer odds of 5,248/1 – so this might be another way to look at matters going forward and see if form is the ultimate indicator of future outcomes – but we’ll leave that for this week (please don’t get over the line)

Although that does at least provide us with four betting options going forward:

PremBot 1 – As Standard Formula
PremBot 2 – More Away Weight
PremBot 3 – As PremBot 1 But Last 6 Home vs Last 6 Away (in a few weeks)
PremBot 4 – The Pure Form Guide
PremBot 5 – To Be Considered

With the lie of the land now set, it is time to see what PremBot 1 has got in his locker this weekend:

A combination of results that is 5,599/1 in odds – so still incredibly big but not as big as the week before. Then again, it probably makes sense to keeping looking at each individual week in isolation for now.

While if we look at PremBot Jr – here is what he is serving up this weekend:

A slightly longer 6,380/1 and once again there is not a great deal of difference, as 8 out of 10 matches have come to the same conclusion. With the additional away weight leaning on victories for both Tottenham and Wolves instead of draw’s in the model above.

HOME OR AWAY

Another thing to factor in, is the current split of results after 88 games played and it is a split that looks as follows:

HW	33	37.50%
AW	38	43.18%
DRAW	17	19.32%
TOTAL	88	100.00%

As you can see, that early away advantage that everyone was proclaiming is not quite as apparent and with some supporters allowed in stadia from next weekend, this may only prove to tip the balance further. Which means that additional away weight, may prove to be heavy in the future.

While another check is to see, how in tune the data driven football predictions are with the bookmakers pre match odds and this weekend, they look as follows:

Today, 27 Nov			1	X	2	B’s
20:00	Crystal Palace – Newcastle		29/25	223/100	279/100	13
Tomorrow, 28 Nov			1	X	2	B’s
12:30	Brighton – Liverpool		111/25	81/25	31/50	13
15:00	Manchester City – Burnley		17/100	701/100	153/10	13
17:30	Everton – Leeds		103/100	11/4	131/50	13
20:00	West Brom – Sheffield Utd		163/100	221/100	48/25	13
29 Nov 2020			1	X	2	B’s
14:00	Southampton – Manchester Utd		291/100	259/100	1/1	13
16:30	Chelsea – Tottenham		111/100	127/50	129/50	13
19:15	Arsenal – Wolves		113/100	233/100	11/4	13
30 Nov 2020			1	X	2	B’s
17:30	Leicester – Fulham		27/50	169/50	529/100	13
20:00	West Ham – Aston Villa		33/25	261/100	51/25	13

There is a fair bit of difference, as the bookmakers have gone for:

Palace, West Brom, Chelsea and Arsenal wins where I’ve picked draws in V1
Palace, West Brom, Chelsea and Arsenal wins where I’ve gone for draws in the first two and away wins in the second two for V2
So in 60% of the matches, the models are in tune with the bookmakers odds – if they can all get over the line, it is the swing matches which will make the difference

Anyway that’s your lot for this article, now its time to see what can be done over the course of a long weekend. So until Tuesday…

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)