Feedback: Units

[Xyth]

Units
Feedback and Development

 

[Howard Mahler]

Grenadier is currently in one later tech column than rifleman (and cavalry).
I think it works better and makes more sense historically if it were in the same tech column or one column earlier than rifleman.
No obvious techs for Grenadier, but maybe Chemistry is best in the current tech tree.

Not a big problem, but I think it should be solved somehow.

 

[Howard Mahler]

Not sure Corporation makes a lot of sense for Ship of the Line.
I do not have anything specific to suggest.

Just something to look at if you get the chance, particularly if you fiddle with this part of the tech tree.

 

[Xyth]

Grenadier is currently in one later tech column than rifleman (and cavalry).
I think it works better and makes more sense historically if it were in the same tech column or one column earlier than rifleman.
No obvious techs for Grenadier, but maybe Chemistry is best in the current tech tree.

Not a big problem, but I think it should be solved somehow.

Not sure Corporation makes a lot of sense for Ship of the Line.
I do not have anything specific to suggest.

Just something to look at if you get the chance, particularly if you fiddle with this part of the tech tree.

This part of the tree needs some significant reworking. There are several problems like this there. Hope to address it for 0.9.5.

 

[Xyth]

Here is a small patch for those interested in testing a new feature. It gives Workers the ability to hurry production in a city like a Great Engineer does, though with fewer hammers (20, 30, 45 or 60, depending on gamespeed). As discussed in another thread this is designed to give workers something to do once they've built everything, it could also help get a new colony up and running much faster.

It turned out to be very easy to implement (though a few tooltips haven't been changed yet) so I thought I'd post it here for people to try. I'm pretty busy messing with terrain and techtree changes at the moment so it would be a big help to get some feedback on this potential change from players.

It contains just one file which replaces the file of the same name: /History Rewritten/Assets/XML/Units/CIV4UnitInfos.xml. Backup the original file in case you want to revert and be aware that you'll most probably be unable to continue any existing saved games once the new file is installed.

 

[Howard Mahler]

It is very good to test things in this way.

While this does indeed give workers something to do later in the game, I suspect it will need some restrictions.

In I believe Civ I, there were caravans, that one could be used to rush production.
When the tech for a new wonder was discovered, one could have all of these caravans waiting to build the wonder in a city in one turn. Unlike Great Engineers which are hard to get and somewhat random, one could specifically build caravans. (You knew you were researching the wonder tech.)
The caravan mechanism was dropped for very good reasons. It was just overpowered and broke the basic framework of the game.

I think the new proposed worker ability would benefit from one or more restrictions such as for example the following:
1. At most one worker can be used per turn by a civ in this way. (Maybe 2 or 3 might work.)
2. Can only be used to rush buildings, not units or wonders.
3. Only becomes available later in the game; requires the proper tech such as Labor Unions.

I await with interest results of tests.

Here is a small patch for those interested in testing a new feature. It gives Workers the ability to hurry production in a city like a Great Engineer does, though with fewer hammers (20, 30, 45 or 60, depending on gamespeed). As discussed in another thread this is designed to give workers something to do once they've built everything, it could also help get a new colony up and running much faster.

It turned out to be very easy to implement (though a few tooltips haven't been changed yet) so I thought I'd post it here for people to try. I'm pretty busy messing with terrain and techtree changes at the moment so it would be a big help to get some feedback on this potential change from players.

It contains just one file which replaces the file of the same name: /History Rewritten/Assets/XML/Units/CIV4UnitInfos.xml. Backup the original file in case you want to revert and be aware that you'll most probably be unable to continue any existing saved games once the new file is installed.

 

[Highwayhoss]

It is very good to test things in this way.

While this does indeed give workers something to do later in the game, I suspect it will need some restrictions.

In I believe Civ I, there were caravans, that one could be used to rush production.
When the tech for a new wonder was discovered, one could have all of these caravans waiting to build the wonder in a city in one turn. Unlike Great Engineers which are hard to get and somewhat random, one could specifically build caravans. (You knew you were researching the wonder tech.)
The caravan mechanism was dropped for very good reasons. It was just overpowered and broke the basic framework of the game.

