Ellis Chaddock has solved UEFA's CL seeding problems.
As a Manchester United fan in my mid-twenties the last Thursday in August has been a joy to experience for as long as I can remember. The UEFA Champions League draw, what a prize for all the hard work I put in by supporting my team through the lean spells and the dark days (we finished 4th one year, it was rough). Which giants of Europe we will be clashing with this year? Will we get to test ourselves against the might of FC Copenhagen, or maybe we might be lucky enough to play Benfica for what feels like the 10th straight season?
To any outsider to the sport, this spectacle should be more than enough to satisfy – How could a three-hour ceremony dedicated to picking the names of 32 teams out of 4 bowls not be "enough"? At least the pay-off is understandable, the grand procession culminating in six potentially intriguing matches per team & a valid excuse to visit three European cities.
As anyone who has spent any time following a team will tell you though, there is more to football than watching your own side; there is also the joy of seeing your rivals fail. I don’t think this ever manifests itself as strongly as it does in the UCL group stage draw.
Somewhat naively, I had high hopes that schadenfreude would get me through this year’s draw; as I’m sure you were all aware Manchester United were not going to be present, but as a Mancunian living in opposition territory I had been gleefully informed, on more than one occasion, that Liverpool would be.
Schadenfreude wasn’t enough. Sure Liverpool have to face Real Madrid and Basel, but they were just so damn happy to be back at the top table that it was hard to laugh at them from outside with my face pressed up against the glass. Even the minor hilarity of Bayern Munich vs Manchester City for the third time in four years couldn’t bring a smile to my face. I felt empty, a husk of a football fan. There must be something wrong with me I thought, maybe I had it wrong? Is there more to football than the tribalism it has become? Should I not be cheering on the other English teams that have made it to UEFA’s top competition?
“Make sure all the English teams perish”
It hit me like a wave. It was so clear. It was UEFA’s fault! They messed up the seedings. If one thing good was going to come of Manchester United finishing seventh then it would be this! I will fix the Champions League seeding system. You’re welcome, UEFA.
*Galatasaray placed as 1st in Turkey following Fenerbache's European ban.
**Monaco given 13.000 as coefficient as they have not played in Europe for the past five seasons. The club recieved Ligue Un's coefficient score.
Above are the UEFA coefficients for all teams entering the CL at the 2014/15 group stage & the pot they were entered into for the draw. UEFA sort all entrants to the group stage purely on the coefficient each club obtained for the end of the 2013/14 season, this covers each club’s performance in European competitions over the past 5 years. When a club has not been in European competition, they simply take the country’s coefficient score, as evidenced by Monaco’s score of 13.000.
What struck me regarding this system is how heavily it is skewed in favour of qualification over the previous season’s performance. Arsenal and Chelsea have consistently qualified, benefitting from England’s high UEFA coefficient and number of qualification spots, which has seen them land in Pot A despite finishing third and fourth in the English Premier League. Meanwhile, every team in pot 4 finished second or higher in their league; including AS Roma & AS Monaco FC.
Don't forget to place your bets for this year's Champions League. Get the latest Guts sports bonus codes and boost your winnings!
So armed with Excel, and a vague memory of my statistic GCSE, I powered forward into the abyss that is UEFA’s ranking system. My aim was to create a system that would have a greater balance between rewarding the previous season’s performance and consistent European qualification (a secondary objective was to stop Chelsea and FC Schalke playing each other again, a match worse than Vorgon poetry). I started by grouping the 32 teams based on previous year’s performance:
2014 Champions League winner: 1
2014 League winner: 18
2014 League Runner-up: 6
2014 League 3rd place: 4
2014 League 4th Place: 3
I used these groupings for my first attempt to rearrange the rankings. Clubs were grouped into their pots based on the previous season performance & then by coefficient. The new seedings based on this can be seen below. Whilst this did give more immediate rewards to clubs who finished higher in their respective domestic leagues, you can see quite clearly it doesn’t reward consistent qualification for Europe.
Alas, all this had done was create two pots (A+D) which were saturated with clubs who had qualified consistently; four of the teams now in pot D were previously in Pot A. Something needed to be done to redress the balance. I jumped headfirst into statistics mode to try and see if the numbers could point me in the right direction. I found that the average coefficient was 75.938. Remembering a combination of my statistics GCSE, and everything I have learnt about business over the years, I decided it was time for some standard deviating. Whilst I am most certainly not using this in the six-sigma black belt defined parameters that my business professor was so fond of, I feel it could be a useful tool to confirm that each pot in my new draw is as competitive as possible.
My business professor
Standard deviation gave me a range of acceptable average coefficients for each pot of 34.62 to 117.26; or to use football terms, an average strength somewhere between APOEL and Arsenal. If I was going to slate off UEFA so vehemently I decided that it was probably worth checking that their current grouping didn’t already meet these criteria. It didn’t. Pot A in this year’s draw had an average coefficient of 135.22 & Pot D had an average coefficient of 26.89. This means a gap of roughly 110 points (55 wins in Europe over the past 5 years) between teams from the top and bottom pot. Whilst my new ranking system gave coefficients of:
Pot A: 109.92
Pot B: 55.80
Pot C: 44.26
Pot D: 93.76
Standard deviation test passed. But the range from group to group still seems too high as we have a case of 2 high quality groups and two lower quality groups. How can I rearrange the groups to address this? This is hard work; I can understand why Platini earns the big bucks now. After much soul searching, I gained inspiration.
A tiered system! That will cure which ails me! This is a little more complex than the previous system but works as such:
Pot A: Winner of previous seasons UCL + Divisional Champions 2013/14
Pot B: 2013/14 League Winners & Second Place
Pot C: 2013/14 League Winners, Second Place & Third Place
Pot D: All teams
SO, to qualify for pot A you must have finished top of your league, for pot B at least second, and so on. If more than eight teams qualify to be entered into each Pot they will then be sorted by coefficients, as before.
At first glance, I think this could be it; there are big clubs in each pot indicating a balance between immediate & sustained success. The statistics confirm this:
Pot A: 109.92
Pot B: 67.83
Pot C: 85.30
Pot D: 40.62
Whilst this weakens Pot D, it does so by improving Pot B & C’s coefficients, which means there is a smaller difference between Pots.
I’ve cracked it & I am now ready to host my own UEFA Champions League draw. I don’t quite have the same budget as UEFA though, so the draw will not be held from Paris, but rather my girlfriend’s bedroom. There will be no luminaries and past stars of the game. We have her cat, Ben. A fine replacement for Gianni Infantino, I’m sure you’ll agree.
Unfortunately, the excitement proved too much for Ben meaning we were left without a star to present the UEFA left-back of the year award.
I thought we might be able to use the little plastic eggs which hold the toys inside Kinder eggs, but again, no budget. It mattered not, this was what it had all been for, the moment for me to prove I could create a better group stage draw than UEFA, one which would leave the neutral and the invested fan positively tingling with excitement for the months ahead. So, without further ado: the draw.
Well, here are the results. I would personally say it was a resounding success as Chelsea didn’t face off against Schalke. There is only a single group (Group H) that may be considered weaker; and even though I swear to god I did not plan this to happen both Liverpool & Manchester City were dealt possible groups of death (I’m sorry City fans, it seems that fate has decided you must face Bayern Munich. I can fight UEFA but I can’t fight destiny).
Anyway, the point is: if I can come up with a better seeding system, why can't UEFA?
If you would like to criticise Ellis' improper use of statistical analysis you can find him on on twitter at @Ellis_bo_fellis.