[KotD] Clash of the Civilizations (Stats)
-
Thread starter
Rorarimbo
- Start date
jojkovusing a parser
Counting them by hand would be overkill
Lanchiusing a parser
Can the parser count total time of the villager working/walking/idling ?
Some more useless stats for numbers addicts 
:
Opponents
[linethru][/linethru]
Of the golden four contenders to the throne of the desert,
_TheMax_ arguably had the toughest route to the semifinals. The combined average of the current and highest ratings reached by his opponents reaches a total of 2435.
Hearttt then follows with an opponent combined rating average of 2391 while 
TheViper and 
Liereyy had it "easier": their opponents averaged 2331 and 2314 respectively.
Considering current rating only (recorded on this day 06-12-17), _TheMax_ still had the toughest opponents with an average of 2362, followed by Hearttt (2310), Liereyy (2291) and TheViper a distant fourth (2236).
Civilizations
[linethru][/linethru]
The 12 best of 5 matches played by our four players led to 50 games played out of a possible 60. Of a total of 31 possible civilizations, only 14 were selected by the players during the captain mod civ draft.
Liereyy and TheViper picked 11 different civilizations during his 3 drafts, while Hearttt and _TheMax_ follow closely with 10 different civilizations each.
Aztecs, Indians, Berbers, Burmese, Huns and Malians were picked at least once by every player.
Civ choices were mostly similar, but Hearttt introduced a bit of variety. He was the only one to pick Franks (and won with them) as well as the only one never to pick Britons or Malay.
When it comes to actually playing, classic AOC civilizations certainly retain a lot of appeal for these four guys. Aztecs and Huns were the most often played civs, with 7 and 6 games respectively. Of the expansions civs, Berbers (played 6 times) were the most popular, followed by Malians and Malay (5 games).
Three civilizations got a perfect 100% winrate with multiple games played. The great winners are Indians with 4 wins out of 4 games, followed by Burmese and Incas (3/3). Aztecs and Huns remain fearsome when played by these four guys, scoring respectively 85.7% (6/7) and 83.3% (5/6) wins.
Malay have seem to be a weaker tier civilization during the tournament and even the four semifinalists of the tournament have not been so successful with them, winning only 1 in 5 games (20% winrate).
Both Liereyy and TheViper have been playing around with different civilizations and neither of them picked one specific civilization for each of their three matches. However, Hearttt has been faithful to Aztecs during the whole tournament (going 2-1 with them) and _TheMax_ to Malay (1-2).
Time
[linethru][/linethru]
The four players spent a combined time of 33 hours, 41 minutes and 54 seconds (game time) fighting in the sands of Arabia to keep their chances of lifting the coveted trophy alive. Having faced off with arguably the toughest opposition, _TheMax_ has played the most games (15) and spent the most time playing (10:49:27) of them. TheViper was the fastest, qualifying for the semifinals playing only 10 games for a total time of 7:18:18.
It should however be noted that TheViper's games were the longest in average, with an average duration of 43 minutes and 50 seconds. TheMax is not far behind with games averaging 43:18 game time while Hearttt played the fastest games of them four at 37:05 in average.
Hearttt was also the only one never to play a game that went beyond 1 hour of game time. His longest game, the fourth one of his quarterfinal encounter, finished with 0:48:50 on the clock. Liereyy and TheViper both played one 1h game while _TheMax_ seemed to enjoyed them so much that he played a total of three - one per set - emerging victorious twice.
The longest game played by these four guys was played (and lost) by Liereyy during his Ro16 match at 1:11:09. The shortest was played (and won) by Hearttt, also during the Ro16, at 22:43. It was also the shortest set of games with a total duration of 1:43:31. The longest set was _TheMax_' quaterfinal victory which required almost two more hours, finishing with a total of 3:45:33 game time.
Various
[linethru][/linethru]
Hearttt was the only semifinalist scoring a 3-0 victory, a 3-1 victory and a 3-2 victory. TheViper was the only one to score two 3-0 victories.
_TheMax_ played the most games where both players reached the Imperial Age, with 6 games out of 15 (40%). TheViper played 5 such games, but only played 10 games (50%). Liereyy had the lowest part of Imp games, with only 2 of his games featuring both players reaching the Imperial Age.
TheViper reached the Castle Age in all his games, while both Liereyy (who won) and Hearttt (who lost) played one game without advancing further than the Feudal Age. _TheMax_ had two such games, losing one and winning one.