I think the new proposed worker ability would benefit from one or more restrictions such as for example the following:
1. At most one worker can be used per turn by a civ in this way. (Maybe 2 or 3 might work.)
2. Can only be used to rush buildings, not units or wonders.
3. Only becomes available later in the game; requires the proper tech such as Labor Unions.

I await with interest results of tests.

I do see your point in that a flood of workers rushing wonders can unbalance the game, so I've been looking over the Civ4UnitInfo.xml modiki for ideas. After looking it over I got a couple of ideas on how to make workers more useful without unbalancing the game while implementing some of your ideas:

1. Instead of giving workers the abilty to rush , use the <Buildings> tag in Civ4UnitInfos to specify certain buildings the worker might build: Granaries, Klins, and other essential infrastructure buildings. This would help cities build faster and soak up excess workers while preventing rushing of units and wonders. By specifying exactly what workers can build you can control exactly how powerful they are.
2. Also give workers the ability to settle in cities as a SPECIALIST_CITIZEN. Another way to soak up excess workers.
3. Create an engineer unit using the Art of the modern worker that workers can upgrade to with getting say the Engineering or some other tech. This unit would be able to build more advanced buildings and rush production. A civilization could be limited to say three engineers (like missionaries) to prevent stockpiling.

So what do you think?

 

[Azoth]

Hmm. I never had the opportunity to play Civilization I. Do you remember, Howard, how much those Caravan units cost, as compared to how much production they could rush? I assumed that 60-hammer Workers that only gave 30 hammers back would not be worth stockpiling. I would never have thought to rush wonders with them; surely it would be more efficient to build wonders directly? Especially if you had access to Marble or Stone, which neither Workers nor Great Engineers take into account?

Anyway, Highwayhoss has some excellent ideas based on your proposed restrictions:

1. Instead of giving workers the abilty to rush , use the <Buildings> tag in Civ4UnitInfos to specify certain buildings the worker might build: Granaries, Klins, and other essential infrastructure buildings. This would help cities build faster and soak up excess workers while preventing rushing of units and wonders. By specifying exactly what workers can build you can control exactly how powerful they are.

If I understand correctly, you're suggesting that workers might instantly build certain basic buildings. It's a good idea but I worry it's too strong. Granaries cost 60 Normal speed hammers: that's twice the proposed 30 hammer rush ability. Instant-Granaries and -Kilns from preexisting Workers, with no hammer loss at all, would seriously change the early game. (I don't even want to think about instant-Terraces.) This might be a better ability for Settlers: at 100 Normal speed hammers, they are more expensive than most basic buildings. And if you lose that last city spot to the AI, there's really nothing to do but disband them.

2. Also give workers the ability to settle in cities as a SPECIALIST_CITIZEN. Another way to soak up excess workers.

A citizen specialist? I like it! This gives Workers a unique ability that doesn't compete with Great Engineers. At the same time, it eliminates the stockpiling problem; a free citizen specialist just isn't worth the trouble. I have only one concern: after your 15 or so Workers finish building Railroads in the Industrial Era, they could all come together and join a single city. That would give the city an immediate 15 hammer/turn boost; 30 hammers/turn if it had a Forge, Factory, and power plant; and even more with an Ironworks or Heroic Epic. I don't think it matters so late in the game, but I thought I should point it out anyway.

3. Create an engineer unit using the Art of the modern worker that workers can upgrade to with getting say the Engineering or some other tech. This unit would be able to build more advanced buildings and rush production. A civilization could be limited to say three engineers (like missionaries) to prevent stockpiling.

