It’s not, atleast not with the current available data
2 Likes
It could be if you have the correct data, and the knowledge to gather, verified, interpret and validate them, otherwise you have to trust others to do this for you, or don’t use them at all.
If someone doesn’t have those knowledges you still can suggest balances, just not based on statistics, but based on your opinion, ideas, experience, not numbers interpreted every time in different ways.
1 Like
Did you even read what I have written?
1 Like
New statistics from aoestats: https://aoestats.io/RM_1v1
The fact that how much different these and the aoe-tc statistics are, shows that we cannot trust either yet.
1 Like
They are actually very similar, however.
Thank you for more stats. I love stats.
1 Like
2-5% differences is not similar in my opinion
Note that even if they were similar we ciukd’t have trusted them.
1 Like
They still have Goths and Indians at over 50% winrates, among other things.
1 Like
Some similar while some not.
1 Like
Again, in there aren’t similar number in statistics, there is a process to see if the numbers are the equals or different.
You have to set an hypothesis.
H0: ua=ub
Then you can proceed to verify that, but to do so you need more data that just a win rate.
Also the problem is the same about the sample, if for example the Italians sample is only (or mostly) water maps, you can’t say that those wins rate is rappresentative of the Italians on land maps.
1 Like
It is one way but certainly not the only way. For the past years balancing has been done according to feedback from (mostly top) players and some casters. This is for a good reason: You expect these people to be able to play civs to their full strenghts. This way you can eradicate random factors as much as possible. Of course, even top players have their particular approaches to playing the game and different levels of experience for different strategies, so we’d need to oberserve games over a longer period of time.
This is where data comes in. Beyond a certain amount of games we cannot go through all of them manually and sum up our experiences of what has proved to be stronger or weaker civs so we start to scale up data (numerical information). That enables people to compare larger amount of games. Doing that, of course, we need to make sure that we understand what these data express.
That’s why a lot of people correctly mentioned the need for proper interpretation. When we get data such as win rates, the idea is to trace these data back to mechanisms in the game that explain why one civ comes on top of others. Example: Franks have been reigning supreme the voobly stats for a long time which could be explained by how they play out (bonuses harmonize well towards certain strategies, intuitive game plans, …) and data context (voobly 1v1 stats primarily stemmed from 1v1 arabia games where Franks are stronger compared to closed or water maps).
So in order for data to be meaningful we need to make sure to make up a plausible correlation between what happens in-game (how civs play out) and what’s the game’s result (loss or win). Of course, the best large-scale way to do that would be to further put the game process into data such as comparing villager amounts, eco stats, unit numbers, researched techs, age uptimes, K/D, and much much more at different time stamps in the game. This is of course not possible unless you are a big company (and MS wouldnt profit from that, at all) or a research institution (that usually do not happen to be interested in optimzing one particular video game). It still wouldn’t be ideal because such a representation cannot understand why things are precisely happening as they are but it gets close enough, I assume.
Data doesn’t speak for itself, it is a representation of something, in this case the strenghts of civs during a match. Without tying back game results to the actual game process we simply can’t know whether a civ, for instance, has weak results because of civ balance, things like path finding, players not having explored the civ enough, or just having a bad matchup against 1 or 2 two civs that are picked a lot more than others. If we have good volumes of data and sound ways to let them explain what we are interested in, they are a nice tool for making arguments why civ a) appears stronger than civ b). If we don’t data are just some numbers that can be fun to look at.
2 Likes
Also, if we consider the leaked patch notes, it seems that they will decreas the FS disc. from 20% (now incorrecly stted 15%) to 15% (so from 60w to 64w), basically now you get 5 FS for the price of 4, with 15% it would be 6 for the price of 5, and for the dark age is a significant nerf.
My initial idea was to make 15% in dark age and 20% in feudal, but I dont’ think it would be that way.
So another way to compensate for this nerf would be to give them back (on top of their 15% disc.) their old bonus of +2LoS (could be even 3) on their FS, making them at least a bit more efficient (but still not on pair with japs FS or galleys).
They left the technology discount as it its so there is no need to compensate for this nerf.
2 Likes
Because the Devs already see that Italians are doing just fine, and are perhaps a bit OP, rather than weak and needing buffs.
1 Like
Btw I realized why GC are slwoer to train than other foot archers: imagine you’re playing random on a water map, and you got let’s say Franks, or Magyar. Of course your opponent, being the tryhard that he is, picked Italian. Trying to outplay him on the water doesn’t look like the best option. However, a surprise cavalry landing would give you a much better fighting chance. But now imagine if the Italian could make genoese xbows too fast, your strat would have a more narrow opportunity window to be successfull.
2 Likes
Is the update already out? Because in leaked ones they fixed it to 15%, that’s why I suggested to give them again the +2LoS.
Yeah I get it, you think that they are OP, I don’t, maybe the devs will decide to buff them, or to nerf them, either way I will still express my opinion about it.
Yes, after I spawned a castle out of the air, and even assuming that I already have one (which Italians don’t need one until late game for trebs) in a minute, with the buff I suggest (16s for a GC) it means that I can create 4 of them, in by the time that I can find and kill the enemy cavalry (that is way more faster) they can still do a lot of damage.
Yeh, and they didn’t mention any intention to nerf the technology discount.
That’s why I said “smaller window of opportunity” because of course all this is irrelevant if the landing occurs before a castle is up. It’s just that the odds are already hard enough for the non water civ guy (especially to sneak against your navy) imo they don’t need an even stiffer obstacle to overcome.
You’re on an island, so they will have less room to run away. And it’s not like you should have several TCs by then to shelter your eco. You also hopefully have more fishing ships and thus more food to replenish your vill count. You’re already advantaged for that, you really don’t need even more.
2 Likes
Can you link it, I couldn’t find it.
The original link was shut down but the changes have been reposted in other places such as here https://www.aoezone.net/threads/early-patch-notes-leaked.164854/
2 Likes
If the castle isn’t already up where is the problem, if I want to counter a cavalry raid I would probably use pikes in this context.
In fact faster GC would give them no advantage in this case.
It’s still the same one leaked, and there still is the nerf.
Now FS cost 60w (so 20% disc. no 20%), now they will fix it, so they will become 64w.