:KOTD ROAD TO THE SEMI FINALS
Opponents
[linethru][/linethru]
Of the golden four contenders to the throne of the desert,
_TheMax_ arguably had the toughest route to the semifinals. The combined average of the current and highest ratings reached by his opponents reaches a total of 2435. Hearttt then follows with an opponent combined rating average of 2391 while TheViper and Liereyy had it "easier": their opponents averaged 2331 and 2314 respectively.Considering current rating only (recorded on this day 06-12-17),
_TheMax_ still had the toughest opponents with an average of 2362, followed by Hearttt (2310), Liereyy (2291) and TheViper a distant fourth (2236).Civilizations
[linethru][/linethru]
The 12 best of 5 matches played by our four players led to 50 games played out of a possible 60. Of a total of 31 possible civilizations, only 14 were selected by the players during the captain mod civ draft.
Liereyy and TheViper picked 11 different civilizations during his 3 drafts, while Hearttt and _TheMax_ follow closely with 10 different civilizations each.Aztecs, Indians, Berbers, Burmese, Huns and Malians were picked at least once by every player.
Civ choices were mostly similar, but
Hearttt introduced a bit of variety. He was the only one to pick Franks (and won with them) as well as the only one never to pick Britons or Malay.When it comes to actually playing, classic AOC civilizations certainly retain a lot of appeal for these four guys. Aztecs and Huns were the most often played civs, with 7 and 6 games respectively. Of the expansions civs, Berbers (played 6 times) were the most popular, followed by Malians and Malay (5 games).
Three civilizations got a perfect 100% winrate with multiple games played. The great winners are Indians with 4 wins out of 4 games, followed by Burmese and Incas (3/3). Aztecs and Huns remain fearsome when played by these four guys, scoring respectively 85.7% (6/7) and 83.3% (5/6) wins.
Malay have seem to be a weaker tier civilization during the tournament and even the four semifinalists of the tournament have not been so successful with them, winning only 1 in 5 games (20% winrate).
Both
Liereyy and TheViper have been playing around with different civilizations and neither of them picked one specific civilization for each of their three matches. However, Hearttt has been faithful to Aztecs during the whole tournament (going 2-1 with them) and _TheMax_ to Malay (1-2).Time
[linethru][/linethru]
The four players spent a combined time of 33 hours, 41 minutes and 54 seconds (game time) fighting in the sands of Arabia to keep their chances of lifting the coveted trophy alive. Having faced off with arguably the toughest opposition,
_TheMax_ has played the most games (15) and spent the most time playing (10:49:27) of them. TheViper was the fastest, qualifying for the semifinals playing only 10 games for a total time of 7:18:18.It should however be noted that
TheViper's games were the longest in average, with an average duration of 43 minutes and 50 seconds. TheMax is not far behind with games averaging 43:18 game time while Hearttt played the fastest games of them four at 37:05 in average. Hearttt was also the only one never to play a game that went beyond 1 hour of game time. His longest game, the fourth one of his quarterfinal encounter, finished with 0:48:50 on the clock. Liereyy and TheViper both played one 1h game while _TheMax_ seemed to enjoyed them so much that he played a total of three - one per set - emerging victorious twice. The longest game played by these four guys was played (and lost) by
Liereyy during his Ro16 match at 1:11:09. The shortest was played (and won) by Hearttt, also during the Ro16, at 22:43. It was also the shortest set of games with a total duration of 1:43:31. The longest set was _TheMax_' quaterfinal victory which required almost two more hours, finishing with a total of 3:45:33 game time.Various
[linethru][/linethru]
Hearttt was the only semifinalist scoring a 3-0 victory, a 3-1 victory and a 3-2 victory. TheViper was the only one to score two 3-0 victories. _TheMax_ played the most games where both players reached the Imperial Age, with 6 games out of 15 (40%). TheViper played 5 such games, but only played 10 games (50%). Liereyy had the lowest part of Imp games, with only 2 of his games featuring both players reaching the Imperial Age. TheViper reached the Castle Age in all his games, while both Liereyy (who won) and Hearttt (who lost) played one game without advancing further than the Feudal Age. _TheMax_ had two such games, losing one and winning one.
smarthy_
swissboyyusing a parser
Can the parser count total time of the villager working/walking/idling ?
Unfortunately not. The recorded games only contain the events generated by the players, e.g. click research, move unit, queue unit, etc. Everything else is computed by the aoe2 engine. It would not even be possible to simply simulate this, because one would need to implement everything, including pathing, destroying buildings, killing units, ...
Be careful for spoilers the matches sheet includes results including the semifinals.
The file from the first post has been updated (only waiting for ELO update). I have also added Draft Matrix sheet. It includes a preference coeficient. The lowest the number is the earlier the players picked the civilisation. Berbers are the best in that regard 2.25 followed by Indians 2.59 and Malians 3.03.
