Saturday 6 February, 2.15pm
Given the difficulties in modelling this year's Six Nations, writes Jack Houghton, an interest in the over-priced Scotland makes most sense...
"A flap of a butterfly's wings in Brazil can indeed set off a tornado in Texas."
Digging out the Elo ratings
Despite not considering myself a rugby punter, several years ago I set up a ratings system for the international game. At the time, I was writing an article about the approaches used by various sporting bodies to rank their competitors, and using a tried-and-tested predictive tool - Elo ratings - seemed the obvious way to compare the value of other approaches.
Given the relative paucity of international rugby games, it's a quick job to update the ratings each year in time for the Six Nations: an event that's proved surprisingly profitable considering how little I know about the game.
This profitability - I've always assumed - is due to a mix of the relative unpopularity of international rugby as a sport to bet on among the statistical class, combined with the tribal nature of fans who become once-a-year punters.
Often, this means England - with the biggest fanbase and a biased broadsheet press over-representing them in previews - are over bet, leading to market ricks.
Can the ratings be trusted?
Predicting this year's Six Nations, though, feels less straightforward than it has in the past. England are short favourites at around 2.3211/8, but my usual methods of assessing whether this is an accurate representation of their chances seem lacking.
For a start, how accurate are my Elo ratings? A decimated 2020 calendar, with hastily rearranged tournaments played amidst player and fan restrictions, make them seem woefully out of date. Are the ratings I have now in any way representative of the teams that will take the field?
Struggling to simulate home advantage
Then there is my model for arriving at odds for the tournament. For this I use a statistical tool called a Monte Carlo simulation, where my computer chugs away, playing out thousands of imaginary Six Nations, based on my Elo ratings, to arrive at a final percentage chance for each team.
The problem here is that the simulation builds in a home advantage: based on several studies that demonstrate its significant impact in international rugby, especially at the Six Nations. But will that home advantage apply as teams perform in front of empty stadia?
Garbage in, garbage out
Statistical approaches in any discipline are easily undermined by small inaccuracies in inputted data. A percentage point or two misplaced at the start of a simulation ripples through it, growing to a wave after a few iterations. Or, to use Edward Lorenz's more elegant analogy, a flap of a butterfly's wings in Brazil can indeed set off a tornado in Texas.
Simulating this year's Six Nations, the windspeed rises quickly.
Scotland the unlikely choice
If forced to bet, I would consider England rightful favourites. My model gives them close to a 50% chance of winning the tournament. This is driven partly by their superior rating: in a match against the two next-best teams (Ireland and France) at a neutral venue, they would have around a 70% chance (1.422/5) chance of winning.
It is also heavily influenced by advantageous scheduling, though: England play three matches at home, to Ireland and France's two, and one of their away matches is to Wales, the second-weakest outfit in the tournament. If we assume that home advantage is less of a factor this year, the 2.3211/8 on England looks about right, and back-to-lay punters can have some confidence that they will be able to get out of any profitable position before that final away match to Ireland.
A more interesting prospect might be Scotland, though, at around 24.0023/1. Whilst England are clearly the best team according to the ratings, there is little between the next three in the list, of which Scotland is one.
The biggest hurdle for Scotland is clearly their opener - away - to England. However, whereas the market makes England the 1.192/11 favourite there, I see the match as being more keenly contested, with England more like 1.321/3 shots, with those odds expanding further without a home advantage factored in.
It's an unlikely prediction, perhaps, but to small stakes, Scotland is the butterfly bet.
You can now bet without leaving Betting.Betfair with our brand new on-site betslip for Exchange markets. You'll see the Exchange back and lay prices at the end of articles - simply login and place your bets as you would do on the main Exchange site
Back Scotland @ 24.0023/1
Saturday 6 February, 2.15pm