It’s the first Friday of the Aviva Premiership and Guinness Pro 12 seasons, and the first chance to fire up the Crystal Egg for new games!
Without further ado, here are the predictions for the weekend ahead:
Pro 12
Home Team |
Draw |
Away Team |
||
| Munster | 94% | 1% | 4% | Edinburgh |
| Ospreys | 95% | 2% | 3% | Benetton Treviso |
| Glasgow | 42% | 3% | 55% | Leinster |
| Scarlets | 27% | 4% | 69% | Ulster |
| Connacht | 76% | 3% | 21% | Dragons |
| Zebre | 26% | 3% | 71% | Cardiff Blues |
Friday sees the Pro 12 kick off with two games that should be easy enough to predict without any kind of stats. Edinburgh improved their standing to eighth in the table last year, but even a relatively weakened Munster are heavy favourites to beat them handily at home – the model gives Munster a 94% chance of winning and a 52% of registering a bonus point win (though of course the stats don’t take into account the potential impact of telling your head coach telling your players they’re on the gravy train). Ospreys meanwhile are also strong favourites against a travelling Treviso team – 95% to win, with a 48% chance of getting five tries while they’re at it.
Glasgow are predicted to just about fall at home against Leinster, though they still have a 42% chance of upsetting the champions. Ulster are also reasonably strong favourites to pull off an away win against Scarlets, while the model seems strangely sure about Connacht beating Dragons at home (I suspect the model is a little biased towards home victories – something I’ll have to work on). Finally, Zebre have only a 26% chance of holding off Cardiff.
Premiership
Home Team |
Draw |
Away Team |
||
| Northampton Saints | 75% | 3% | 21% | Gloucester Rugby |
| Leicester Tigers | 96% | 1% | 4% | Newcastle Falcons |
| Saracens | 95% | 1% | 4% | London Wasps |
| Sale Sharks | 38% | 4% | 58% | Bath Rugby |
| London Irish | 33% | 4% | 63% | Harlequins |
| London Welsh | 39% | 6% | 55% | Exeter Chiefs |
The first game of the Premiership season is set to be a little less one-sided, though Northampton at home to Gloucester are still favoured to win, with a 75% of winning, but just a 20% chance of doing so with a bonus point. Leicester against Newcastle and Sarries against Wasps are predicted to be walkovers for the two big teams.
The remaining three games are set to be a good deal more competitive, with three narrow away wins predicted for Exeter, Quins and Bath against London Welsh, London Irish and Sale Sharks.
Bets
As always, a reminder that this is just my own playing around with figures for fun – I have no great claim to expertise, either in stats or rugby, but hoping to have some fun with this. Any suggestions for improvement on the model (details here) are more than welcome.
One interesting point that came up on Facebook when I first launched the model was how well or how badly its predictions matched up to historical bookies’ odds. It’s a subject I’m planning on writing about next week, but the short answer seems to be that they match up quite closely – very high levels of correlation and very similar levels of accuracy in prediction.
That appears to be borne out by the odds this week, which for the most part track quite closely to my predictions. This is both good (a sign the model might be accurate) and bad (hard to get any value). In any case, it looks as though the ebst value is to be found where favourites are undervalued. Saracens at 1.2 seem like a decent bet – the odds imply an 83% chance of winning, but the model predicts 95%. London Welsh might also be a good bet at 3.75 (implied probability 27% vs predicted 39%).
For the Pro 12, Munster at 1.18 also seem worth a punt (85% bookies, 94% Crystal Egg) whatever about their recent leaks, while unfortunately several bookies don’t have odds up at the time of writing and Betfair doesn’t have much value on offer!
For my first week I’ll stick with expected winners – £3 on Sarries and £2 on Munster. I suppose at least if I start making money from Saracens it will ease the pain of their winning…
EDIT: Ulster now seem pretty good value at around 2.1, so I’ll put a pound on them as well.
That’s all for now folks. Tune in again next week for a review of how the betting went, and some more thoughts on the model.
PS: It’s becoming apparent to me that it will be a bit of a rush to get an article out each week between 12pm on a Friday, when the last teams are announced for the Saturday and Sunday games, so there may well end up being two updates in future weeks – one on Thursday for the Friday games, and one on Friday evening / Saturday morning for the weekend games.