2024-11-29, 12:41 PM - Word count:
After a few seasons of not having made an article, I have finally found a fun article to make (and also found some time to make it). This article is very heavily based on Canadice’s old ELO ranking articles.
First off, how did the ELO come to be. For this I mostly used the ELO system FIFA also uses for its man’s international team ranking. Which means that after a match the new ELO for a team gets calculated by the formula:
P = PBefore + I * (W - We)
P – Points (After the game)
PBefore – Points before the game
I – Game Importance
W – Result of the match
We – Expected Result
P and PBefore are pretty self-explanatory.
The game importance is slightly changed compared to the FIFA scoring, as for FIFA international teams this is heavily build around the world cup (and continental cups). This means that instead of going from 5 (friendlies outside the window) to 60 (FIFA world cup after quarter finals), I have decided on a more easy system of, 10 for friendlies (same as FIFA), 20 for shield, 30 for D2/Minor League, 40 for Cup/D1/Major League/The League.
The result of a match is also slightly different from the actual FIFA calculation, I am keeping the same 1 for a win, 0.5 for a draw, 0 for a loss, however to not introduce the total points increasing I am not going to take the 0.75 for a penalty win (still 0.5 for a penalty loss). For cup ties if the game goes to penalties the result before the penalties will be taken.
Expected result is once again a calculation, this being We = 1/(10^ - (dr/600) +1), where dr is the difference in rating between the teams (so for TeamA this would be PBefore TeamA – Pbefore Team B).
Some other changed compared to the official FIFA one is:
1 – No safety for teams after the quarter final. For the FIFA one from the quarter final onwards you can’t lose any ELO, which would introduce the ELO going up over the seasons which is something I don’t want.
2 – All teams start on 1200 ELO, in the FIFA ranking the starting ELO was based on the old system, which I don’t have with this, so I made all teams start on 1200 (also for the teams who were not in the original 6).
So now onto the actual classification
Each team has their own little story, but let’s first look at the strongest team over the seasons. First up we see Cairo being the strongest in S1 (with London very closely behind them). Hollywood however quickly went over them both and was the strongest team in S2, before barely ending up behind London in S3. S4 was once against a Hollywood season with barely behind them Reykjavik (back then Sydney) before in S5 Reykjavik/Sydney went barely above Hollywood. Then the Hollywood domination started, being easily the best team in S6, S7, S8, and S9, before giving it up to Tokyo for S10. Then since S11 Reykjavik has been the team with the highest ELO throughout S11, S12, S13, S14 S15, and S16. However in these final few seasons Schwarzwälder was not far behind, before overtaking Reykjavik and being the team with the highest ELO after S17 (which is where we are right now).
On the opposite side of the coin we have the teams with the lowest ELO, which after S1 was surprisingly enough Hollywood, Seoul followed in S2, before Cairo got there in S3. After that we had a bit of a Montréal “era” being the worst team for S4, S5, and S6 before giving it over to Kaapstad for S7, S8, S9, S10, S11, and S12 where it went to Athenai for S13, and S14, before a Montréal season in S15, and a Paris period in S16, and S17 which is once again where we are right now.
Now onto some fun stats right now.
First off let’s take a look at the biggest improvements between 2 seasons. In 5th we see Schwarzwälder in S15 increasing 201.88 points over the S15 season (compared to the S14 season). Next up is Tokyo in S4 increasing 202.11 points, Reykjavik in S11 starts the top 3 with 203.21 points, then in 2nd we have Buenos Aires in S9 increasing 212.35 points, and finally the biggest improvement over a season was Hollywood in S2 with 231.13 points.
The 5 biggest fallers meanwhile, got in 5th Reykjavik with a 194.33 point loss in S6, North Shore (Red Star Laos back then) with a 194.45 point loss in S7 (although this was their first season), Athenai in S11 and S10 are 3rd and 2nd with a 197.85, and a 200.95 point loss respectively. Finally Sao Paulo in S15 was the biggest point loss, losing 222.67 points in 1 season.
The most stable teams ever meanwhile were in 5th, Buenos Aires in S6 gaining 2.01 points, then in 4th we find Hollywood in S14 losing 1.3 points, Hollywood in S12 comes into the top 3 with a 1.24 point loss, and finally S8 Hollywood finished off this top 5 with a 0.11 point gain over the full season.