As Xyth hinted earlier, this is a fine idea. I would avoid the name "Engineer" simply because we already have Engineer specialists and Great Engineers. What do you think of "Architect," unlocked at Architecture? The Architect would cost 60 hammers, be built with food, and be available in limited numbers. Workers could upgrade to Architects at zero cost, since both units would have an equal hammer value. You would also be able to build Architects directly. A building requirement, like Monastery for Missionaries, might be a good idea. How about the Kiln, which doesn't have a distinct purpose at present?


To conclude, I think some combination of (2) and (3) would be best.

 

[Howard Mahler]

I do not remember what caravans cost in hammers versus their yield.
Key features I do remember:
1. They could be built in unlimited quantities just like any other unit.
2. They could move.
3. They could be used in unlimited quantities to rush wonders (and probably other things as well.)

The key point is you could build lots of caravans, move them to the appropriate city, and on the very turn you discovered a tech that unlocked a Wonder you could build the Wonder. (World Wonders, I believe that Civ I had no National Wonders.)

On the one hand, you probably spent more total hammers building the caravans than you would have by building the wonder the regular way.
On the other hand, you could build caravans in several different cities.
Most importantly, you got to build the wonder before somebody else did. Slow and steady would almost never win the Wonder race.

This whole discussion I believe started as an attempt to give workers something additional to do in the latter parts of the game. This whole thing started as a way to introduce a minor tweak or two in the later game, not to come up with a game altering mechanism.

The ability of settling as a "special citizen" I think is the best idea so far.
I like it and subject to testing I think it should be added.
I would suggest testing this without the rushing idea.
(Assuming the AI would understand it and it could be programmed.)
It is unlikely to be game busting and presumably would make more sense once workers are not needed to do other things such as build roads, mines, etc.

For example, a special scientist gives more than an ordinary scientist but no GPP. A citizen gives +1 hammer. While a "special citizen" might give something extra such as +1 coin, I lean towards just the +1 hammer.

The rushing idea would probably work with severe restrictions. Some good ones have been thrown out there.
However, this has the inherent potential to get out of hand.
Also it overlaps to some extent the existing abilities to rush via slavery or with money, not to mention Great Engineers.

 

[Xyth]

Here's a new test patch for the worker, this time with the ability to settle as citizen specialists (with no restrictions). The file goes in the same location as the previous one.

 

[Keinpferd]

I suggest we move the rest of the discussion on ranged combat, Siege units, and the like to the aforementioned Units thread.

Done.

I maintain that ranged combat was a spectacular failure in Civilization III.

Right.

Granted, some mods might be compatible with ranged combat, especially if they rework other aspects of the Civilization combat engine.

Not might, they are compatible. The only thing about ranged combat, the AI in PAE won't understand is how to load units with ranged ability in boats or fire at boats. In practice, your concern about an exaggerated advantage to the defense doesn't prove true. No reason to be skeptical about the idea of airplanes anymore, if they are circling above your head.

Your remarks about PAE seem to be talking about an entirely different mod. Very little resemblance to the PAE of late, as I know it. But with your conclusive remark you hit the spot what PAE is about:

At that point, we may as well be playing a different game.

It is a different game. And no further mention of PAE from my side.

As for carving out a distinct personality, I will always think of History Rewritten as the sensible mod.

Absolutely.

So what if there are a hundred mods out there "more advanced" than HR: it doesn't mean they're better.

I wasn't trying to say, HR were less advanced than other mods. On the contrary, at a very early stage, HR 0.2, let's say, you could already see, that it was considerably more advanced than other mods at a comparable stage of development and easy to predict, it will be competing with only the best mods, one day. But it takes some time, until a mod has a kind of an unmistakeable "trademark" feel to it, a cetain atmosphere that commercial products are usually lacking.

At this stage, a hundred mods are more advanced than yours, like, as a yet different example, this one. I'm assuming here the mod keeps the promises its description raises. This reads fun, idea-rich, innovative, phantastic. I'm convinced, at future stages, HR will get there, too.

The quoted mod certainly doesn't fall under the category kitchen sink, how we call it here. I think there is little chance to confuse, which mods are additive-but-adding-nothing and which aren't.

