So in what is now becoming an annual exercise let’s review the 2017 season through the prism of Expected Wins (Expt Wins).The 2015 and 2016 versions of this article can be found here and here
For the uninitiated Expt Wins uses bookmaker’s odds (note 1), as a sort of independent arbiter, to see which teams over (or under!) performed versus what was expected on a game by game basis. It is a much better fairer view than sheer win percentages given (a) how relatively short the season is for most teams and (b) how uneven the Championship can be in terms of the quality of teams facing off against each other.
Table 1; 2017’s best and brightest
Unsurprisingly the top10 is peppered with teams that were promoted. This makes sense as for the majority of teams the league makes up at least two thirds of their season.
Carlow had an excellent season – but it was no fluke; they were also top5 in 2016. Over the past two seasons they have won 11 games when they were only expected to win ~7.7. And they managed to top the 2017 table despite losing as 1/16 home favourites against London. 1/16, without accounting for the bookmaker’s margin (see note1), implies a 94% win probability. Win that game and their “above Expt Wins” total would be twice that of second placed Louth. That loss is the shortest price loss in the database and must be one they desperately want back. There is no guarantee that Carlow would have gained promotion had they beaten London – as Wexford no doubt would have put greater emphasis on their final two games – but they must be absolutely kicking themselves every time they think of that game. And yet – they still topped the 2017 table despite this loss.
Now I am by no means an expert on the ins and outs of Louth football – and you have to think they have a good ‘un in Pete McGrath – but you have to feel for Colin Kelly. Back to back promotions. A 59% win rate over that period which lands them 5th on that metric behind the likes of Dublin, Tyrone, Kerry and Kildare (themselves aided by back to back promotions) and along with Carlow the only team to finish in the top5 Expt wins both years.
The two Championship campaigns were poor in comparison (played 6 won 3 with three loses of 4, 6 and 9 when stepping up against Derry and Meath) but still … be careful what you wish for.
In 2016 Tyrone and Cavan were in the top10 following successful promotions from Division2 and the trend holds true for Galway and Kildare. Division2 is always very tight – just under half (27 of 56) of all games in the last two years had a zero or one point handicap. Extend that to two points and 80% of the games are covered. Win enough games to gain promotion in these tight contests and you are well on your way to outperforming expectations for the season.
Table 2; 2017’s laggards
This is a mixed bunch of
1. Division4 teams who struggled to register wins and who are perennially down the bottom of these rankings – Limerick, Waterford, Wicklow
2. Teams that had a disastrous season – Cavan, Laois, Derry
3. Very good teams that didn’t get the job done enough – Kerry & Mayo
4. Cork!
Taking the four cohorts in order
1. The worst team in 2017 was (subjectively) Wicklow but no matter how bad you are when you play your peers in the league you are always given “some” chance. Wicklow’s seven league appearances saw them chalked up at odds of 8/11, 3/1, 5/6, 10/3, 11/2, 6/5 and 6/1. When we remove the bookmaker’s margin that equates to an expectation of two wins. And that’s for the “worst team”. Limerick’s odds were 8/13, 11/10, 13/8, 1/10, 8/15, 11/8 and 1/10 which comes out at just over four wins.
No matter how poorly you are viewed under Expt Wins you will always be expected to notch up at least two wins and maybe four or five … if you struggle to win games full stop you will always be down the bottom end of this table.
2. All three of Cavan, Laois and Derry were relegated and whilst combined they won 4 of 10 Championship games three of those victories came against Division4 teams when they were heavily favoured. At a very high level this cohort win the games they are expected to win, lost the ones they were expected to lose and came out the wrong side of way too many 50:50 calls
3. Mayo being so low on the table is easy enough to explain; in the three games that they drew Mayo were 1/5, 1/6 and 23/10 – those three games alone account for their negative Expt Wins. Kerry are slightly different. They may have finally managed to beat Dublin in the Division1 final but outside of that they failed to win half their games – with three of those games coming against Mayo when a good favourite (2/5, 1/2 and 8/13). They were almost prohibitively favoured at 1/20, 1/5 & 1/6 in the three Championship games that they won. That mixture (winning when big favourite, losing/drawing when favouritism is less obvious) is a recipe for a poor Expt Wins season
4. Cork. Ah Cork. For the second year in a row they appear in the bottom5 but can you imagine how poor they would look had Waterford managed to tack on one more point when Cork were 1/50? Cork were middle of the pack on win ratio (winning 41% of their games) but were overturned by Tipperary as a 7 point favourite in the 2016 Championship whilst also losing at odds of 1/3, 4/11 and 4/11 over the two league campaigns. They never won a game as underdog to balance these losses.
Is it predictive?
Although there are outliers – notably Carlow, Louth and Cork – I would lean towards no. There is just too much volatility as teams yo-yo up and down the table; Kerry from 27th in 2015 to equal 6th in 2016 and then back down to 28th in 2017; Cavan from 9th to 31st, Armagh from 32nd to 5th. Good luck trying to pick which of this year’s top5 will stay there!
Note1; calculating Expt Wins
Using the All Ireland final as an example. Paddy Power’s odds for the game were Dublin 4/9, Mayo 3/1 with the draw being 9/1. All that these fractional odds are is another way of expressing probabilities. To work out the probability any odds equate to you use the following formula (B/ (A+B)). For Dublin’s 4/9 the B here = 9 and the A = 4 so the probability of a Dublin win = (9/ (4+9)) which equals 0.692 or 69.2%. Do this for all three odds and you get
Dublin = (9/ (4+9)) = 69.2%
Mayo = (1/ (1+3)) = 25.0%
Draw = (1/ (1+9)) = 10.0%
The total percentages add up to 104.2%. Now we know that there are only the three outcomes for any game – team1 wins, team2 wins and draw – so anything above 100% for these three outcomes is the bookmaker’s margin. To get a truer understanding of the probabilities we strip out the margin equally across the three outcomes and come up with an Expt Win for each team. Dublin in this instance = 67.8% or 0.678 (69.2%-((104.2%-100%)/3)); Mayo = 23.6% or 0.236 (25.0%-((104.2%-100%)/3))
Note2; the odds
All odds are taken from Paddy Power and tend to be taken towards the back end of the week (Friday night/Saturday morning) to let any movements settle down. It is possible that injury news etc. changes the odds between what was taken and what they were at throw in but I’m comfortable enough that this would be a rare enough occurrence not to have too big an impact.
dontfoul 