Next up what were the biggest gains in a single game (this has been standardized for a importance to 10, so we can compare Friendly/Shield/Minor/Major/Cup game). In 5th we see Athenai winning against London with 2-1 in the S14 season, in the 10th matchday of the minor league, gaining Athenai 8.653 (25.96) ELO. In 4th we go back to S11 where we see Kaapstad beating Paris in the Group C matchday 1 of the cup with a score of 1-0 to gain 8.913 (35.63) points. In 3rd we once again go to Kaapstad, this time a 5-2 win against Buenos Aires, once again in Group C but this time Matchday 2, and the S13 variant, this gained Kaapstad 8.913 (35.65) points. In 2nd we have the S16 friendly between Montréal and North Shore, which was won by Montréal with 5-3, gaining Montréal 8.94 points, and the biggest upset according to this system, is Kaapstad beating Reykjavik with 1-0 in the S13 cup of, once again group C but this time matchday 3 (this was a big upset cupgroup for Kaapstad), gaining them 9.56 (38.24) points. Fun fact, the biggest draw upset was the S10 D2MD12 game where Kaapstad drew against Paris with 1-1 giving them 3.83 (11.49) points.
Now for the 5 most predictable games. In 5th we have S2 D1MD7 were Cairo and Catalunya drew 1-1 with 0.01 (0.04) ELO going to Cairo. In 4th we have S8 D1MD11 were Athenai and Cairo drew 1-1 with 0.007 (0.03) ELO going to Cairo, in 3rd we see Reykjavik playing Montréal drawing 1-1 for 0.005 (0.02) ELO going to Reykjavik in S9 group B matchday 1. Then finally we have a semi-shared 1st place, with Red Star Laos – A.C. Romana in the S7 friendly game 2 being a 1-1 draw with 0 ELO moving, and then in 1st as this one was a bit closer, being the 1-1 draw between Romana and Paris in S11 D1MD2.
For the closest games ever, we of course have the Laos-Romana, and Romana-Paris games mentioned before, both ending in a 1-1 draw with 0 ELO points moving. Of course we also have all S1MD1 games (and 2 S1MD2 games) as teams back then all started on 1200 points, and either had 1180 or 1220 after MD1. Other wins between 2 equal teams, who weren’t teams in their very first game (or S1MD2) are, FC Rio 0 – 1 Tokyo S.C. in S1MD4, Tokyo 2-0 Seoul in the S3 first round leg 1 of the cup. Next up we have S5 D1MD13 between Tokyo 3-0 Athenai. Finally the last equal teams game was the S9 Friendly between Reykjavik 1-0 Montréal. All of these games got 5 points from 1 team to the other (times 4 for the league/cup games).
I hope some people found these stats interesting. Want to check it in more detail, see this google sheet: Google Sheet
First off, how did the ELO come to be. For this I mostly used the ELO system FIFA also uses for its man’s international team ranking. Which means that after a match the new ELO for a team gets calculated by the formula:
P = PBefore + I * (W - We)
P – Points (After the game)
PBefore – Points before the game
I – Game Importance
W – Result of the match
We – Expected Result
P and PBefore are pretty self-explanatory.
The game importance is slightly changed compared to the FIFA scoring, as for FIFA international teams this is heavily build around the world cup (and continental cups). This means that instead of going from 5 (friendlies outside the window) to 60 (FIFA world cup after quarter finals), I have decided on a more easy system of, 10 for friendlies (same as FIFA), 20 for shield, 30 for D2/Minor League, 40 for Cup/D1/Major League/The League.
The result of a match is also slightly different from the actual FIFA calculation, I am keeping the same 1 for a win, 0.5 for a draw, 0 for a loss, however to not introduce the total points increasing I am not going to take the 0.75 for a penalty win (still 0.5 for a penalty loss). For cup ties if the game goes to penalties the result before the penalties will be taken.
Expected result is once again a calculation, this being We = 1/(10^ - (dr/600) +1), where dr is the difference in rating between the teams (so for TeamA this would be PBefore TeamA – Pbefore Team B).
Some other changed compared to the official FIFA one is:
1 – No safety for teams after the quarter final. For the FIFA one from the quarter final onwards you can’t lose any ELO, which would introduce the ELO going up over the seasons which is something I don’t want.
2 – All teams start on 1200 ELO, in the FIFA ranking the starting ELO was based on the old system, which I don’t have with this, so I made all teams start on 1200 (also for the teams who were not in the original 6).
So now onto the actual classification
Each team has their own little story, but let’s first look at the strongest team over the seasons. First up we see Cairo being the strongest in S1 (with London very closely behind them). Hollywood however quickly went over them both and was the strongest team in S2, before barely ending up behind London in S3. S4 was once against a Hollywood season with barely behind them Reykjavik (back then Sydney) before in S5 Reykjavik/Sydney went barely above Hollywood. Then the Hollywood domination started, being easily the best team in S6, S7, S8, and S9, before giving it up to Tokyo for S10. Then since S11 Reykjavik has been the team with the highest ELO throughout S11, S12, S13, S14 S15, and S16. However in these final few seasons Schwarzwälder was not far behind, before overtaking Reykjavik and being the team with the highest ELO after S17 (which is where we are right now).