It looks there is no correlation between how prefered a civilisation is compared to its actual win ratio. The Aztecs 70% win ratio, The Huns 69% win ratio, the Incas and Britons 67% win ratio have 5.55, 7.53, 6.67 and 7.6 preference coeficient. At the same time the most prefered civ The berbers (2.25) has 43% win ratio. As the tournament progresses there are some trends that change in preference but its hard to evaluate how much is dew to the bans.
Also in the overall standings 4 civilisations have 65 points in the top ( Aztecs, Indians, Mayans, Berbers) each of them has chances to win before the Grand Final. It is interesting to know each of them gains advantage in different aspect. The Aztecs have the most number of wins ( Strongest). The indians have been selected in drafts the most ( Most Wanted), The Berbers have been played the most (Most Popular) while the Mayans have been banned the most compared to the other 3 (Most Feared - only Incas being banned more times). This means should the formula for calculating points be different each of them could be ahead.
Once again as said above its obvious there are civilisations that are prefered more while its other civs that perform better in games. This can be seen in all performance related tables, too - The soccer table and the ELO compared to the Draft Matrix.
The file from the first post has been updated (only waiting for ELO update). I have also added Draft Matrix sheet. It includes a preference coeficient. The lowest the number is the earlier the players picked the civilisation. Berbers are the best in that regard 2.25 followed by Indians 2.59 and Malians 3.03.
It looks there is no correlation between how prefered a civilisation is compared to its actual win ratio. The Aztecs 70% win ratio, The Huns 69% win ratio, the Incas and Britons 67% win ratio have 5.55, 7.53, 6.67 and 7.6 preference coeficient. At the same time the most prefered civ The berbers (2.25) has 43% win ratio. As the tournament progresses there are some trends that change in preference but its hard to evaluate how much is dew to the bans.
Also in the overall standings 4 civilisations have 65 points in the top ( Aztecs, Indians, Mayans, Berbers) each of them has chances to win before the Grand Final. It is interesting to know each of them gains advantage in different aspect. The Aztecs have the most number of wins ( Strongest). The indians have been selected in drafts the most ( Most Wanted), The Berbers have been played the most (Most Popular) while the Mayans have been banned the most compared to the other 3 (Most Feared - only Incas being banned more times). This means should the formula for calculating points be different each of them could be ahead.
Once again as said above its obvious there are civilisations that are prefered more while its other civs that perform better in games. This can be seen in all performance related tables, too - The soccer table and the ELO compared to the Draft Matrix.
misafecoELO is here:
1. Aztecs 1692
2. Huns 1662
3. Incas 1659
4. Burmese 1634
5. Mayans 1627
6. Britons 1623
7. Indians 1622
8. Spanish 1614
9. Mongols 1606
10. Malians 1602
11. Goths 1599
12. Celts 1594
13. Vikings 1584
14. Japanese 1571
15. Berbers 1564
16. Magyars 1556
17. Italians 1550
18. Vietnamese 1532
19. Ethiopians 1529
20. Malay 1502
Less than 2 games: Koreans, Franks, Byzantines, Chinese
1. Aztecs 1692
2. Huns 1662
3. Incas 1659
4. Burmese 1634
5. Mayans 1627
6. Britons 1623
7. Indians 1622
8. Spanish 1614
9. Mongols 1606
10. Malians 1602
11. Goths 1599
12. Celts 1594
13. Vikings 1584
14. Japanese 1571
15. Berbers 1564
16. Magyars 1556
17. Italians 1550
18. Vietnamese 1532
19. Ethiopians 1529
20. Malay 1502
Less than 2 games: Koreans, Franks, Byzantines, Chinese
I also updated my probabilistic model. It can be found in the last column of the civ stats sheet by the way.
Here are the results:
AZTECS 10.72%
INDIANS 9.80%
MAYANS 9.62%
BERBERS 9.25%
BURMESE 7.07%
MALIANS 6.63%
INCAS 5.74%
HUNS 5.64%
MALAY 5.46%
MAGYARS 4.31%
BRITONS 3.79%
ETHIOPIANS 3.56%
CELTS 3.03%
BYZANTINES 2.69%
SPANISH 2.26%
MONGOLS 2.23%
KOREANS 1.62%
GOTHS 1.54%
CHINESE 1.10%
FRANKS 0.99%
VIKINGS 0.98%
VIETNAMESE 0.66%
ITALIANS 0.66%
JAPANESE 0.66%
I find it quite interesting how it differs from ELO. Both models take into account the strength of the enemy. I do think my model is able to cope better with less data but ELO might get better if all civs have played a significant amount of matches.
