temple

I shouldn't have ended that with a question mark. simulation/data/technologies/trade_nerf.json { "genericName": "Trade Nerf", "autoResearch": true, "description": "tn", "requirements": {"any": [{"civ": "athen"},{"civ": "mace"},{"civ": "spart"}]}, "icon": "ionic_column.png", "tooltip": "Trade income is reduced by 50%", "modifications": [{"value": "Trader/GainMultiplier", "multiply": 0.5}], "affects": ["Trader"] }

The trade nerf only applies to Athens, Macedonia, and Sparta? (I have my own ideas about reworking trade, haven't posted them yet.) Haven't tested, but I think the vision change will be nice.

This is now in the game (rP20071). Stan did a nice job.

This is an annoying bug (#4725). He does benefit from speed improvements (at the corral), but for some reason hero cavalry are slow (16.5) compared to other cavalry (17.5, 20, 22, and higher for champions and veterans).

That's a good point, slingers have half the crush damage and bolt shooters have twice the health, so that's a factor of four altogether. A picture's worth a thousand words, (2) is a22 and (3) is svn. (Actually it's been changed so that there's no splash damage behind where the bolt lands, so ignore the northern-most damaged unit in (3).)

You have some nice questions about the role of the bolt shooter. But I want to say that the a22 bug is enormous. You should be basing your thoughts on how bolt shooters were in a21, because unless your enemy's really packed together, that's what it's like in svn. Testing now, even against box formation they only do about 50% more damage than in a21. Well, I guess they have extra health and better accuracy, too. But my point is that a22 is very misleading.

They haven't been updated, see the second graph. I'll make a patch for this. Edit: D834.

Veteran units get accuracy upgrades (see the thread I linked to), but it's very tiny now. We could make that a larger bonus. (Also a previous comment in this thread: we could add it to the phases, like how citizen-soldiers get +20% hp in town, +10% hp in city.)

The fact that you've only been playing for a few weeks and already are beating pros who've played for months or years, should be enough indication that something's seriously wrong. I played a game last night (rated, oops), dog + 12 jav cav vs my 4 infantry + 7 spear cav, I lost everything, he lost one horse. GG after 3:28. I'm saying change it. That's the difference between a21 and a22, so that should be the first step in rebalancing. Maybe use an accuracy halfway between the two.

I can't parse this. My point was that historically it might be that sword cav > spear cav, so the bonus should go the other way. I found the comment, it was in your spear cav thread. And things could get interesting with trample damage and charge attack, or even secondary attacks (cavalry could carry spears + swords, or javelin + swords, etc.). I'm not a big student of history, so I can't really say what a realistic countering system should be.

Thanks Hidan. I've created a ticket #4733.

I think the general problem is that it seems impossible to have the same unit weak vs women in village phase but also decent against champs in city phase. Currently citizen-soldiers get +20% hp in town and +10% hp in city. I think we should we should make that more like the veteran upgrade, i.e. have phases give increases in health, armor, attack, accuracy, range, vision, and speed. If citizen-soldiers are weaker, then women and towers are stronger. (We could even make the cavalry bonuses bigger than the infantry ones, if that was necessary to further weaken early raiding.) Another idea is ballistics, to make it a technology to have ranged units aim at moving targets just as accurately as stationary ones. Since fleeing women are moving while hunted animals are largely stationary, this would allow skirm cavalry to be effective hunters while at the same time being ineffective raiders (unless the opponent's not paying attention). It would also making fighting more micro-intensive. This would need testing and maybe some tweaks to projectile speeds. I agree with the need to distinguish sword and spear cavalry, for gameplay reasons. Although I think others have said that actually sword cav should be the counter unit?

Pizza? I assumed there was a chance we weren't the first. I didn't play Empire before last week so I was expecting 16 slices rather than the cc's being placed together like this. (Maybe that could be a variation on this map? I think there's other maps that have some probability of the cc's being placed together, if I'm remembering that right.)

