Well there are these guys who just claim objectivity doesn’t exist.
But usually they just hate it because they only want others to concede to their opinions.
I’m glad there is objective data cause with it you can defy these kind of people quite easily.
Objective data can be interpreted subjectively and twisted.
Usually, that’s what happens. I’d rather tell someone “This civ has this unit to do these things which makes this happen” to their “But I don’t like how it feeeeeeeeellls” and win the argument with, y’know, an argument that’s reasonable, rather than giving them “But they’re at 47% so clearly…” and then they just double, triple, quadruple-down on that number instead of making an actual argument.
The only way you get 50% is by making everyone the same. And when you make everyone the same, you make everyone else quit.
Nobody demanded this. So why do you even speak about it?
Because so long as you have winrates, you’ll have people who complain about civilizations at 48% and 52%. I play the game. I don’t need a statistic to agree with me when I think a civ is strong and I certainly don’t need it to tell me a strong civ is weak.
Statistics in a field you can experience, gives the power of the most superficial insight to those with no experience, and little else.
I think saying that winrates are completely useless for anything at all would be a stretch. They just have to be combined with other factors, such as pick rate, win % for specific maps, win % on different time periods of the game, etc.
As experience is only valuable from a certain point in skill. A lot of people complain that XYZ is OP based on their experience, when a lot of the times, it’s just because they don’t have the proper skillset/don’t know the proper counter to something.
So balance should be done on experience?
Probably most devs don’t even play the game at competitive level, so what should we do? Trust whose experience?
Ask one of the pro players? Which pro player though? The current #1 in the leaderboards? The first ten?
Win rate is not something you can base EVERYTHING on, but it’s a starting point, without it, we might just balance things according to the season.
It’s summer, buff Mayans who lived in hot and humid climate.
I prefer to not call you at all, you criticize without offering any viable alternative.
Subjective experience is not a viable one.
From my experience as a moderate noob I regard Chinese difficult to play, so I demant buffs for Chinese.
Is it okay to you? It’s my experience. Should be good. 
Yes, that’s preferrable. You come in as chinese hater and say “onagers are too good for chinese and I need BBC to counter it” and the next person who comes along says “don’t rely entirely on chuks, go for your hussars, they have a very open tech tree and the upgrades are super cheap.”
This is what actually happens, except currently someone who doesn’t like the answer that’s been given says “BUT 47% SO CLEARLY WEAK.”
which is when someone who knows something comes in and gives you an alternative. Experience is always valuable. Ignoring experience is how you miss half the picture. Not confronting experience with further experience is how you allow the truth to lie unfound on any subject.
This is general good advice, but technically in this game any civ more or less has answers for anything, so the logical consequence should be stop any further buff/nerf for good.
Why they nerfed Arambai so hard? Double castle Arambai was too strong? If people just scouted more they could have spotted the double castle in time!
Why were towers nerfed in feudal? Scouts/drush and you should have seen that coming.
You can always find someone who says “everything is fine as it is”, even 1000HP war elephants, because monks exist. On the other side some people would cry that near everything is OP, plenty of example on this forum.
So who is right? OP criers or buffs askers?
Also, win rate has some confidence interval, according to the devs, iirc is 47-52% or something similar, that would encompass almost all civs at the moment, so it can’t really be used as hard proof that something is UP/OP unless the civ is way above/below that, and even then, it’s not automatic.
Good advice is always welcome, but not everyone can put it to use, expecially if noob (and I suspect most players are at least moderately noob).
My archers keep dying to onagers!
“Just micro better.”
My onagers can’t hit anything!
“just aim better!”
I couldn’t run as fast as Usain Bolt the other day!
“Just train more!”
Indeed. I do not disagree with you, but just that experience is also only valuable in certain cases. I remember seeing threads like “Onagers are overpowered” (they are not), “FR is overpowered” (it isn’t), “walls are too strong” (they aren’t), etc.
And yet people have had years sometimes decades to generate such proofs and the proofs or required tests take about an hour. The fact that it’s not common knowledge whether the statements are true or false says a lot about how players make assumptions/don’t do tests and the community as a whole is not very good at disseminating relatively critical information for certain civs. If nothing else it reinforces my point about likely suboptimal play. After all if players don’t immediately know with high confidence whether an alternative is viable or not how can they choose between something and that alternative reliably/optimally?
On an optimal decision making basis players should not be averse to spending stone on walls though. Just because a resource is scarce doesn’t change that its optimal use is determined by opportunity cost. By not spending villager time to mine stone for walls you are leaving it in the ground. This means the opportunity cost occurs in the future, when you mine the last stone.
It requires a model with a grossly unrealistic discount function to prove that this is optimal play or a stone price at the market which is similarly unrealistic.
Consider the following hypotheticals:
The marginal effect of building walls on stone usage was that you lost a castle that would have been finished 10 minutes time into the future. But this has to be discounted for a variety of reasons meaning the economic profit (benefit less cost and opportunity cost) from building the stone walls is still the highest possible profit obtainable from using stone in a wide variety of cases.
Again it’s a statistical thing. It’s possible to generate isolated examples where indeed palisade walls are a good idea and stone walls aren’t. However the assumptions required to have palisade walls be meta and stone walls be used extremely infrequently in general start to violate optimal decision making properties like consistent discounting under risk, marginal cost <= marginal benefits, lowest opportunity cost is best, etc.
