ChronA

I do believe a quiver might be somewhat encumbering when running at a full sprint, but that would depend on the precise gear arrangement. If the quiver is open the arrows might bounce out. Likewise, if the quiver were loosely strapped to the wearer, it might flap around and disrupt their gait. However a closed or snuggly fitted quiver, which was well secured to the wearer or supported with the dominant hand while running, would sure be no more encumbering than a skirmisher's small shield and javelins. So, yeah, there might be an argument for a very small speed disadvantage for archers vs slingers and skirmishers, but overall I think giving everyone the same speed is probably the more authentic choice. Yes, I think discipline and training is the key consideration. Putting myself in the sandals of an ancient infantry man, I think I could easily buy the idea of "point this long stick at the charging horsey and he won't run over you because he doesn't want to die." I would be much more skeptical of the instruction "don't worry, the horsey won't run over you because if he does run over you then your friends will maybe try to chop him up with their big knives while he is tanged in your corpse (and he doesn't want to die)." The former promises me that if the horseman decides to test the theory then he will definitely die, but I might live. The latter tells me I will definitely die if the horse decides to test it, but he will only possibly die. Both actually work equally well provided your mounted opponent is not a suicidal idiot, but the spear works out much better for the front row infantry man on those occasions when the enemy does field suicidal idiots. The problem is that in 0 AD every unit is a suicidal idiot. That kind of ruins the whole historical concept of the cavalry charge, which was always just a weaponized game of chicken. I'm not sure there is a clean way to fudge the dynamic in the simulation without resorting to a whole aura-based morale system.

I never said that ranged infantry would be able to beat cavalry. Neither did I. However, I see I was in error because you specified that foot javelinists should be "situationally useful against melee infantry"... meaning that civs with javelin skirmishers could build them to counter foot swordsmen, spearmen, & pikemen (& elephants). Then in turn cavalry could be produced to counter them. That still leaves no clear reason to produce foot archers or slinger, but it is better than I thought. Yes, I've looked at them. Those shields cover about 60% of the rider and 10% of the horse... which leaves the horse looking mighty squishy. Agreed. And yet: the Athenian Hippeus b...

I broadly agree with Thorfinn's preferences. I have a few comments though. 1. If melee infantry categorically beat both ranged infantry and cavalry, there really is no gameplay reason to ever build anything except melee infantry. For the sake of gameplay diversity, ranged infantry and cavalry really need to provide some sort of value-added beyond the capabilities of the standard swordsmen, spearmen, & pikemen. In this writeup I don't see that they do, except possibly acting as a superior raiding force (which is easily countered by static defense, and so will only see a little bit of early-game play). 2. Giving cav and foot archers minimum range could be a double edged sword. It can make them automatically kite any unit substantially slower than them, which is a pretty huge advantage. My experience playing around with unit modding is that a unit with a minimum range does not always perform worse than the same unit without minimum range. Often it actually becomes MUCH better, sometimes even reversing an intended counter. (But on the other hand, I was just complaining that ranged infantry don't add any value in this schema. Maybe they could all be configured to auto-kite the melee infantry, which would in turn give the cavalry a unique role to play by chasing the pests down.) 3. While I do have any principled objection to providing units intended to counter ranged attack with high pierce resistances, that decision really ought to be reflected in their artwork. An unarmored, tunic-wearing farmer on a horse should not be made an arrow-sponge just because a balance plan dictates that sword cavalry counter ranged.

I second this.

I would not consider the problem of the missing rush solved until rushing is viable for almost all civs, and not just those few with access to mercs. Unless the plan is to give every civ P1 access to mercs or champions as a tool to enable rushing, any effort you put in now to balance this will only be a stop-gap. It will all have to be redone from scratch eventually, as part of the general overhaul of unit roles and relationships that it will take to bring this game into a semblance of polish. That's not to say the effort would be wasted, but this is a bandage on a cut that needs stiches.

