Well let met introduce to you "The Nyls".

The Nyls is my Poisson based prediction model for the Belgian Jupiler Pro league. It tries to predict the games for this competition. I was doubting about giving it a name after @beat-the-bookies has the Reverend, I couldn't stay behind!
The first idea was Nostradamus but like @stimp1024 correctly wrote: Nostradamus never has predicted any thing nice, only disasters and that's not what you do want to predict!
So after some deliberation, I decided to name it after my youngest son: Nyls!

What does it predict?

Well the model is actually 2 models. Both Poisson statistical models, but one is based on the results of the last 3 season and the current season, the other is based on the results from the current season only. This last one has no use until at least 5 games are played in the Jupiler Pro league, so for the moment it is using the results of only the past season!
For each game it does predict the chance that the home or away team will win or of course the draw.
It will predict the chance that both teams will score and if more or less than 2.5 goals will be scored.
As desert it will also try to predict the exact outcome of the predicted games.

How is it going so far this season?

Well in Belgium already 2 rounds are played and the results are looking promising!
If have taken the average odds from oddsportal.com.

There are 2 ways to use the model. One is to follow the percentage, the other is to follow the correct score. So a 1-1 prediction would result in betting on a draw, on both teams to score and an under 2.5 goals.

Home, draw or away

Or also know as 1X2 betting. Here you do bet on the winner of the game. This is the most common betting type used by bettors.

1x2 on percentage

Number of bets	Won	Pay-out	P/L
14	8	14.3	+0.3

This means that if you would have bet on each game according to the model, you would gave gained 0.3 units. This reflects a ROI of 2.14%!

1X2 based on the correct score prediction

Number of bets	Won	Pay-out	P/L
14	8	19.84	+5.84

This means that if you would have bet on each game according to the model, you would gave gained 5.84 units. This reflects a ROI of 41.71%!

The difference with the % betting is rather big. This is normal because the % prediction rarely predicts a draw, while the correct score does (maybe even predicting too many draws)!

Both teams to score (BTTS)

Here you do predict if both teams will find the net or not.

BTTS on percentage

Number of bets	Won	Pay-out	P/L
14	11	120.07	+6.07

This means that if you would have bet on each game according to the model, you would gave gained 6.07 units. This reflects a ROI of 43.36%!

BTTS based on the correct score prediction

Number of bets	Won	Pay-out	P/L
14	8	12.92	-1.08

This means that if you would have bet on each game according to the model, you would gave lost 1.08 units. This reflects a negative ROI of -7.71%!

Over or under 2.5 goals

Here you do predict if more or less goals than 2.5 will be scored. The 2.5 number of goals sound strange because it is not possible to score 0.5 goals but it is to make it clear!

Over/Under on percentage

Number of bets	Won	Pay-out	P/L
14	9	15.04	1.04

This means that if you would have bet on each game according to the model, you would gave gained 1.04 units. This reflects a ROI of 7.43%!

Over/Under based on the correct score prediction

Number of bets	Won	Pay-out	P/L
14	7	13.75	-0.25

This means that if you would have bet on each game according to the model, you would gave lost .25 units. This reflects a negative ROI of -1.79%.

The Correct score prediction

The holy grail among sports bettor. Crazy odds sometime, but so hard to predict!

Number of bets	Won	Pay-out	P/L
14	2	17.29	3.29

Only 2 winners but still good enough for a profit of 3.29 units, reflecting a ROI of 23.50%. Amazing!

How do I use it

No, I don't bet blindly according to the model. But if you do present it, you should also give some metrics about how the model is doing!
I do use it as a reference for placing my bets. First I will make up my mind about the bets I do want to play and next I do consult my Poisson model. This could cause me to place the bet, trow the bet away or change it!

Round 3

Here is the Nyls for round 3 of the Belgium Jupiler Pro League!

---

The Lars

I also have a Poisson Model available for the English Premier League. This one is called the Lars, named after my oldest son!
If you aren’t a sports bettor but am playing in the Steem Fantasy League organized by @acidyo, you can use it as reference for setting up the squad for the next round!

Here it is for the first round!

Cheers,
Peter

Sports Bettors On Steemit Unite and Join the SBC !

@steemitbloggers