• redone from scratch

D3664 isn't committed yet; and even if it were, camels would still be a bit slower than they are right now. But yeah, you're right, they'll probably be more dangerous in the early game; a different cost could work (food, wood, or time).
I would actually prefer putting it in an Arab mercenary camp. To the best of my knowledge, there is no evidence for the Ptolemies fielding camel archers. That's a different discussion, though.

• forgot mercenary costs

It can be done, though I'm unsure it's necessary. Their lower movement speed makes them less useful, as does the fact they don't benefit from cavalry-specific technologies.
Separately, we could consider removing camels from the stable, making it harder for the Ptolemies to mass them.

People should play and test and see how it works out in practice.
Other camel attributes can be differentiated too (e.g. cost), if necessary.

The reason I added it to the forge is because it's a town phase technology and it affects all types of archers. For comparison, the steel working technology (increasing the attack damage of all swordsmen) is also researched at the forge.
The archery tradition technology is at the civic centre only because it's a village-phase technology.

• now with technology (spotted by @borg-)

Oops, you're right, I forgot to svn add it.

See D3735.

Even more health? Each rank already adds 25%, which means basic elephants have 400, advanced 500, elite 625.

Yes, one can select a target, however, that doesn't affect additional arrows, though. Try it yourself (e.g. with a siege tower).

This is a great improvement, thanks a lot!

1. The higher population was added at the explicit request of @wraitii. I don't particularly care either way. Elephant archers are significantly harder to kill than other basic archers on the one hand, on the other hand, they're less dangerous than champion cavalry.
2. That's actually possible by using <BuildingAI>; cf. ships. The downside is that they will then shoot at any enemy target within range, unlike other soldiers, which aim at only one target (the one you order them to or the nearest).
3. Having visible slots (“turret points”; cf. walls) on movable units is my preferred solution too. It would also be useful for chariots and ships. However, right now it isn't possible; it requires both code and new animations.

You're right, they do. The proposed tooltip is correct. The description could be rephrased, its purpose is to teach people something about history or give a justification.

In D3353#161971, @Freagarach wrote:

While resource cost and population cost may be closely related for the end user, they are not code-wise. In the code is it more related to entity limits than cost. Most importantly having an effect after the entity has been created.

That still doesn't explain why it's an improvement to have it under the separate population node.
Yes, I agree population is conceptually related to entity limits. Currently those are located under <TrainingRestrictions> for units and for structures under <BuildingRestrictions>, whereas <RequiredTechnology> is under <Identity>. Actually, I wouldn't mind putting all of them under <Cost>. Also, the entity limits are actually defined elsewhere, under <EntityLimits> in the special/player/player.xml templates.

How do you mean more work for future additions? And for mods I assume you mean they need to change templates again?

No, the need to change templates again is a one-time event, that shouldn't matter. I meant what I wrote: when adding a new template, you now need two cost-related nodes instead of one.

Elephants used to have a vision range of 0, but I see you changed it recently (D2946 / rP24953) to 60.

Why not just give all animals a vision range of 10?

it makes sense to also revert food eco techs.

Look at the gather_farming_*.json changes (below).

Let's keep it simple then: ±10% health and wood cost, no population changes.

In D3668#161685, @borg- wrote:

I said only for consistency, since the project limit per unit is the same as the maximum garrison, If we keep it that way, the tower can garrison 3 soldiers, but one is completely useless for arrows, just for capture.

Actually my point is it's fine to limit either the garrison capacity or the arrow count (cf. army camp, siege tower); doing both is unnecessary.

New mercenary technologies can be added later. Currently it's not even clear how much this patch would alter balance.

The purpose of this patch is removing the training time from the archery tradition technology, because it essentially gives those with the technology an economic bonus over others.
Other technologies can be added, of course, though preferably in separate patches, since such additions could make sense independently of the training time removal.

@azayrahmad, if you type “whaling Romans” or something in a search engine, you'll get dozens of hits from July 2018, all of them covering a single publication, the one linked above. Please read the article in question carefully and critically.
What the authors did was analyse eleven ancient whale bones, only one of which is actually from 0 A.D.'s timeframe (WH812; see the supplement). Based on those they concluded grey whales and right whales used to live in the Gibraltar region. That part is clear.
As for the Roman whaling industry hypothesis, it should be read as a call for further research and is rather speculative; as the article says in section 4c: “None of this demonstrates that a Roman whaling industry existed”.

Indeed, that shouldn't have been deleted.

Try playing Polar Sea.