We are getting close to the alpha 25 freeze date now. Much ink has been spilled on a variety of worthy topics, hopefully not too much of it in vain. Happily, I've got the privilege of an unusually light work week on my plate starting tomorrow. If there are any last minute tasks that need doing, I can help out. I'm guessing I should join onto the designated IRC to coordinate efforts and get help with using the version control system. Or is there a better way that you would prefer?

Sounds like a good outcome. I shall watch your replay with interest.

Really? The whole reason I made this post was that I was getting tired of always changing units' armor values from 1 to 0 in my personal balance experiment mod, and I have never had a template return invalid as a result. Maybe you made a mistake in the syntax? Admittedly this is a question of balance philosophy, but I think if an 11% change in damage received or dealt substantively alters the balance of the game it is a sign of a design problem. If your balance needs to be that finely tuned to work correctly, then almost every change in unit stats or meta will necessitate massive balance revisions to maintain a playable experience. In contrast, a balance painted in bolder strokes will maintain its general shape even as the relative popularity of certain tactics rises and falls in proportion to their current strength (a la game theory).** When I'm doing my experiments, I actually try to start by working only in powers of 2: doubling or halving damage, attack multipliers, and HP until the units have the qualitative relationships I want. Then I try to make minimal fine adjustments necessary to restore game feel. I'll grant that this technique can make for some very hard counters, but it makes maintaining balance easier, and every unit gets a very distinctive role and identity. ** P.S. This approach can definitely cause problem with civ balance when factions are differentiated primarily through their preferred playstyles within a common pool of archetypical strategies. If a certain tactic becomes too niche, civs that specialize in it will get much less play. I would argue that this just means each civ needs to have a couple of viable playstyles in its pocket, but reasonable minds may differ.

Ah! Now that's a good theory! Maybe there was thought of having a hero with an armor-debuffing aura. If every unit has at least 1 armor they could have it reduced by 1 and not fall into negative resistances. (I don't think negative resistance would intrinsically cause any issues, math-wise, but I can see how someone would worry about it.) Likewise maybe someone was worried about divide-by-zero problems if there was an armor multiplying/dividing aura. These are plausible explanations. Unless such auras are already in the game and are see a lot of play, I'm not sure I buy that they are worth forgoing the benefits I listed... but it would at least explain things. Also, thanks for your encouraging words, @hyperion. When messing with gameplay and balance I definitely do believe in a conservative approach. I think there was an XKCD to the effect that any time you find some human made thing that looks weird and stupid, it's a sign that a brilliant engineer spent days trying to optimize an impossible set of demands, before throwing up their hands in frustration. It's a bit hubristic to suppose one can simply do better without large portions of effort and luck.

I am a subscriber to the logic of Chesterton's fence. As much as I don't like this design choice, I must assume it served some purpose at some point in time. Thus it is imprudent for me to invest resources into developing a new solution until I can account for that original purpose. Possibly. However there are some concrete reasons to make a change. It would make it easier to understand the number of hits it takes to kill a unit, which is arguably the most important statistic for understanding unit balance. It would make the GUI slightly simpler and more informative for new players. It would make it easier to test and observe observe in-game how modification that affect damage are working. And it would shorten and simplify some of the template xml files. Is there an easy way to do this? It seems like it is probably not the result of a single patch, and who knows when it was introduced. Querying the development log for "armor" turns up a mountain of irrelevant tickets.

This is something that really confuses and annoys me. Why do women need a 10 percent damage reduction against every damage type? Why do animals and ranged units need a 10 percent damage reduction against every damage type? Would it have been so game breaking to just let them take the amount of damage it says the attack is supposed to do?

Hm. If arrows are doing 6 damage that means the new templates aren't loading. If it were just the scripts that were broken, the archer should be doing 20 damage per shot. I'm guessing the mod isn't loading at all. For additional confirmation, could you check the hack armor on the pikeman: it should be 18. Just a stupid hunch here, but try reinstalling the my mod by unzipping it in your mods folder. I gather properly built mods are zipped, but I didn't go through that whole process. (I am not a proper computer sciency type--sorry.) Also, is it working for anyone else?

