https://snag.gy/iNfrRw.jpg

So here goes. I've used my [**EPL Indexes**](https://steemit.com/epl/@nanzo-scoop/nanzo-s-epl-indexes-2016-17-matchday-4) to build an algorithm that predicts the outcome of upcoming Premier League fixtures. 

# SCORE PREDICTIONS

Here are the first set of score predictions.


|  Home Team | Home Goals | Away Goals | Away Team |
|  ------ | ------ | ------ | ------ |
|  Man United | 1 |0  | Man City |
|  Arsenal | 3 | 2 | Southampton |
|  Bournemouth | 0 | 2 | West Brom |
|  Burnley | 0 | 1 | Hull |
|  Middlesbrough | 1 | 0 | Crystal Palace |
|  Stoke | 1 | 3 | Tottenham |
|  West Ham | 1 | 0 | Watford |
|  Liverpool | 2 | 1 | Leicester |
|  Swansea | 1 | 1 | Chelsea |
|  Sunderland | 1 | 2 | Everton |
 
 
I will compare my predictions to the actual results on Monday.
 
# BEATING THE ODDS
 
Just for fun, I’ve decided to place some bets on the outcomes. I’m using Betfair that allows me to back and lay teams. Amounts are in £ sterling. As you can see I’m not a big gambler!
 
### My bets
 
|  Bets | Type | Odds | Liability | Return |
|  ------ | ------ | ------ | ------ | ------ |
|  Arsenal | Back | 1.6 | 2 | 1.2 |
|  West Brom | Back | 4.1 | 2 | 6.20|
|  Burnley | Lay | 2.42 | 2.84 | 2 |
|  Tottenham | Back | 1.85 | 2 | 1.70 |
|  Liverpool | Lay | 1.67 | 1.34 | 2 |
|  Chelsea | Lay | 1.63 | 1.25 | 2 |
|   |  | **Totals** | **11.44** | **15.0** |
 
 
### My Rationale
 
I’m backing **Arsenal** because even though my algorithm predicted a tight home win, the range shows that the score could easily be 3-1 or 4-1 to Arsenal. Arsenal are also playing at home. This should be a banker. **West Brom** to beat **Bournemouth** is my value bet. My algorithm points strongly towards an away win and the bookmakers are offering generous odds. I was tempted to stake a little more on it, however as it is early in the season, I’ll play it conservatively. I’m also getting decent odds on **Burnley** not beating **Hull**. Given my algorithm has an away win, this bet plays itself. **Tottenham** should beat **Stoke**, so no controversy there. **Liverpool** not beating **Leicester** is a risking bet. Liverpool at home is an unknown quality. I’m betting this one on value (and Liverpool’s general inconsistency over the years). Also my algorithm has Liverpool only just winning, this game could very well be a 1-1 draw. **Swansea** vs **Chelsea** promises to be closer than advertised according to my algorithm. I’m playing this one as Chelsea are probably over-priced.
 
### Bets I passed on
 
The following bets, I’ve decided **against** playing as I feel there is little value at the odds given,
 
|  Bets | Type | Odds | Liability | Return |
|  ------ | ------ | ------ | ------ | ------ |
|  Man City | Lay | 3.55 | 5.1 | 2 |
|  Crystal Palace | Lay | 3.8 | 5.6 | 2 |
|  Watford | Lay | 4.7 | 7.4 | 2 |
|  Sunderland | Lay | 3.9 | 5.8 | 2 |
|   |  | Totals | 23.9 | 8 | 

# UP NEXT...

I’m actually looking forward to seeing how my algorithm fares! Even though it’s the first iteration and has a limited dataset, it's fun to put the theory into practice. 

Later in the week, I’ll be using these predicted results to determining the best players to pick for my FanDuel fantasy football team. 

Game on!