Players: Aska (1310), temple (1733), Naw (1168), Cubidevin (1247) vs Emperior (1445), reggiemarr, phoenixdesk (1229), camelius (1452). Map: Empire, Tiny, 200 pop, 21:31. Summary: Not a great game, but we made history last night. 2017-08-13_0003 - empire pizza.rar

I think in general I would prefer to have gameplay that naturally results in the counters we want, rather than having explicit bonuses. But of course some might be necessary. We already have hack/pierce/crush damage and armor which takes care of some balance. E.g., siege is good against buildings and other units aren't, ranged units are terrible against siege, etc. There are planned things that can take care of other issues. Secondary attacks (#252, D368) would for example give a minimum range to ranged units, which I think is a better solution for melee > archers rather than the current way of giving them less dramatically less health and armor (it's a 3x difference). The idea is if melee units get close enough to ranged units then they'll be forced to use their inferior secondary melee attack. Trample damage (#995) aura for cavalry, so they are better against infantry. Also charge attack (#994), but maybe the effects are similar enough so that trample damage would be enough. Or maybe not. (Charge attack could be a way to differentiate spear cav and sword cav, which are currently the same unit just with different stats.) But also make cavalry obstructions bigger (#4516 is chariots) so that massed infantry can surround them. Imagine how good elephants would be if you could fit them as closely together as infantry! This goes in the opposite direction, and not as extreme. But along with attack/armor/health adjustments, maybe we wouldn't even need the spear bonus against cavalry. Note that already melee units have different ranges (sword 2m, spear 4m, pike 8m), so spears and pikes should be able to attack two or three deep. But in practice that doesn't really happen because units stop at their max range rather than walking forward to make room for units behind them. Some changes to ranged units: elevation bonus (#4028, which I've just submitted D781 for) and ballistics. Currently units aim at moving targets just as accurately as stationary ones, which really shouldn't be the case. If we change that then ranged units would naturally be less effective versus fast-moving cavalry. (Cavalry have large footprints, so right now archers hit them more easily than infantry.) This would help against early raids, since for example jav cav wouldn't hit fleeing women as often. And it would probably help with hunting, too.

I would say the problem was more that they were good vs buildings and good vs units, so you could survive with an army of pure slingers. One way to fix that is to make them weaker vs buildings like in a22. But another way to fix that is to make them weaker vs units, so that a pure slinger army could be more easily countered. I think this second option might lead to more interesting gameplay. (Edit: Maybe slingers could do circular splash damage? Maybe javelin could do linear splash damage? Just offering some ideas.) (Just to do the math here, looking at the stats in a21: A fortress has 40 pierce and 6 crush armor. A ram does 150 crush damage with repeat time 1.5s, which is 53.1 hp/s against a fortress. A slinger does 9.5 pierce and 2 crush damage with repeat time 1s, which is 1.2 hp/s. So one ram is worth 44 slingers. It seems the only change in a22 was slingers' crush damage being cut in half, which brings their rate down to 0.67 hp/s. So now one ram is worth 79 slingers.) Archers shoot faster than skirms and slingers, so they can be better against moving targets. Here are the projectile speeds (m/s): a21 a22 unit 120 75 archer 60 62.5 slinger 56 62.5 skirm 75 75 cav archer 50 62.5 jav cav Archers shoot slower than in a21, along with being less accurate (and promotions having less of an effect). Numerically it seems archers are slightly worse this alpha, unless I've made a mistake. I don't use them often so I can't say my own feeling. (There were some changes with the prepare time of melee units, maybe that could make ranged units seem better.)

The damage in the graphs is the equivalent pierce damage. I took crush damage (of slingers, cats, quins) to be 1/3 pierce damage which corresponds to about 10 extra levels of armor, since infantry and cavalry have about 10 more crush armor compared to pierce armor. Yes, it doesn't take into account moving targets. The thing about ranged units is that they can all attack at once, so they work even better when massed. And since archers have longer range, they can be massed more effectively. I'm just showing the differences. I think these graphs help in understanding the effect of "spread". I like that skirms were made more accurate in a22, but obviously it was too much. And for example, instead of reducing slingers' crush damage from 2 to 1, maybe a better solution would've been to reduce their accuracy. Hopefully we can make a23 better balanced.

