Food for Thought

Age of Empires II: DE II - Discussion
1

These questions are just food for thought.

  • Is 50% accuracy (keeping in mind that accuracy starts at 100% at 2 range and drops off to around 65% at 7 range; 50% is the asymptote) the optimal (in terms of balance and unit design) default accuracy for heavy cavalry archers? Why not 60% or 40%? Think of how it affects civs like the goths or khmer.
  • Are stone prices for buildings optimal? Does the revealed preference of players toward castles and away from stone walls and towers inform us of something or is it simply just a reflection of habits or preference? Why not have towers cost 75w 100s? Why not 25w and 150s?
  • are stone mining techs really optimally priced? What about the default stone mining rate?
  • are pike and skirmisher upgrades optimally priced given the units inflexibility and design goal of being cost effective counters? Consider the counterplay available from militia, siege, poorly upgraded xbow, and freely available knights.
  • is the training time difference of 4.5 res/sec and 2.7 res/second between knights and pike appropriate given the role and relative inflexibility of pikes?
  • are mills, farms, and wheelbarrow optimally priced given the dynamics we observe with food costs and food heavy units between feudal and late castle age? What makes having mills cost the same as lumber camps good design? What makes 60f the optimal price for farms?
  • given that unit line costs stay the same across time, is it really prudent to have significant divergence between food and wood/gold economy in the mid game relative to dark age and imperial age? Doesnt this imply that actual villager time costs also diverge in the midgame?
  • was supplies at -15f really necessary in the late game or was its a band-aid to fix the food divergence above? If such divergence was not present what would the appropriate discount be?
  • what differences between units or civs are necessary to prevent 30 or 40% price differences in the same unit line from being overpowered? Or from forcing the non-discounted version from being underpowered? Does the blacksmith and UT parameter space even allow such balance to occur? Keep in mind upgrades and unit combinations are not free but discounts are usually provided freely.

Meta questions:

  • given that any of these parameters need to be changed what is the most appropriate updating proceedure for said change? Large discrete changes? Forward guidance to players on direction followed by frequent small changes in that direction?
  • What relation exists betweent the answers to these questions? Between the answers to these questions and other game parameters?
  • what do the civ bonuses which affect these parameters imply regarding sensitivity of the game toward these values?
  • what are risks associated with getting any of these questions wrong? What might occur if one begins hardening these decisions by optimizing conditional on these specific parameter values? How large of a difference might develop between the “perfect” AoE2 game and the local maximum being moved toward via naive optimization.
  • Can gradient descent with respect to win rate or civ balance totally ignore these questions and not run the risk of killing certain important game design? How would one identify if such risk were being realized?

Meta meta question:

  • What discipline or profession would be best suited to answering these types of questions? Where might combinations of knowledge be required to tackle this problem?

For those thinking of answers:

  • How many answers to the above end up being “I dont know the answer but I’m too afraid to touch it?”
  • What does uncertainty with respect to answering these questions imply about knowledge of the game and its dynamics? About the ability to optimize the game balance in general?
  • Is there a way to link these answers in a way that makes answering the entire collection much easier?
8 Likes
2

Love this bizarre post.

4 Likes
3

Big Napoleon dynamite stretch and warm up. My safe word is WHISKEY

I think it’s too low. And I think CA power ceiling is too high.

Iaw CA should be easier to use earlier/smaller numbers, but more difficult to reach crit mass or said crit mass should be weaker

But that’s in comparison to xbows that simply mass sooner, making even CA civs tend towards xbows

I think this. At that stage of the game the multi tasking workload is high enough that people are distracted enough not to consider stone walling properly

In answer to this and a lot of the following. Many people believe the game is balanced around aggression>defense

Intentional changes were made that supports this idea.

It’s also why early eco instead of late game fire power is generally the overriding factor in win rate. Aggressive games make for better viewership

Pikes, skirms, camels are all ultimately less cost effective than their singular roles should be

That’s what the PUP is for. But all of these things take time to balance and time requires money. So probably cheaper to do easier/simper fixes /balance changes and not necessarily deep dive these things, rather spend money on new content and making the game more playable/stable

Lost time , upset player base

as stated on Reddit, I hope they make an aoe2.2, with a new engine(fix pathing,elevation etc), that simply uses buffed isometric gfx, regional vil skins/buffed architecture, tweaks to the older/polarising/more boring civs, and hopefully tweak the meta so that there is more variety/easier to maximise almost all the units.

And to answer the actual question: And I’m assuming you would need that kind of money (from a new title) to be able to tweak the meta enough to achieve this

1 Like
4

This often comes down to not making more work for yourself than you have to.

E.g. which is less costly: going through civ by civ every 3 to 6 months and making small changes, or making more global changes that dramatically reduce sensitivity of balance to civ design or small bugs?

Its the same principle as code reuse.

Plus youll notice most of the questions here deal solely with parameter values, not mechanics. The cost to change them is relatively cheap.

Im not saying answering these questions conditional on wanting to achieve a specific design objective is easy. But realistically we can equally wonder: were these considered prior to DE? If not what does that imply about future expectations for fixing the meta even with a re-release?

5

Answering the last question only,

I dont know if there is a way to properly “balance” the game. For me, I trust experiment only, thats why I prefer some constant PUPs and stirring the meta changes (but not giant steps, unless with reasons).

I dont believe any mathematical systems can describe the balance well - such that physics can yet be fully explained by maths and can only be discovered by experiment. Thats why I support the way LoL does balance - keep on small nerfs and sometimes big reworks.

1 Like
6

The link doesnt have to be so precisely described as mathematics might describe it.

It could be anything which constrains one answer based on the answer to other questions. For example something as simple as “changes to food economy affect units in proprortion to their food %. This in turn has knock on effects.” We might not know the functional relation between these two things, it could be linear, non-linear, etc.

While this doesnt seem helpful it dramatically restricts the family of possible optimal solutions. You couldnt e.g. get away with having farms cost 200 wood and we know this based on the combination of current data + the relationship expressed above.

7

This post is a treasure in the thousands of miles of piles of garbage on this forum. All the costs in AoE2 reflect the indev meta, which had nothing in common with the meta at release, which has nothing in common with the meta now.

And the post is very on point. Unfortunately, now it’s too late to change anything, as the game has been established. Maybe in AoE5 things will be different.

Spoiler: net -60w +175f for the same time a net +300 w/g can be made is not a optimally designed economy, especially one in which 300 gold can net you 57g 200f.

2 Likes