In D3704#161579, @chrstgtr wrote:

Good points, but metal cannot be easily worked around the way that wood shortages can be worked around. For example, you can make a lot of slingers/mercs for Sahel. You can use your metal to make traders to get wood. On the other hand, once metal is out, it is extremely difficult to get more of.

Not every faction has slingers or mercenaries, though. Moreover, having no wood is extremely annoying, since practically all structures cost wood and most units do as well.

Also, mainland (and similar maps) are by far the most commonly played.

Perhaps, but that doesn't mean gameplay should be balanced entirely based on that kind of map.

It depends on what maps you're playing. Resource shortages are a part of map design. On Mainland metal is the most serious, while on e.g. Northern Lights or Sahel wood is the limiting factor.
Besides, the grain gather technologies already cost 200, 300, 400 wood.

In D3706#161565, @wraitii wrote:

? Long walls have the most HP. Gates are indeed weaker for historical/encouragement reasons, but long walls have more HP exactly to compensate their larger obstruction size.

No, wall towers have more health (4000) than long walls (3000). That's always been the case, for good reason, and should stay as is.

We ought to increase the general cost of walls following this diff, so I don't think this is particularly relevant, but you're certainly correct that it makes walls overall much cheaper.

Well, I wouldn't mind doubling their costs and time, however, even it would be increased a hundredfold, my previous remark remains true: wall towers are the single most important segment that determines the total costs of walls.

Gameplay is exactly the reason why it's important that wall towers have more health than other wall segments. As you pointed out, targetting wall towers can be a waste of time. Targetting gates or long walls is more efficient, therefore that should be encouraged, and having less health does exactly that. Also, the AI doesn't take footprints or obstruction sizes into account.
Furthermore, when building walls, there are more wall towers than other wall pieces taken together. Consequently, those having the highest health and consequently the highest costs works as a check on building walls. Reducing it can only result in making walls being built even more quickly than they are already.

In D3706#161503, @wraitii wrote:

I see no reasons why wall towers should keep 4K HP (equivalent to long wall pieces) when short-wall pieces, which block more ground, are 1000 HP. I would suggest reducing their HP to 1000 or 1500, as I did in D3466/D3465, and their cost accordingly to 12 or

See my answer at https://code.wildfiregames.com/D3465#153373
(Actually it's 4000, 3000, 2500, 2000, 1000 health for the tower, long, gate, medium, and short wall segments.)

I also see no reason they would hold more garrison than equivalently sized wall pieces

It depends what you consider “equivalently sized”; wall towers are thicker and taller. Currently long wall segments can host 8 soldiers, medium 4, short 0. There is room for 12, 8, 4 (or even more), but the length of the short segment is left empty because of overlap with wall towers. 4 seems a reasonable number for wall towers.

(IIRC women can use wall turrets to garrison inside, which I think is _also_ poor design and should be removed).

It is removed in this patch; see line 42.

Yes, I understand how population works. However, what happens after an entity is destroyed is irrelevant here: a cost is something that is needed to obtain something; and population very much fits that definition. Moving it elsewhere just means more work for future additions and mods.
D2948 / rP24394 was very much an improvement: increasing the population limit only happens after an entity is created and is clearly not a cost, it's far more similar to e.g. <ResourceSupply>, unlike the population cost.
Again, could you explain why this patch is an improvement?

Thank you for your response! Your first link is about something in the Arctic, which is irrelevant for 0 A.D. The other two are news articles covering a publication; it's advisable to read further than the headline and look up the actual research article. In this case it's https://doi.org/10.1098/rspb.2018.0961

Does that also mean treasures or resources disappear when depleted?

Could you explain what this patch does?

Right now this patch doesn't touch technologies. However, @borg- is not entirely satisfied with the small tower, hence why I'm making some additional suggestions. (Perhaps Crenellations ought to affect fortresses then too?)
And yeah, the hellenistic_metropolis technology is quite problematic.

Maybe the Crenellations technology (+40% arrows per soldier) ought to be removed? It kind of defeats the purpose of this patch and also makes fully garrisoned towers quite dangerous.

In D3668#161279, @borg- wrote:

For consistency, should the limit of garrison units in the sentry tower decrease to 2? This would affect the capture for sure.

With technologies, a fully garrisoned small tower could fire 6 arrows, a large tower 9, an Iberian tower 14. With this patch the small tower gets a hard maximum of 3, as you requested earlier. The garrison capacity could be lowered instead, but I don't think it's a good idea to do both.

To make the rush a little more viable in the beginning, wouldn't it be interesting to give the civic center arrow default 1? This would make future technologies more interesting as well.