@alre That's odd. I've checked it again on my side and it is definitely working for me. The gazelle takes damage from every shot, hits and misses. So to try to figure this out, a few questions: Are you sure the gazelle isn't taking damage? If it was very far away the arrow might have a very long flight time before the hit registers. Also, if the gazelle is very far away the damage falloff will reduce the damage to low single digits. (At max range each arrow should only do 5 damage, if he is shooting beyond his max range--because that is something units can do when leading a target--it will be even less.) Were you using a citizen soldier archer, or a champion/hero archer? I didn't get a chance to propagate the changes to champions or heroes, so they can still miss. Do you have any other mods enabled in your load order? If so, try disabling them. Are you using the alpha 24 build, or have you upgraded to the a25 preview? I prototyped the mod from a24. Its possible a25 breaks something.

If you want to test this you will probably want to open up a text editor and make some adjustments. I put in the bones of an interesting balance concept with the infantry templates I included. However their main purpose was to demo some of the things one can do with the new features. Thus the ranged infantry in particular are a hodge-podge of different concepts. (Archers have fully deterministic accuracy with the maximum amount of damage falloff allowed. Slingers are fully spread-based with damage falloff starting at 50% of their max range, and skirmishers are 50% spread and 50% deterministic with no falloff.) For my own tastes I would probably drop the archer and skirmisher <AccuracyOverride> down to 0.3 and 0.4 respectively, then bump up the archer's <OptimalRange> to 20. To complete things for multiplayer testing, the changes would also need to be extended to cavalry, champions, and static defense buildings. I didn't have time to work on them yet. If anyone is especially interested in my take, I can probably dedicate a few hours tomorrow to whip a polished version up.

Okay! At some point I've got to stop playing with different stat values on my own and share this darn thing! This mod contains support for a deterministic damage model, via accuracy overrides and distance based exponential damage falloff. It also adds support for directional armor/resistance values. The included unit templates demonstrate the grammar for these features. (I think they also strike an interesting new balance between unit roles... but I'm biased obviously.) Let me know your thoughts. damage_mechanics_mod.zip

@Grapjas Ah, I see what you are saying. Yes, it is possible that my premise is flawed. I don't actually play competitively, so I am drawing from a biased sample of what vocal multiplayers are saying on the forums, and my own inferences from studying unit stats and casted matches. Perhaps I am looking at the wrong games, and in more even matchups people are rallying more. Still, I think it is still valid to acknowledge the merit of a patch that would encourage more volatile games.

Very true. And a good community can often salvage a flawed game by adopting rules of sportsmanship that counteract balance problems. But a craftsperson should strive to produce an ideal product regardless.

I did a bit of experimenting with modifying resource drop distances in A23. I didn't bother with auras or anything: just loosened the leashing on the am-I-close-enough-to-my dropoff-point-yet check. Unless that part of the code has received heavy modifications since then, it should not be hard to implement variable drop distances, and the performance hit should be very-small to non-existent.

Lower damage increases the opportunities for micro; increased damage gives more value to initial positioning. Low damage games have more tendency to snowball, as players will be more able to escape from bad fights before suffering significant damage. Personally I like micro opportunities, but I must say, EA already suffers from a problematic amount of snowballing. RTS is at its most interesting when there is a sustained back and forth (see AOE2 and SC BW). This requires the ability to either trap or immediately kill large numbers of units when they get out of position. EA's pathfinding doesn't really allow much trapping, so buffing melee DPS is probably the better option over all. Edit: But on the other hand, it might also be that slow, grindy fighting has become a staple of 0AD's identity. Changing it to prevent snowballing could alienate current players. In that case nerfing ranged DPS combined with other measures to reduce snowballing might be the better solution... in short, IDK.

