Great site!
Based on my win rates we need to buff Persians and Khmer 11
That is the issue, your win rates mean nothing. Global winrates are the indicator.
Exactly. Itâs almost ridiculous. How can you say that persians and khmer are average? They are sick civs.
Inlcuding britons and chinese.
Not what the data shows. Khmer and Britons are average, with Chiniese and Persians being a bit stronger, but not over 55% win rate.
The data is too general. I mean noobs say teutonic knights are sickâŠ
There should be a data only for pros to really know the real and good win rates.
There is, go to 1600+
You know ingame taunts? What does 11 means? 
I am joking. I know both civs are OP.
2 Likes
But they are not, and what you âknowâ is anti-scientific, emotional beliefs.
I guess we need to wait a couple of months until the data get real informative value - the data only considers games played from April 1 on. In the 16+ list we currently have less than 8000 games played. So a civ with a pick rate of 1% has been analyzed based on 80 games. So Goths being top 5 here actually means very little. I really like the site, as well, but I donât think there is any reason to let these data influence balance for the upcoming patches.
4 Likes
I deeply agree on this.
Youâre not wrong but I think with Goth something is different, they are supposed to be so, so bad that basically they will get civ winâ d to oblivion by anything remotely meta. Just by looking at the bonuses there is no way they should be better than other rare civs like Teutons, Viet or Ethiopian.
Interesting data. Keep in mind that this reflects less than a monthâs worth of games, so best not to draw sweeping conclusions about civ balance, especially if you canât tell me why these results are the way they are. The meta will change over time, and I fully expect that many of these civs will shift places, some dramatically. The previous data from HD had months or years worth of games, so letâs not pretend that a few weeks of info carries the same weight.
That said, I am also surprised to see the Goths ranked so highly at the moment, considering theyâve gotten nothing but nerfs since AoC, and the only buff theyâve received that might make a difference is their staggered infantry discount that now begins in Dark Age. My hypothesis is that, while that bonus has made the Goths a little stronger in the early game, the results we are seeing are largely due to what Iâll call âthe Hoang effect.â Think about Hoangâs early success with the FDrush - Did he win because his strategy was inherently superior to the contemporary meta? Not at all, but rather, his ability to execute an uncommon, aggressive early game strategy caught a lot of players off guard. Over time, however, players learned to deal with this (see: Daut vs. Hoang, among others), and Hoangâs strategy, rather than redefining the meta, was relegated back to a niche, bordering on obscure strat, like the Inca vill rush, Lithuanian insta-drush or the ancient Saracen monk rush. I believe a similar thing is happening with Goths - their early game is catching players off guard (like it did in the previous update, but to a lesser extent), and many players donât yet know how to deal with early aggression beyond the standard drush, or otherwise havenât taken the proper precautions against early game Goths. Granted, the Gothsâ Dark Age might be a little more user-friendly than some of the aforementioned strats, but I believe once players become used to it, the Gothsâ winrate will plummet to average at best.
Now, I could be wrong, and time will tell, but this seems to be the only way to reconcile Gothsâ apparent success here with the high-level consensus of them being trash tier (along with my own experience).
So this is a good start, but it really doesnât tell us much thatâs definitive, except
regarding the civs with abysmal win rates. A win rate of 1 or 2% higher than average isnât worth freaking out about, but civs that are near 10-20% below average are clearly not great for 1v1s. Patience is needed to make solid conclusions about the relative power levels of most civs.
Edit:
Right on. I know we get a lot of reactionary, âshoot from the hipsâ balance suggestions that are based on fleeting experiences or minimal data (e.g. Tatars are useless because their hill bonus is too situational! Vs. Nerf Tatar hill bonus, itâs too strong! - A week later), so best not to draw wild conclusions based on limited data.
2 Likes
Are you ever going to disappoint me 11 Tbh your take is what makes the most sense right now. Also Iâve thought some more about the whole thing and maybe seeing these surprising results will make people want to play the top 1600+ civs more, and thus we might end up with some kind of auto-fullfilling prophecy where Celts/Franks and the like remain on top.
1 Like
Using the law of large numbers, and calculation with the bernuilli process, assuming the number of matches played and the number of wins is reliable data, the follwing is true:
There is a 6,15% chance that the probability of goth winning a game is 53,0% as the stats say.
If youâre interested theres a 4,57% chance that the probability of goth winning a game is 50%, and theres a 0,72% chance that the probabily of goth winning a game is 45,75%, as aoestat says.
sources: https://en.wikipedia.org/wiki/Law_of_large_numbers
https://en.wikipedia.org/wiki/Bernoulli_process
Of course these chances are still not perfectly accurate, since each match doesnât have the same chance of goths winning (depends on map, and enemy civ).
With this I just want to show that the winrate is NOT equal to the probability of goths winning a match, but not anything more, do not use these stats for concluding anything more than that, because it might not be usable for that.
1 Like
It is the rate at which they win matches, however, which is the best rate for balance discussion.
The game must be balanced around the probablity of civilizations winning. Every civ has to have around 50% of this probability, because thatâs when the game becomes perfectly balanced. But let me emphasize, this is NOT the same as the winrate (stats only show winrates).
1 Like
Every civ having 50% is unrealistic. i would be satisfied if every one of the gaging civs had at least 45% winrate.
Explain the difference. It is not clear to me what you means.
For example, 1 game franks versus britons. The briton player wins⊠Their winrate is 100%, frank winrate is 0%. However, the probability of britions winning is definitely not 100%, franks can win aswell, it is just this time they couldnât.
With coinflips:
The probability if getting head in a PERFECT coinflip is 50%. However, the percentage of heads you got (for example in 10 you got 6 - then its 60%), is not the same (this is winrate in aoe)
1 Like
By a large amount of games we can assume that the winrate is almost equal to the probability. If we would add confidence intervals, they wont be big. So current winrate are a pretty good estimate for the probability of winning. There isnt really a difference in the outcome assume a large amount of games. It is not like Goths has winrate of 53,4%, but there probability of winning is about 40%. The difference is pretty small.
1 Like