See D2854.

In D3319#161243, @Angen wrote:

can you please make fastest rams as fast as now?

Yeah, that's the idea

also i would prefer seeing ram stats supported by history -> change art and not art -> stats based on art

As for the art, the Persian actors are based on Assyrian reliefs depicting movable ram-towers. The Macedonian actor looks reasonable, while the Roman seems more mediaeval. I don't know about the rest.
We know from detailed Roman texts that rams and siege towers were constructed on the spot, within sight but out of reach of the defenders on the city walls. When finished, they were slowly rolled to the enemy walls; this could take days; when they reached the walls, they were fixed to the ground before the actual ramming started. Rams and siege towers were massive objects that could have crews of a hundred or more. People advocating “two men and a log” don't realize how heavy trees are.
Having large groups of small rams is unrealistic, however, making them massive objects that move at e.g. 5% the current speed would make them unplayable.

Three against.

Three against.

Allow me to explain my reasoning.
Earlier in this discussion (January 16 and 17) you suggested having three levels of rams with greater differences and also different population costs (2, 3, 4). Rams can already raze structures very quickly and are very dangerous in large numbers; reducing their population to 2 would mean even more of them can be fielded, further unbalancing the game; hence why I opted for 3, 4, 5 instead.
Furthermore, rams can be quite hard to destroy, so giving them a lot more health is probably not a good idea; your earlier suggestion (January 17) equated to 320, 400, 480; I lowered it slightly, hence 300, 375, 450; I suppose we could consider 300, 400, 500 too. However, 400, 525, 650 is really far too much. (For comparison, siege towers are a lot more expensive and have only 500 health.)
As for capacity, it ought to be an even number (rams are symmetrical); I prefer 4, 6, 8 but can accept 8, 10, 12.

To me, population is very much a cost. First you need to pay ordinary resources, then you need to have sufficient population, and finally you need to wait a certain amount of time for units to be trained (or structures being built). And if you cancel training or construction, the time is lost but the population is retrieved, just like other resources. Basically population is very much a resource, in between gatherable resources and time. I really don't understand why moving population outside the <Cost> node is an improvement.
For comparison, the <Cost>, <Loot>, <Looter>, <Player/BarterMultiplier>, <ProductionQueue/TechCostMultiplier>, <ResourceGatherer>, <ResourceSupply>, and <ResourceTrickle> nodes all contain resource lines and nobody is suggesting moving all those into a single unified <Resources> node either.

Actually I'm fine with these changes being given a try. Further adjustments can be made later, if and when necessary.

Well, I'm not sure I really want these changes, though I don't really like D3665 or the current situation either. More patches with other ideas are welcome.

In D3699#161153, @borg- wrote:

I would not remove the technology, I would make it exclusive for Carthage. maybe for rank 3?

That can be done.

Yeah, see D3699; it's probably not perfect, yet it doesn't hurt to have more alternatives to choose from.

Why civic structures?

Keep in mind trade income does not scale linearly with the distance.
Right now each trader boading a ship gives 20% of 0.75 = 0.15; with this patch it's 50% of 1.2 = 0.6, so if the distance is somewhat longer, it is profitable.

same speed (proposed by this patch), total cost and base trade gain multiplier

In other words, making merchant ships a lot less cost-effective than they currently are. It could be done, easily, though I think it would not be an improvement.
[EDIT] Here you go: D3697.

Why not do that then, perhaps in combination with a further reduction of their training time?

The current situation is certainly not great and while this patch softens it, it's not really a long-term solution. Mercenaries basically stay “citizens that cost metal”. More important is what mercenaries must be in 0 A.D.; perhaps their templates should get fully separated, as is the case with champions.
(Also keep in mind mercenaries are not supposed to be cost effective.)

Ignore the summary, this patch is heavily outdated. Before I redo it from scratch, it would be nice to figure out reasonable numbers, hence this discussion.
The current ram is roughly equivalent to the proposed medium ram. And no, metal cost stays the same: basically the idea is to have the wood cost proportional to its health, while the metal cost is proportional to the attack damage.

Hence the detailed explanation in the summary.

@Freagarach (or any other team member)?

• rebased

With this patch merchant ships get exactly the same gain to cost ratio as land traders. This means which of the two is more profitable depends on the total distance and how direct the route is. Moreover, merchant ships are much harder to destroy than land traders.
As for traders boarding merchant ships, right now it's rather inefficient, but by raising the factor per trader from 20% to 50%, it becomes more attractive. It doesn't have to be a viable option in every game, but it ought to be usable in some specific situations; right now it isn't; with this patch it is (e.g. if sea routes are clearly longer than land routes).