If economic parity is adopted as a permanent fixture of the balance design, then yes, adding this feature to the code base would be a good investment of effort. However, my expectation is that this will only be a temporary condition. I'm hopeful (perhaps delusional is the better word) that in another 12 months or so--around release 28 or 29--the core archetypes and balance philosophy will have been refined to the point that they are broadly accepted as finalized. At that point balance discussion can productively move on to civ balancing and economic diversification. And then the skirmishers and the slingers can finally get their speed buffs (with compensatory eco and combat nerfs as necessary) and all will be well with the world again. Of course, besides my insane optimism about the 0AD's future development rate, this plan assumes that finding our desired role definitions for skirms and slingers is not predicated on them having some mobility advantages over archers. I think this is the heart of your concern... and you may very well be right. However I am not yet convinced enough to invest into a major coding project over it.

I support the uniformity of citizen soldier movement stats as a balance testing expedience, for the time being. It is much harder to evaluate which unit types are over or under performing in real games if you also have to account for variations in economic strength as a confounding variable. Would we have realized just how powerful the archer unit archetype is if slingers and skirmishers had retained their mobility driven economic bonuses? Once the project is in mid-to-late beta I will definitely be disappointed if variations in unit mobility are not reintroduced (along with plenty of other civ specific economic bonuses), but for now I think standardization is necessary to keep the project moving forward. I second this sentiment. Given the time period, the gameplay focus should absolutely be on melee infantry. The first thing we should try is (3) adding some inaccuracy to archers. It's what's already been patched in for A25 unless I'm much mistaken. Let's see how it works. If it falls short I think the next best suggestion is (4) damage drop off with range. As of today I've now got the code for this all working, although I'd like to brood on it one more day just to be sure there are no well hidden bugs. I don't like the idea of adding pierce armor to skims or slingers because it doesn't fit their game art and it's dubiously historical. I actually really like this concept. Unfortunately I can also imagine it being a huge PITA for code maintenance, so probably not something likely to win approval.

Indeed. assuming we can scale them all without remaking them (some cannot be imported in blender), which we cannot. And that sunk cost fallacy is precisely why I argue so loudly against adding more civs (and art assets in general) while the game remains in alpha. It is a bad situation when art assets start dictating what is possible in core gameplay and balance decisions. But it is what it is, and we are stuck with it. This very thread is a prime example. A lot of people want pikemen to fit into the role of the tank, but it simply does not fit their art direction. Moreover, the fact that each civ gets exactly one P1 melee infantry unit, and for some civs that one is a pikeman, prevents us from exploring some of the more interesting possibilities for diversifying their role: Making pikemen slow super-tanks doesn't work because (apart from not fitting their in game depiction) lowering their movement speed would nerf the economy of pike civs... Making them melee superiority with high hack resistance but low pierce armor won't work because it would make an anti-cav counter, which the pike civs depend on, too vulnerable to ranged infantry... etc. I don't want to point fingers. It seems to me that everyone working on this project does so out of genuine enthusiasm and good faith. But we should acknowledge that the art contributors, who appear to enjoy an privileged role in the governance of this project, are likewise culpable in an outsized fraction of its balance and gameplay issues. Maybe it is not unreasonable to ask those same contributors for some outsized efforts in service of solving those problems?

Like I've said elsewhere, halving the attack speed of ranged fighting assets (buildings included) would probably improve a lot of balance problems.

I can't actually argue against switching elephant attack damage from hack to pierce. I think that's an indictment of how arbitrary EA's damage types are, more than anything else. More importantly though, I argue that this issue is merely symptomatic of the deeper problem that we've been discussing elsewhere: you guessed it... ranged vs melee balance and static defense vs unit balance. Elephants would have a harder time killing rams if those rams were actually escorted by meat shield to their target. However, no one escorts their rams because there is no unit type with sufficient arrow resistance to dive against static defense and ranged fire support.

Drat, I was afraid that might be the case. Based on my experiments, I think "completely broken" is hyperbolic, but I was definitely seeing more problems with units getting stuck on things. I was hoping there was some other parameter that just needed to be tweaked. I guess it is more complicated than that. Sounds like you think there would be performance costs too? I must say that is disappointing. It's these kinds of unpolished little gameplay details that undermine the appeal of 0 AD to me. Perhaps another solution would be to scale down the models, but with the amount of art assets I'm guessing that would take months of continuous work.