Here are the results:
AZTECS 10.72%
INDIANS 9.80%
MAYANS 9.62%
BERBERS 9.25%
BURMESE 7.07%
MALIANS 6.63%
INCAS 5.74%
HUNS 5.64%
MALAY 5.46%
MAGYARS 4.31%
BRITONS 3.79%
ETHIOPIANS 3.56%
CELTS 3.03%
BYZANTINES 2.69%
SPANISH 2.26%
MONGOLS 2.23%
KOREANS 1.62%
GOTHS 1.54%
CHINESE 1.10%
FRANKS 0.99%
VIKINGS 0.98%
VIETNAMESE 0.66%
ITALIANS 0.66%
JAPANESE 0.66%
I find it quite interesting how it differs from ELO. Both models take into account the strength of the enemy. I do think my model is able to cope better with less data but ELO might get better if all civs have played a significant amount of matches.
Yeah, that's really quite interesting
Hi, I quite like the ideea of the ELO civ ratings. We could make it independent of the order of games played if we did the ratings updates in batches.
Meaning, asume all civs have rating 1600 in R1 and make all the updates as if they were always winning and losing against ELO 1600s.
In R2, we use the ratings after R1 updates for all the matches and so on
Meaning, asume all civs have rating 1600 in R1 and make all the updates as if they were always winning and losing against ELO 1600s.
In R2, we use the ratings after R1 updates for all the matches and so on
I think that has been suggested before but I don't think it would help. R1 has the highest number of matches. If we calculate all matches with rating 1600 (or at least something between 1580 and 1620 after R1), we would reach something that is almost the same as the win ratio (e.g. final rating ~ 1600 + 15 * (overall wins - overall losses)).
Elo is supposed to consider the previous strength of the civs, a civ has to lose fewer points when losing to Indians than when losing to Byzantines. Therfore you would want the civs to change their ratings as soon as possible to obtain something that is at least somehow meaningful after the tournament.
It's true when you consider that between rounds, experts could have reconsidered their choices of civs depending on how well they fared at the previous round.
However, within the same round, it does not make as much sense since matches are being played in arbitrary order (due to scheduling). So I think an averaged elo would be better than something that is being updated after each game.
However, within the same round, it does not make as much sense since matches are being played in arbitrary order (due to scheduling). So I think an averaged elo would be better than something that is being updated after each game.
Yes, and the Elo is calculated based on the actual order of these games. If a game is played on Tuesday, players that play on Saturday can indeed react to that as well.
(Also I agree with misafeco that it will not be a huge difference probably)
Can we change the Finals settings to make sure these 4 civs are played so that they can be added to the ranking? 8-)
What about those poor 7 remaining civs, which weren't played at all?
It's not just the ES, you can add EW and EEW too. Meso civs trained almost 100 of them per match on average.
AndreasCan someone tell me why malay score so bad?
I would expect them to be one of the best civs: 4 extra vils if u reach fuedal cause of the faster age bonus,
and OP unique unit. Even the tech which makes champ line not cost gold anymore is OP.
Why did they perform so badly? 11
I would expect them to be one of the best civs: 4 extra vils if u reach fuedal cause of the faster age bonus,
and OP unique unit. Even the tech which makes champ line not cost gold anymore is OP.
Why did they perform so badly? 11
Well Cysion needs to do something about them, and the civs with less than 2 games to some extent. And Vietnamese also. They suck.
All civs being balanced on 1v1 arabia is a must. At least 21 out of 31 are fine now
Time
Your time
G M T
Your zone
AoE Live-Streams
Voobly Top 5 RM 1v1
MUHAMMEDD DE Top 5 RM 1v1
Hera 2644
_Barles_ 2639
Villese 2617
GL.TheViper 2602
ACCM |AOEbuilds.com 2594
Voobly Top 5 RM Team Game
DE Top 5 RM Teamgame
2583
痛大师 2245
HGB_AOE 2181
chaos_2_win 2009
正义的威震天 2008
Voobly Top 5 DM 1v1
Voobly Top 5 DM Teamgame
DE Top 5 Empire Wars 1v1
[aM]_MbL40C_ 2053
ACCM |AOEbuilds.com 2031
Hera 1999
Yo 1964
CDUB.dogao 1935
DE Top 5 Empire Wars TG
mYi.Sitaux 1752
__BadBoy__ 1726
ELEOS | ElNoniro 1690
Lauth3 1687
BlackRock 1668
