Greetings Fellow International Soccer Enthusiasts,
My name is Ron Kessler and I have been publishing my own international
soccer rankings (men and women) since May of 1997. During this time,
I have repeatedly touted my rankings to be superior to FIFA's own
men's ratings. It has finally dawned on me that talk is cheap and it
was time for me to prove my boasts or shut up.
The results, in brief, are as follows: My rankings will produce a
a better relative rating of the involved teams for more than 60% of
all matches. When considering only competitive matches where there is
a significant difference between my and FIFA's ratings, that number
jumps to over 67%, or my system prevailing in more than 2 out of every 3
matches.
I have developed my system and am publicizing it and this study to
the best of my ability so that fans of the beautiful game can have a
reliable alternative to FIFA's sometimes mystifying rankings. The first
incentive to do this came in 1995 when I saw FIFA had Norway ranked 2nd
in the world and thought, "Geez, even *I* can develop a better system
than this!". My rankings (and this study) can be found online at
http://home.sprynet.com/sprynet/ronkessler/rankings.htm
So without further ado, here are the details of the study:
I compared matches that fit one of four scenarios as follows:
1) Neutral site matches where FIFA and I disagreed on which team was
ranked higher. I could use straight Ron Win/FIFA Win/Draw results here.
2) Non-neutral site matches where FIFA and I disagreed on which team was
ranked higher. I sorted these into two categories - matches where
FIFA's pick was playing at home, and matches where my pick was at home.
I then compared the home winning percentages for the two systems.
3) Neutral site draws. For these matches, I checked to see which system
had the two teams ranked closer together. Note that there is some
overlap with draws that occur in 1) above. I count them for each
scenario because they are being analyzed in two completely different
ways. These are the only repeating matches in the study sample.
4) Neutral site matches where both systems had the losing team ranked
higher. Here, I also checked to see which system had the two teams
ranked closer together.
I compared teams position numbers, comparing actual team ratings was
not possible due to the fact our systems are totally different. I used
a sample from December, 1995 to August, 1999. I run my rankings on the
same days as FIFA so that our rankings can be directly compared.
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
ALL MATCHES (no minimum difference between systems, includes friendlies)
# of
Mtchs Results Ron % FIFA% Diff (R-F)
1) 104 Ron 43, FIFA 24, Draw 37 59.1 40.9 + 18.2
2) 269 Ron Home - W 84, D 36, L 13
FIFA Home - W 61, D 36, L 39 76.7 58.1 + 18.6
3) 168 Ron 97, FIFA 64, Same 7 59.8 40.2 + 19.6
4) 103 Ron 68, FIFA 33, Same 2 67.0 33.0 + 34.0
---
644 Wgtd Diff = + 21.2% - Ron% = 60.6, FIFA% = 39.4
Chance of equal systems being this far apart at random: 1 in 14 million
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
MINIMUM OF 10 PLACE DIFFERENCE BETWEEN SYSTEMS' RELATIVE PLACEMENT OF
TEAMS (eliminates randomness of small differences) - Includes friendlies
# of
Mtchs Results Ron % FIFA% Diff (R-F)
1) 56 Ron 27, FIFA 13, Draw 16 62.5 37.5 + 25.0
2) 166 Ron Home - W 52, D 17, L 8
FIFA Home - W 39, D 26, L 24 78.6 58.4 + 20.2
3) 82 Ron 57, FIFA 25, Same 0 69.5 30.5 + 39.0
4) 54 Ron 44, FIFA 10, Same 0 81.5 18.5 + 63.0
---
358 Wgtd Diff = + 31.6% - Ron% = 65.8, FIFA% = 34.2
Chance of equal systems being this far apart at random: 1 in 580 million
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
COMPETITIVE MATCHES ONLY (no minimum difference between systems)
# of
Mtchs Results Ron % FIFA% Diff (R-F)
1) 70 Ron 28, FIFA 19, Draw 23 56.4 43.6 + 12.8
2) 133 Ron Home - W 45, D 23, L 5
FIFA Home - W 27, D 15, L 18 77.4 57.5 + 19.9
3) 105 Ron 58, FIFA 44, Same 3 56.7 43.3 + 13.4
4) 77 Ron 48, FIFA 27, Same 2 63.6 36.4 + 27.2
---
385 Wgtd Diff = + 18.4% - Ron% = 59.2, FIFA% = 40.8
Chance of equal systems being this far apart at random: 1 in 2870
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
COMPETITIVE MATCHES AND MINIMUM OF 10 PLACE DIFFERENCE BETWEEN SYSTEMS'
RELATIVE PLACEMENT OF TEAMS ONLY
# of
Mtchs Results Ron % FIFA% Diff (R-F)
1) 34 Ron 15, FIFA 11, Draw 8 55.9 44.1 + 11.8
2) 78 Ron Home - W 30, D 11, L 3
FIFA Home - W 14, D 8, L 12 80.7 52.9 + 27.8
3) 47 Ron 32, FIFA 15, Same 0 68.4 31.6 + 36.8
4) 39 Ron 32, FIFA 7, Same 0 82.1 17.9 + 64.2
---
198 Wgtd Diff = + 34.2% - Ron% = 67.1, FIFA% = 32.9
Chance of equal systems being this far apart at random: 1 in 650000
* * *
One last thing to look at is how FIFA has performed since
January, 1999 when they changed some of their parameters. The
sample is smaller than ideal, but yielded some meaningful
numbers:
ALL MATCHES AFTER JANUARY 26, 1999
# = 77 Wgtd Diff = + 29.0% - Ron% = 64.5, FIFA% = 35.5
Chance of equal systems being this far apart at random: 1 in 75
When compared with the weighted difference of + 21.2% for all
matches through the whole sample, it appears FIFA is doing even
worse than before - although again the sample is still too small
to say that for sure - but there is only a 1 in 75 chance that
my rankings are ahead of "new" FIFA by sheer random chance. One good
thing to say for FIFA's new parameters - our rankings disagree
much less frequently than they previously did. Before FIFA's
change, we disagreed on a whopping 15% of all matches. Since
then, the number has been reduced to 8.5%. So it is edging closer
to my system I guess but is still getting beat regularly when it
strays.
My regret with this study is that it is not being done
by a neutral third party. I can assure you that I have posted
EVERYTHING that I have tried and that I did not tamper with the
sample or any numbers (other than to round off and that was done
in an unbiased manner). Nonetheless, I welcome any third party to
compare these rankings themselves or to suggest new methods
to add to the study.
Thank you for your interest in my work, I hope this study
has proved to be interesting. Sincerely,
Ron Kessler
Prepared and maintained by Ron Kessler for the Rec.Sport.Soccer Statistics Foundation
Author:
Ron Kessler (ronkessler@sprynet.com)
Last updated: 19 Aug 1999
(C) Copyright Ron Kessler and RSSSF 1999
You are free to copy this document in whole or part provided that proper
acknowledgement is given to the author. All rights reserved.