`ScoreSight weekend price checks - 4 to 6 September

scoresight

Casual Punter
Joined
Aug 26, 2026
Messages
11
Reaction score
1
Points
3
URL
ly winner is not automatically a good bet, so I am now recording the model probability, fair price, minimum acceptable price and the market price.

Saturday: Lens-Lorient X2. Model 46.7%, fair 2.14, minimum 2.25, market around 2.60-2.70. Again, 0.25u at most.

Sunday: Marseille-Paris FC draw. Model 27.9%, fair 3.58, minimum 3.75, market around 4.00. Small stake only.

The passes matter too: PSG at 1.36-1.42 is too short for my 53.9% estimate, and Arsenal at roughly 1.67 is below the 1.82 fair price from my 54.9% estimate.

These are provisional model estimates, not certified hit rates. Odds move, so I will record the actual price taken and skip anything below the stated minimum. 18+, gamble responsibly
 
Weekend price check for 12-13 September. I am logging three model views, but none as a full bet yet.

Atalanta are rated at 54.9% to beat Cagliari, which translates to fair odds of 1.82 and a minimum price of about 1.89 with a 4% margin. The 1.56 snapshot is too short, so this is a pass.

The Sheffield Wednesday-Wigan draw is estimated at 28.7%: fair 3.48, minimum 3.62, versus a 3.95 snapshot. The number is interesting, but the League One validation is inconclusive, so it remains a watchlist rather than a pick.

Trabzonspor are rated at 57.4% away to Konyaspor: fair 1.74, minimum 1.81, versus 1.98. That looks attractive on the raw calculation, but the separate market-comparison check failed, so I am not promoting it.

The main lesson for me this weekend is to publish the price needed as well as the likely winner. Has anyone found team news that materially changes one of these three estimates?
 
Sunday 20 September: Manchester City-Sunderland.

My updated model gives City a 70.3% chance of winning. That translates to fair decimal odds of about 1.42, compared with 1.29 in today's data snapshot. The model clearly favours City, but that estimate does not support taking such a short price.

These are provisional estimates while historical validation is being rebuilt, not a verified betting edge. What would change your assessment most here: City's starting XI or Sunderland's ability to slow the game down?
```
 
Back with an update before adding more price checks. I have been checking whether the probabilities behind them actually earn their confidence.

In our separate October validation sample, the model's most likely result was right in 32 of 62 matches. Simply taking the market favourite also gave 32/62. More importantly, the market's probabilities scored better: Brier 0.588 versus 0.612 for the model (lower is better).

These are 62 matches from seven leagues, played on 5-12 October, with results available in the current evaluation. The sample is incomplete and includes passes. It is not the settled record or ROI of the September ideas in this thread.

That means I cannot call a disagreement with the odds a proven edge. The estimates I posted here remain provisional. I will keep the misses and passes visible alongside the useful observations.

For the next entries, I want each price comparison to stand on its own: match, time of the odds snapshot, model probability, market price and the eventual result. What makes a prediction journal most useful to you: a clear match-by-match record, or the running comparison against the market?
 
Three matches from the refreshed weekend model, with an important distinction between the likely winner and the price:

Saturday, 17 October: Manchester City v Ipswich - City 75.1%; Bayern v Leipzig - Bayern 66.7%.

Sunday, 18 October: Twente v Utrecht - Twente 64.1%.

All three are outside the model's abstention flags, but that does not make every price attractive. Their calculated fair prices are approximately 1.33, 1.50 and 1.56 respectively. The odds in the 16 October model export are 1.22, 1.21 and 1.30, all shorter than those estimates.

So these are the model's preferred outcomes, while my price-check conclusion is to pass at the quoted prices. The probabilities remain provisional model estimates. No bet or profit is being recorded here. Which of these favourites do you think the market is pricing most aggressively?
 
Quick Champions League price check for tonight. I rechecked the market before posting, so these prices are only a snapshot.

- Fenerbahce-Liverpool: Liverpool 48.2% in the model, fair price 2.07, market roughly 1.53-1.61. That is a pass on price, even though Liverpool is the likely winner.
- Dortmund-Betis: Dortmund 56.2%, fair 1.78, market roughly 1.70-1.80. This is close to fair rather than an obvious edge.
- Viking-Sabah: Viking 59.0%, fair 1.69, market roughly 1.41-1.48. Again, the favourite looks right but the price is too short for this estimate.

The other six matches were left out rather than forcing a number. This is a price check, not a claim of a betting edge and not a recommendation to stake.

Would you treat Dortmund at around 1.80 as a fair confirmation, or is the margin still too thin?
 
Back
Top
GOALLLL!
Odds