I thought I'd post about the accuracy changes in alpha 22, i.e. why skirmishers are so good. Here's a graph of the differences in average damage per second between a21 and a22 for the basic ranged units. Their accuracy depends on the target's footprint (larger things are easier to hit), so I made graphs for both infantry and cavalry targets. (The calculations are straightforward but I won't detail them here.) Javelin cavalry versus infantry at 20-28m now do about twice the damage they did in a21: 12-14 instead of 5-8 hp/s. Their base damage was reduced slightly (the dotted lines are above the solid lines from 0-10m), but that didn't make up for the big increase in accuracy. For a23, we'll definitely want to adjust the numbers. Another difference in a22 is that veteran ranged units aren't boosted as much as in a21. Here's a graph showing that, along with the Mauryan and Persian "Archery Tradition" technology. (Archers' accuracy was reduced slightly in a22, so the dotted lines are above the solid ones.) Promotion doesn't come with a decent increase in accuracy like it had in a21, and the archery tradition difference is really stark: against infantry at 80m, archers only do half the damage they used to. We'll want to adjust these numbers for a23 too. Here's a few more graphs. When fighting against other ships, ships are much more accurate than shown above, since their footprints are so large. The bolt shooter (ballista), catapult (onager), and quinquereme have splash damage, which I haven't taken into account here.

I like all of those ideas, causative. (I never (purposely) use the second alert level.)

And here's some comments and my replies: If you only used the first alert level, then the garrisoned soldiers would stay garrisoned since they wouldn't be put under alert. To me, there's no reason (apart from healing) to have units garrisoned if the enemy's not around; I'd rather have those units working. But I understand your view too. Hopefully in this thread we can get a sense of what the community prefers. That may not be a good thing. I often will set a rally point for a building before it's destroyed, so that the units inside will run away to the desired location when the building is destroyed. If you ungarrison them before the building's captured or destroyed then they'll walk to the rally point. If you're only at the first alert level, then soldiers will move to the rally point while women will run and hide. That might be nice in some situations, but it seems like there are too many complications. (What happens if there's not enough garrison spots for everyone, do they try to garrison somewhere anyway, or just stand still? What happens if some units die on their way to a building, do other units continue on towards a building that's farther away, or do they turn around and try to garrison in this building that now has free spots, potentially running into the enemy and dying themselves? What happens if you manually change some of the garrison orders, because you care more about elite priests than expendable women, do other units continue with their orders or do they alter them now that the "reserved spot" numbers have changed?) So in this case I probably prefer dumb and predictable ("find the closest building that's not full") rather than smart and unpredictable.

Here's a couple improvements (D681) I thought about: 1. #4457 "When the bell is rung, units ejected from a destroyed building should garrison in another" This makes a lot of sense to me. If the units are "running and hiding", then they should do that after the building they're garrisoned inside is captured or destroyed too. 2. Put garrisoned units under alert too. In light of the above, this would mean pre-garrisoned units would also run and hide when their building's captured or destroyed, rather than walk to the building's rally point. A lot of players garrison women just in the area where the attack is happening, and only use the town bell if the attack becomes larger. Currently the women garrisoned first won't be put under alert, so when you lift the alert they'll stay garrisoned rather than returning to work. It's very easy to forget about them in that scenario, so they end up staying inside their house for the rest of the game! (There's also a timesaver here if you want to clear out garrisoned buildings: quickly raise an alert then lift it. Units garrisoned will eject and return to work, while non-garrisoned units will stop for a second before resuming their work.)

I thought it'd be good to have a topic on this issue, maybe to build a consensus. The current system (minus bugs): There's two levels of alert, the first for women and the second for all units. The alert affects all units within the area (except those garrisoned). When put under alert a unit stops what it's doing and tries to garrison in the nearest building, or continues on to the next one if that's full. When the alert is lifted, the unit ungarrisons (if needed) and returns to its pre-alert work orders. The philosophy behind this, as I see it: Units put under alert should drop what they're doing and run and hide. When the alert is lifted they should come out and return to what they were doing previously. So the questions are: Is that a good philosophy for the town bell/alert? And if so, can we make improvements to the current implementation that are consistent with it? Keep in mind that since the town bell is an emergency measure, its behavior should be relatively simple and intuitive -- we don't want to rewrite Petra.

Pathfinding improvements. It's always sad when units get stuck on something or line up in a long row to walk anywhere. Territory/capture rethink. I haven't looked into the history, but the current system seems pretty arbitrary. Market/trading rethink. This also seems pretty arbitrary. For example, trading routes half as long generate income at half the rate, which makes trading useless on small maps and between nearby allies.

One thing that's bugged me is that cavalry is the same size as infantry. Imagine if elephants were that size and never bumped into each other or had trouble pathfinding (relatively speaking). I think if cavalry were fatter (to match their art), then they'd do worse against large infantry armies since more spearmen could attack cavalry at a time than cavalry could attack spearmen.

Gabriel was host and I was an observer, and I saw things fine from his perspective.