On the opposite side of the coin we have the teams with the lowest ELO, which after S1 was surprisingly enough Hollywood, Seoul followed in S2, before Cairo got there in S3. After that we had a bit of a Montréal “era” being the worst team for S4, S5, and S6 before giving it over to Kaapstad for S7, S8, S9, S10, S11, and S12 where it went to Athenai for S13, and S14, before a Montréal season in S15, and a Paris period in S16, and S17 which is once again where we are right now.
Now onto some fun stats right now.
First off let’s take a look at the biggest improvements between 2 seasons. In 5th we see Schwarzwälder in S15 increasing 201.88 points over the S15 season (compared to the S14 season). Next up is Tokyo in S4 increasing 202.11 points, Reykjavik in S11 starts the top 3 with 203.21 points, then in 2nd we have Buenos Aires in S9 increasing 212.35 points, and finally the biggest improvement over a season was Hollywood in S2 with 231.13 points.
The 5 biggest fallers meanwhile, got in 5th Reykjavik with a 194.33 point loss in S6, North Shore (Red Star Laos back then) with a 194.45 point loss in S7 (although this was their first season), Athenai in S11 and S10 are 3rd and 2nd with a 197.85, and a 200.95 point loss respectively. Finally Sao Paulo in S15 was the biggest point loss, losing 222.67 points in 1 season.
The most stable teams ever meanwhile were in 5th, Buenos Aires in S6 gaining 2.01 points, then in 4th we find Hollywood in S14 losing 1.3 points, Hollywood in S12 comes into the top 3 with a 1.24 point loss, and finally S8 Hollywood finished off this top 5 with a 0.11 point gain over the full season.
Next up what were the biggest gains in a single game (this has been standardized for a importance to 10, so we can compare Friendly/Shield/Minor/Major/Cup game). In 5th we see Athenai winning against London with 2-1 in the S14 season, in the 10th matchday of the minor league, gaining Athenai 8.653 (25.96) ELO. In 4th we go back to S11 where we see Kaapstad beating Paris in the Group C matchday 1 of the cup with a score of 1-0 to gain 8.913 (35.63) points. In 3rd we once again go to Kaapstad, this time a 5-2 win against Buenos Aires, once again in Group C but this time Matchday 2, and the S13 variant, this gained Kaapstad 8.913 (35.65) points. In 2nd we have the S16 friendly between Montréal and North Shore, which was won by Montréal with 5-3, gaining Montréal 8.94 points, and the biggest upset according to this system, is Kaapstad beating Reykjavik with 1-0 in the S13 cup of, once again group C but this time matchday 3 (this was a big upset cupgroup for Kaapstad), gaining them 9.56 (38.24) points. Fun fact, the biggest draw upset was the S10 D2MD12 game where Kaapstad drew against Paris with 1-1 giving them 3.83 (11.49) points.
Now for the 5 most predictable games. In 5th we have S2 D1MD7 were Cairo and Catalunya drew 1-1 with 0.01 (0.04) ELO going to Cairo. In 4th we have S8 D1MD11 were Athenai and Cairo drew 1-1 with 0.007 (0.03) ELO going to Cairo, in 3rd we see Reykjavik playing Montréal drawing 1-1 for 0.005 (0.02) ELO going to Reykjavik in S9 group B matchday 1. Then finally we have a semi-shared 1st place, with Red Star Laos – A.C. Romana in the S7 friendly game 2 being a 1-1 draw with 0 ELO moving, and then in 1st as this one was a bit closer, being the 1-1 draw between Romana and Paris in S11 D1MD2.
For the closest games ever, we of course have the Laos-Romana, and Romana-Paris games mentioned before, both ending in a 1-1 draw with 0 ELO points moving. Of course we also have all S1MD1 games (and 2 S1MD2 games) as teams back then all started on 1200 points, and either had 1180 or 1220 after MD1. Other wins between 2 equal teams, who weren’t teams in their very first game (or S1MD2) are, FC Rio 0 – 1 Tokyo S.C. in S1MD4, Tokyo 2-0 Seoul in the S3 first round leg 1 of the cup. Next up we have S5 D1MD13 between Tokyo 3-0 Athenai. Finally the last equal teams game was the S9 Friendly between Reykjavik 1-0 Montréal. All of these games got 5 points from 1 team to the other (times 4 for the league/cup games).
I hope some people found these stats interesting. Want to check it in more detail, see this google sheet: Google Sheet