Underdog

Everyone loves an underdog, bar the fans of the opposing team. This is true of any sport, including hurling. In hurling, three teams have dominated historically, with Cork, Kilkenny and Tipperary taking the lion’s share of championships, with over 70% of the total between them, and over 80% since the ‘back door’ system was implemented in 1997. However, this still leaves some room for wins from less decorated counties, many of whom have come from nowhere to unexpectedly take hold of the Liam McCarthy. In this article, we’ll be looking at each All-Ireland champion, from the beginning of the back-door era in 1997, all the way up until 2017, and we’ll attempt to determine how much they exceeded expectations.

Determining performance versus expectation

Obviously, expectations are subjective. Dedicated fans of even the weakest county may hold on to some hope that this could still be their year, while others are resigned to the same two or three teams winning year in and year out. As such, we’ve turned to maths to give us a less biased idea: using the BAINISTEOIR rating system’s database, we investigated the expected odds of winning it gave each year’s champion in each of their provincial and championship matches. This was subtracted from their actual result, and the results from each match was summed to give us a total wins above expectation’ figure. Of course, the championship structure in these years was unusual, and could result in winning teams having wildly different total numbers of matches. As such, a final figure was also calculated; ‘performance over expectation per match’, in order to let us view expectation through a slightly different lens. For example, a team who won a match with even odds would get a score for that match of +0.5, and the loser would get a score of -0.5. The team with the greatest performance above expectation can be considered the greatest underdog champion.

Overall results

The two charts below give us results for the two metrics for each year’s champion from 1997 up to 2017, ordered from the greatest overachiever to the smallest. While each team did perform above expectation (even the biggest favourite was never predicted be more likely than not to win the championship outright from the beginning), there is a large gap between the largest and smallest number. The first chart displays each team’s ‘wins above expectation’, while the second displays their expected performance exceeded per match.

There is, of course, a story behind each of these results, but we’re only going to go into detail on some of the extreme values:

Notable results

Clare, 2013: Greatest wins above expectation

Record: 6/1/1 W/D/L

Wins above expectation: 2.38

Performance above expectation per match: 29.7%

Context coming into the year

Coming into the 2013 All-Ireland Clare were a team finding their feet. After several underwhelming years, they had gained promotion to 1A in the league in 2012, and managed to cling on to this position in 2013, despite being in the relegation playoff. Their championship performance was still poor in recent years, as they consistently failed to either win in Munster or progress through the qualifiers to the quarter finals. The championship favourites were still Kilkenny, with people optimistic for an upset looking towards either Galway or Tipperary.

Match by match

Game 1: 02-Jun-2013, Munster QF, Semple Stadium, Thurles

Clare 2-20 (26) vs 1-15 (18) Waterford

Expected odds to win: 38.8%

 

Game 2: 23-Jun-2013, Munster SF, Gaelic Grounds, Limerick

Clare 0-15 (15) vs 0-23 (23) Cork

Expected odds to win: 48.7%

 

Game 3: 06-Jul-2013, All-Ireland R1, Cusack Park, Ennis

Clare 1-32 (35) vs 0-15 (15) Laois

Expected odds to win: 78.5%

 

Game 4: 13-Jul-2013, All-Ireland R2, Semple Stadium, Thurles

Clare 3-24 (33) vs 1-20 (23) Wexford

Expected odds to win: 63.9%

 

Game 5: 28-Jul-2013, All-Ireland QF, Semple Stadium, Thurles

Clare 1-23 (26) vs 2-14 (20) Galway

Expected odds to win: 34.5%

 

Game 6: 18-Aug-2013, All-Ireland SF, Croke Park, Dublin

Clare 1-22 (25) vs 0-18 (18) Limerick

Expected odds to win: 52.2%

 

Game 7: 08-Sep-2013, All-Ireland F, Croke Park, Dublin

Clare 0-25 (25) vs 3-16 (25) Cork

Expected odds to win: 47.6%

Game 8: 28-Sep-2013, All-Ireland F (Replay), Croke Park, Dublin

Clare 5-16 (31) vs 3-16 (25) Cork

Expected odds to win: 47.8%

 

Notes on their path to the All-Ireland

Aside from the qualifier rounds, Clare were never considered true favourites, with their next best odds occuring in the semi-finals against Limerick, where the odds were roughly even. Going by expectation once they were in a knockout situation (i.e. From the qualifiers onwards), their rating yielded only a 1 in 20 chance to win the All-Ireland, in a year where only 13 teams were competing. Of the three ‘big’ teams coming into the tournament (Kilkenny, Tipperary and Galway) Clare only had to face Galway on the way, but even with this favourable situation, their odds were not good. Despite this massive overperformance in terms of wins over expectation, their performance above expectation per game was more in the middle of the pack. This was due to their initial loss against Cork in Munster. This produced a negative score which took time to compensate for, especially given their easier route through the qualifiers. However, the losses gave them a larger total number of matches than the average All-Ireland winner, allowing them to further boost their wins above expectation. Despite not winning all of their games, Clare stumbled at fortunate moments, in situations where they wouldn’t be knocked out immediately.

Cork, 1999: Greatest per-match overperformance

Record: 4/0/0 W/D/L

Wins above expectation: 2.08

Performance above expectation per match: 50.7%

Context coming into the year

The 90s had been a decade of upsets, of underdog teams coming to the fore; championship wins from 94-98 by Clare, Offaly and Wexford upset the normal balance, and the big three were struggling to right things. Tipperary had reached a Munster and All-Ireland final two years before, but failed to win either. Kilkenny reached the previous years’ final, but failed at that too, and so took the step to replace their manager with a former star player named Brian Cody. Meanwhile, Cork were nowhere to be seen; they couldn’t even reach provincial finals, let alone All-Ireland finals, and hadn’t contested either since 1992. Their team was totally inexperienced, with an average age of 22. In 1998, they were beaten by 8 points, only scoring 13 themselves, to be knocked out of Munster and the championship.

