A week of mixed results. The model may have only correctly predicted 8 from 12 games (for an overall accuracy this season so far of 73%), but my bets all came good – including the first of the year where I disagreed with the bookies on Gloucester toppling Leicester at home. My stake is now up to £12.71, a high so far.
Here’s what the weekend’s results have done for the Crystal Egg’s predictions for final league tables:
| 1 | Leinster | 73 |
| 2 | Ospreys | 72 |
| 3 | Glasgow | 72 |
| 4 | Ulster | 69 |
| 5 | Munster | 67 |
| 6 | Scarlets | 56 |
| 7 | Cardiff Blues | 46 |
| 8 | Connacht | 44 |
| 9 | Edinburgh | 40 |
| 10 | Dragons | 36 |
| 11 | Benetton Treviso | 29 |
| 12 | Zebre | 25 |
| 1 | Saracens | 78 |
| 2 | Northampton Saints | 75 |
| 3 | Bath Rugby | 65 |
| 4 | Leicester Tigers | 63 |
| 5 | Harlequins | 59 |
| 6 | Exeter Chiefs | 56 |
| 7 | Gloucester Rugby | 53 |
| 8 | London Irish | 51 |
| 9 | London Wasps | 50 |
| 10 | Sale Sharks | 44 |
| 11 | Newcastle Falcons | 28 |
| 12 | London Welsh | 17 |
Of course a weekend of upsets is bound to have implications down the line. Not only have some teams won points they would not have been expected to last week (Munster!), it also means a good model should re-evaluate how good it thinks the team is. It’s also important not to over-do this – even excellent teams still have upsets and runs of bad form. This is why the Crystal Egg still predicts Leinster will end up topping the table come May – although only just, given that their chances have fallen from 49% last week 28% this week, just one percentage point behind the Ospreys.
Elsewhere, Munster have given their chance at the playoffs a real shot in the arm – their combined odds of coming in the top four are up to 64% from 50%. In short, the business end of the table is a lot more open than it has been so far, with Ospreys, Glasgow, Ulster and Munster all benefitting. Interestingly, the second most certain outcome for the season (after Zebre in last) is Scarlets coming sixth. The model seems to predict four pretty well-defined tiers which may explain this. At the top, there are the big five – Leinster, Munster, Ulster, Ospreys and Glasgow. Next there is the solidly mid-table, but not able to rise above it Scarlets. Then we have the mediocre to poor sides of Cardiff, Edinburgh, Dragons and Connacht. This tier used to also include Treviso, but unfortunately they seem to be joining their Italian friends Zebre in the “terrible” bottom tier. So while there may be a lot of competition within a tier – for example Munster or Glasgow could reasonably end up anywhere between first and fifth – there is much less chance of a team jumping up a tier – e.g. Connacht placing sixth.
This seems to be a bit of a contrast to the Premiership where there seems to be much more even spacing between the teams – with the exception of Newcastle Falcons and London Welsh, who handily sit in the terrible tier below everyone else in the Premiership.
Sarries also took a blow to their chances of topping the table this weekend, although to a much lesser extent than Leinster, dropping 12% to 63%. Bath have overtaken Leicester to third place, but still have only a 3% chance of topping the table – Northampton and Leicester between them accounting for 94% of chances to finish first.
At the other end of the table, another 52-0 pasting for London Welsh makes it look like its going to be a long season for their fans – though they should enjoy Premiership rugby while they can, with the model predicting a last place finish for them is a 90% certainty.