I mean we see this play out in high level games all the time. High level players lose a significant fraction of their games due to raids, rather than always due to losing their army followed by an overrun. The expected marginal economic profit (again economic profit just incorporates opportunity cost, it doesn’t have to do with aoe2 economy) of stone walls in many cases is going to be positive yet the player data would suggest it isn’t.
I would say running test can takes years if you need 100% accuracy - take a look on Go, after 2000 years it’s not known the optimal handicap from black to white alone. When we talk about test, we must talk about how much accuracy from optimal we want to obtain from tests. For a complex game like AOE2, there are just too many civs to run tests so that all pairs of civs are tested with high confidence for most of the tests. So the question is not “is it suboptimal play?” but rather “how far can it be from optimal play”?
I’m not talking about civ v civ tests. That’s not really relevant to the point I’m making. I guess I shouldn’t have said optimally but I can clarify that. By optimal play I meant play which is unlikely to be significantly improved upon to a degree such that it affects balance without extremely high search costs.
The point I’m making is that players have to be at the bare minimum taking advantage of a significant % of the advantages which the civ design affords them. Advantages which have reasonable search costs. This is because if you buff a civ with:
You will create an incentive to play the civ which increases the chance that players find the underused advantage. This can effectively create 2 buffs from 1 buff. C.f. Saracens and their archer upon DE release.
Since the things I listed are not-too-hard to find advantages that are likely underused due to the lack of marginal extra use in appropriate contexts it creates problems for data interpretations as the win rates will be biased downward for some civs relative to if the civs were played ‘optimally’. The true optimum doesn’t really matter if it’s exceedingly difficult to find or unlikely to affect balance over the slightly suboptimal play.
I perfectly agree with op, experience is a very good way to know which civs have problem for multiple reason imo:
First, it’s completly fine to have confroncted experience, it creates debates and opinion which is the best thing to know if a civ should be changed and how, while the number well not really…
Second, even if a civ have good winrates and are not op or weak still certain bonus or mechanic are not healthy for the game (cumans’s 2tc , flemish revolution, first cruisade etc…)
Obsidian arrow wasn’t changed because it was op just because this mechanic was completly bad for the game especially in tg.
The only problem i found with experience is the person which make these discussions should have really experience about that and not just see viper doing an exotic strat which worked very well and so ask a nerf.
I think that’s completly hypocritical to say that a 800 elo could know how for exemple fix franks being op in tg while not nerfing them in 1v1 when they are balanced. And not because that is elitist just because if someone is 800 elo, he doesn’t know most basic and so cannot have knowledge deeper enough to make balance changes (and the argument which it’s possible that player can have a good theorical knowledge and just bad execution of it is very bad, knowledge is by a very large portion maked by experience and if your execution is enough bad to be 800 elo then you cannot have enough knowledge that’s completly impossible), i see it a lot here and the discussion is not really possible sometimes because of the not enough high knowledge and that’s very hard to know yourself when you have enough or not knowledge about it, so its leads of balance changes which have absolutely non sence like teutons’s buff on tc range or reducing elite elephant archer price by 62%.
If balances changes discussion for poles is made i cannot really make opinion of it because my opinion changes every time i played it , this is not stable nor even determinated enough so yes it’s better yourself to know what you’re talking about before talking about.
Again, this is not elitist, that is really a thing which exist and not only for low elo as for myself, but peoples should question themself before making balance suggestions before saying thing like “britons op”
Did you really adaptating well ? Is is possible you loose all the time because of a mistake you make yourself and not because the thing is op ?
For example eagles on EW is too strong because no matter that you try to do you’l be in trouble but there are not on rm when you have time to react and make something to handle it like longswords or cavalry archers or castle and uu for a lot of civ like ethiopians, malay, malians etc.
A thing is not op when that’s strong, only when a thing is un dealable correctly in too many cases like the cumans on realease whith their undealable steppes lancers.
I think was both OP and bad design, I can recall Viper using it at Arena in one tournament and bypassing walls and even castles, was stupid.
I agree, tho I dont think sarcaren markets should be used as farm replacement… its pretty easy to see that the cost function is convex, therefore you should seed X farms to archieve optimium, while X is highly unlikely to be 0.
Great posts!
I think devs are doing a great job balancing the game overall.
We don’t really care about the optimum number of farms for using the Saracen market. We care about the domain over which it is better than the alternative which is almost guaranteed to be farming. This is usually nearly completely determined by the price of food at the market. So usually as long as the price is below some price P, one can obtain a benefit by buying food rather than farming. Sometimes to meet constraints you will need to seed a few farms in feudal age. Still doesn’t change the break-even price.
You will have a tough time proving that it’s optimal to farm over using the Saracen market while P is relatively low (like 140 or below) without making mistakes like ignoring the wood cost of farms.
its about 160 in castle age. Buying food below that gives you a benefit. And the bigger the difference the more you gain from it. So it’s easy to see why the saracen market is such a strong eco bonus - if you use it to it’s full extend you gain a ressource equivalent of about 400-500 w. other civs can only get about 100-150 w equivalent with full market abuse. (This doesn’t even include skipping the farm upgrades)
And ofc this is somewhat problematic to balance cause only a small margin of players is capable of cheesing it to that levels, but the benefit of doing so is quite high.
It’s just difficult to oversee that you claimed saracens don’t need to place farms in optimal play which I highly doubt… or rather, say something vague is optimal
I never claimed that Saracens don’t need to place farms in optimal play. You are welcome to try to find the quote where I said such a thing.