Match by match

Game 1: 13-Jun-1999, Munster SF, Semple Stadium, Thurles

Cork 0-24 (24) vs 1-15 (18) Waterford

Expected odds to win: 58.2%

 

Game 2: 04-Jul-1999, Munster F, Semple Stadium, Thurles

Cork 1-15 (18) vs 0-14 (14) Clare

Expected odds to win: 44.2%

 

Game 3: 09-Aug-1999, All-Ireland SF, Croke Park, Dublin

Cork 0-19 (19) vs 0-16 (16) Offaly

Expected odds to win: 44.2%

 

Game 4: 12-Sep-1999, All-Ireland F, Croke Park, Dublin

Cork 0-13 (13) vs 0-12 (12) Kilkenny

Expected odds to win: 50.3%

 

Notes on their path to the All-Ireland

Cork’s All-Ireland win in 1999 was a total shock to many, as they had been so uncompetitive over previous years, and their team was mostly composed of extremely inexperienced players. Though their route to the championship only comprised of four matches, preventing them from scoring as highly on the ‘wins above expectation’ metric, their per-match performance over expectation was phenomenal. Their game against Waterford was the only match where they were really favourites; the final had them at roughly even odds, while they were underdogs in the remaining two matches.

Kilkenny, 2012: Winning an All-Ireland by performing to expectation

Record: 4/1/1 W/D/L

Wins above expectation: 0.05

Performance above expectation per match: <0.01%

Context coming into the year

Kilkenny, at this point in time, were very firmly established as the best team in hurling, possible ever. They had won three quarters of the All-Irelands in the previous dozen years, and they had been finalists in 11 of the previous 14 years. Including the 2012 season, they had won 6 out of the past 11 leagues, and been finalists in 8. Until proven otherwise, they remained the most dominant team in the country, and despite strong showings in recent years by Galway and Tipperary, Kilkenny were still expected to win both Leinster and the All-Ireland in 2012.

Match by match

Game 1: 23-Jun-2012, Leinster SF, O’Moore Park, Portlaoise

Kilkenny 2-21 (27) vs 0-09 (09) Dublin

Expected odds to win: 75.7%

 

Game 2: 08-Jul-2012, Leinster F, Croke Park, Dublin

Kilkenny 2-11 (17) vs 2-21 (27) Galway

Expected odds to win: 81.6%

 

Game 3: 29-Jul-2012, All-Ireland QF, Semple Stadium, Thurles

Kilkenny 4-16 (25) vs 1-16 (19) Limerick

Expected odds to win: 83.1%

 

Game 4: 19-Aug-2012, All-Ireland SF, Croke Park, Dublin

Kilkenny 4-24 (36) vs 1-15 (18) Tipperary

Expected odds to win: 61.9%

 

Game 5: 09-Sep-2012, All-Ireland F, Croke Park, Dublin

Kilkenny 2-13 (19) vs 0-19 (19) Galway

Expected odds to win: 72.0%

 

Game 6: 30-Sep-2012, All-Ireland F (Replay), Croke Park, Dublin

Kilkenny 3-22 (31) vs 3-11 (20) Galway

Expected odds to win: 71.3%

 

Notes on their path to the All-Ireland

Strangely, there are several parallels between the greatest (2013 Clare) and smallest (2012 Kilkenny) overperformance by total wins over expectation. Both occurred within a two year period, both teams failed to win their provincial championship, and both teams won the All-Ireland in a replay against the team who initially beat them. Despite these similarities, the 2012 Kilkenny win stands out as a huge outlier compared to the other results. The key difference was expectation: Kilkenny weren’t meant to lose any of their games. This was their 6th championship in 7 years, and their lowest odds, against Tipperary in the semi-final, still had them winning 3/5ths of the time. In contrast, the 2013 Clare team were underdogs in the majority of their games, and it was considered unlikely for them to progress beyond the quarter finals. Even with a perfect record, Kilkenny were too dominant to ever get much of an ‘overperformance’ rating, so any deviation from perfection was able to destroy their overall score. As such, this year, where they stumbled twice against Galway, losing one game and drawing another, they barely performed to their original expectation by winning the All-Ireland.

Conclusions

As mentioned above, the definition of underdog can be highly subjective, and each championship carries its own narrative with it. However, it’s very interesting to see that, with nothing more than a mathematical formula and a database of match results, the BAINISTEOIR rating system stumbled upon results that would likely line up with answers given by of a large proportion of the population. The 1999 Cork and 2013 Clare teams were both young and promising, but few predicted them to have championship success as early as they both did, and their victories stunned the dominant teams of the time. When people think of surprise, dark-horse championship wins of the last 20 years, these two both spring quickly to mind.

On the other side of things, the rating system also made an accurate assumption in the other direction: that by 2012, everyone except Kilkenny were sick of Kilkenny, and them winning another championship was nearly a foregone conclusion.

The rating system cannot construct or interpret narratives, and the questions will always exist, as to which championship winner had the easiest or hardest path, or was the most or least surprising, or was the best or the worst. However, by filtering through the data, and assigning these mathematical values, it’s possible to find statistical support for many of the popular arguments to all these questions.