Week 3 Pre Game J League modelling predictions

What follows are the predictions for the two probability based models from the current goals scored and XG in the Japanese J League. A detailed description can be found on the following link as well as the results from the past two rounds of games; 

https://anegitsguidetofootballbetting.blogspot.com/2024/07/how-to-price-up-football-match.html

Disclaimer; This blog is not a football tipping service. The modelling shown below is for information purposes only. The data will not be altered post event but if you're of a distrusting nature feel free to screenshot the post and compare.

In the model we are using decimal odds post results and an example of £1 bet on each game result. 

The first prediction is what the bookmaker makes the game favourite. The next is the probability taken from the result based on a goals scored model. The last is the probability taken from the result based on an XG Model. 

The basis of the two models is the same, only the base data between actual goals scored and XG changes. 

Predictions are shown as the side favourite with the bookmaker - Goals scored model - XG Model

Midweek predictions

Yokohama v Sagan 1 - 2 - 1

Weekend fixtures 

1 Sanfrecce v Vissel 1 - 1 - 1

2 Kashima v Hokkaido 1 - 2 - 1

3 Machida v Nagoya 1 - 1 - 1

4 Tokyo v Cerezo 2 - 1 - 2

5 Albirex v Sagan 1 - 2 - 1

6 Jubilo v Kawasaki 2 - 2 - 1

7 Urawa v Shonan 1 - 1 - 1

8 Gamba v Yokohama 1 - 2 - 1

9 Kashima v FC Tokyo 1 - 2 - 2

10 Avispa v Kyoto 1 - 2 - 1

Prices that will be used to calculate returns on any correct results were taken from Bet365 Tuesday July 2nd. We will be keeping a running total each week to see how each model is performing.

For exact current totals see the link above but at present the results are as follows;

Simply betting all favourites with the bookmaker each round ❌

Betting each round using the GS Model predictions ✅

Betting each round using the XG Model predictions ❌

Update 1 at 18:15 hours Friday. Neither bookmakers favourite or XG model has won. The GS model is 1 from 2 and the experiment like you'd find on a tipping service is whether betting on all outcomes would it return a profit.

Just a reminder after the opening two rounds of data with the hypothetical examples used we were left like this;

Accumulative totals

Bookmakers £20 stakes £18.39 (-£1.61)

Goals scored model £19 stakes £22 (+£3)

XG Model £20 stakes £17.52 (-£2.48)

Round 3 results

Bookmakers favourite winners;

2, 3, 8 and 9 @ 1.42, 1.95, 2.62 and 1.95. Returns of £7.94 (-2.06)

Goals scored winners;

3, 5 and 10 @ 1.95, 3.60 and 3.60 Returns of £9.15 (-0.85). First negative returns for the goals scored model. Note if you'd bet the midweek fixture it would have been £11 for returns of £10.81 (-£0.19).

XG winners;

1, 2, 3 and 9 @ 4.33, 1.42, 1.95 and 2.62. Returns £10.32 (+£0.32)

New Accumulative totals

Bookmakers £30 stakes £26.33 (-£3.67)

Goals scored model £29 stakes £31.15 (+2.15)

XG Model £30 stakes £27.84 (-£2.16)