For comparison, resource gather rates are also numbers typically below 1, not percentages.

It isn't.

If you mean how easily elephant archers can be killed compared to champion elephants, a difference of 9 resistance levels corresponds to a factor 2.581× in health, i.e. 825 health of a champion elephant is worth 2130 health of b/a/e archers (with this patch elephant archers have the same resistances as ranged infantry or cavalry).

In D3679#160832, @ValihrAnt wrote:

I'd prefer all of the attack techs remaining at 15% and just having the last armor upgrade provide +2 armor to avoid the kill times becoming too low after all techs.

Is the idea to have the damage and resistance technologies cancel each other out or to give favour one slightly more when researching them all?

Or an entirely different idea: how about allowing Athenian elite spearmen to promote to champions?

• population to 2, per @wraitii
• lower resistance, health to 400 to compensate
• lower experience

Yeah, I prefer 0 too, however, that makes foundations about 10% weaker; I think it should be done together changing the entity crush resistance of structures from 1 to 0 while compensating their health, i.e. something for a future patch.

@wraitii?

• rebased
• excluded attribute changes (“stats”)

Some people argue they ought to have different attack times, see e.g. https://wildfiregames.com/forum/topic/36639-a24-feedback/?tab=comments#comment-417530
I gave all soldiers the same attack time in my 0abc mod, to make things easier for myself. I haven't reimplemented that in the latest iteration, though. I'm increasingly of the opinion that prepare and repeat times should be based on what looks best for the animations. Artillery has much longer attack times than soldiers, and rightly so. Crossbowmen are somewhere in between artillery and archers. As for archers (this patch), I chose 1.5 s because 1 s simply looks too fast for me and 2 s is too close to the 3 s crossbowmen have.
Displaying damage-per-second values in the user interface would certainly be helpful, though.

Personally I don't really care whether attack technologies give 10% or 15% each.

Which again show we all have different play styles! For me, a longer range means more units can be massed to shoot at a single target.

Yes, the same costs for melee and ranged damage and hack and pierce resistance technologies (the grand total is for everything). Currently attack technologies cost more resources, because they increase the damage by 15%; however, if they're lowered to 10% each, they can have the same costs as their resistance counterparts.

Perhaps you could write a patch for that? I'm not sure I fully understand what you mean.

How about the following?

         small, medium, large
wood      : 240 , 300 , 360
population:   3 ,   4 ,   5
health    : 300 , 375 , 450
capacity  :   6 ,   8 ,  10
movement  : 80% , 75% , 70%

The time reduction looks good and can be committed as is. If and when a future patch tweaks training times of other units, then cavalry values can be further adjusted then and there, but right now there is no point in waiting for that.
As for the movement speed change, I have no objections to it and it could be given a try. However, I think adjusting unit turn rates is more important and might have a greater effect. Perhaps we should wait for such a patch.
And maybe we ought to differentiate between horse archers (×0.8) and javelineers (×0.9).

In D3668#160647, @wraitii wrote:

I would like to keep a late-game tech to increase tower base-attack, as it now counters stronger units, so maybe I'd favour reworking techs in a later patch.

Something for a later patch, yes. If a technology increases tower attack damage, then I think fortresses etc. ought to benefit too. Or perhaps structures should benefit from the archer technologies available at the forge.
Another idea is raising the attack spread of defensive structures from 1.5 to 2 and introducing a technology that improves it by 25% (i.e. returning to the current value). Right now structures are much more accurate than basic troops; a higher spread would make raiding a bit easier.

How about the following costs?

1. 200 f/wood + 100 metal + 40 s
2. 350 f/wood + 250 metal + 50 s
3. 500 f/wood + 400 metal + 60 s

Grand total: 2100 food, 2100 wood, 3000 metal, 600 s,

• add a max arrow count to small towers, per @borg-
• keep sentries technology and keep default arrow count of large towers at 1, per @ValihrAnt

You mean using the already existing skirmish placeholders? I tried, and while it doesn't cause errors, they're displayed as black squares and attribute values aren't displayed either.
Hence this approach with more files.

No, civic centres don't count towards the town phase, and I believe that's intentional and shouldn't change. To start with, it's the structure that researches the phase technologies, so you always have one. Moreover, it's technically a village phase structure; if you don't have one at game start (e.g. nomad), you can build one; and on some maps (e.g. Empire) you start with two.