Although there was a quick turn-around time between game weeks 31 of the 32 sites monitored met, or exceeded, the 40% benchmark I have been using. The top 4 sites came at 90%. and the RMT Opti-bot predicted at a 70% level with 5 exact scores. Had you chosen to use the bot predictions it would have yielded a haul of 40+ points.
Maybe a full week to generate predictions gives the prognosticators too much time to make last minute adjustments to their processes. It could be that quick and dirty is the way to go. We will get the chance to see if the shortness between game weeks and good results hold up over the next few weeks.
This week presents another challenge since there only 2 days to gather enough data to make the statistical projections meaningful. With that in mind here is what the sites predict.
- 19 of 22 sites are predicting a Chelsea away win v Palace and 43.0% of the Opti-bot simulations agree with a 12.0% chance of a 1-2 score.
- 11 of 22 sites are predicting a Boro win at home v Swansea and 43.6% of the Opti-bot simulations agree with a 17.2% chance of a 1-0 score.
- 11 of 22 sites predict a Stoke win at home v Leicester but 37.4% of theOpti-bot simulations predicts a Leicester win. There are conflicting score predictions from the bot for this match with 10.2% for a Stoke 2-1 score and 10.4% favoring a Leicester 1-2 score.
- 10 of 22 sites predict a Sunderland win at home v Watford and 44.0% of the simulations agree with a 10.2% chance of a 2-1 score.
- 20 of 22 sites predict a West Ham win at home v Hull and 51.4% of the simulations agree with an 11.4% chance of a 2-0 score.
- 10 of 22 sites predict a Man United win on their visit to West Brom but with 34.0% of the simulations predicting a West Brom win and 34.8% predicting a United win it appears this match could go either way. The bot also predicts in 12.6% of simulations a 1-0 West Brom score.
- 11 of 22 sites predict a Bournemouth win at home v So’ton but the Opti-bot disagrees with 41.8% of the simulations predicting a So’ton win with a 15.0% chance of a 0-1 score.
- 10 of 22 sites predict an Arsenal away win v Man City and again theOpti-bot disagrees with 38.4% of the simulations predicting a City win with an 11.8% chance of a 2-1 score.
- 21 of 22 sites predict a Spurs win at home v Burnley and 62.8% of the Opti-bot simulations agree with an 18.6% chance of a 1-0 score.
- 11 of 22 sites predict a Liverpool win at Everton while 11 of 22 also predict a draw and 42.6% of the Opti-bot simulations predict a Pool win with a 10.8% chance of a 1-2 score.
As you can see, there are always a few matches where the sites and the Opti-bot disagree. I guess this will always be the case when trying to predict scorelines. “You pays yer money, and you takes yer choice”. It’s always the position we are going to find ourselves in.
The statistical predictions for GW17 are –
|TOT||v||BUR||2.4-0.1||3-0||2-0 or 3-0|
The top sites for GW16 were –
|# of Predictions||# correct wins/draws||# Correct Exact Scores|
|3 sites tied||10||7||2|
|4 sites tied||10||7||1|
|2 sites tied||10||6||3|
|2 sites tied||10||6||2|
|3 sites tied||10||6||1|
|Poisson Model 1||10||3||0|
No claims are made that the results posted here will outperform the processes you currently use, or that they should replace what has worked for you in the past. Basically, I look upon them as a time-saver for many managers who lack the time to surf the Web for this data.