A designer knows he has achieved perfection not when there is nothing left to add but when there is nothing left to take away.

Quotable!

 

[Xyth]

I decided to take a look at how Siege units and ranged combat is implemented in PAE IV, to see what Pie had done to balance it there. Obviously it's difficult to compare this one aspect in isolation as HR and PAE are very different mods but here's a summary:

&#8226; Like HR, there are Ram units that cannot damage units, only bombard cities
&#8226; Catapults and the like can both bombard cities and attack units
&#8226; Archery/Skirmisher units can also do ranged attacks
&#8226; Archery/Skirmisher units still have first strikes, but also collateral damage and withdrawal chance
&#8226; Ranged combat is restricted to 1 tile only
&#8226; Units that do collateral damage do it to fewer units in a stack compared with BTS
&#8226; Ranged siege can't kill units, only reduce them to 80-90%
&#8226; Archery/Skirmisher units can kill
&#8226; The only promotions available to siege units are Accuracy, Barrage and City Garrison
&#8226; Some of those promotions give a wider scope of benefits though (e.g City Garrison I is +20% City Defense, suffer 10% less collateral damage and +10% vs Melee)
&#8226; One siege unit, the Onager, seems to be specifically designed for city defense
&#8226; A few siege and archery units have a maximum number that can be built at one time
&#8226; Siege units cost extra gold per turn to upkeep
&#8226; There are no changes at all to the AI regarding ranged combat​

This gives us a starting point for focussed discussion. My view is that suicide siege might be balanced but it's somewhat nonsensical and not very interesting. Ranged combat is worth pursuing (it's realistic and it's fun) but only if we can arrive at a model that is balanced and that the AI will be able to cope with.

 

[Simon_Jester]

The real key, I think, is "ranged combat is nonlethal." This means you must commit units to 'brute force' combat in order to decide a battle in your favor, and also means that to easily overpower a stack of X units by firing a ranged bombardment and then finishing it off close up, you need to deploy 2X units.

And using 2X units to attack X units always works.

 

[Keinpferd]

• There are no changes at all to the AI regarding ranged combat[/INDENT]

This gives us a starting point for focussed discussion. My view is that suicide siege might be balanced but it's somewhat nonsensical and not very interesting. Ranged combat is worth pursuing (it's realistic and it's fun) but only if we can arrive at a model that is balanced and that the AI will be able to cope with.

May I correct this? The reason why ranged combat works in PAE and the AI knows perfectly well how to use it is a small change to the AI's UnitAIs. It's that sequence (related to archers in this example):

Spoiler :

<UnitAIs>
<UnitAI>
<UnitAIType>UNITAI_COLLATERAL</UnitAIType>
<bUnitAI>1</bUnitAI>
</UnitAI>
<UnitAI>
<UnitAIType>UNITAI_CITY_DEFENSE</UnitAIType>
<bUnitAI>1</bUnitAI>
</UnitAI>
<UnitAI>
<UnitAIType>UNITAI_CITY_COUNTER</UnitAIType>
<bUnitAI>1</bUnitAI>
</UnitAI>
<UnitAI>
<UnitAIType>UNITAI_ATTACK</UnitAIType>
<bUnitAI>1</bUnitAI>
</UnitAI>
<UnitAI>
<UnitAIType>UNITAI_ATTACK_CITY</UnitAIType>
<bUnitAI>1</bUnitAI>
</UnitAI>
</UnitAIs>

The sequence is the crucial point. If changed, the archer would be less keen to apply ranged mode or wouldn't at all. That was the outcome of patient experimenting, I guess. Even if it's not a complex SDK flic flac but just an ordinary XML-change, it does the job.

And suicidal catapults, balanced or not, are always bloody stupid.

 

[Azoth]

The real key, I think, is "ranged combat is nonlethal." This means you must commit units to 'brute force' combat in order to decide a battle in your favor, and also means that to easily overpower a stack of X units by firing a ranged bombardment and then finishing it off close up, you need to deploy 2X units.

And using 2X units to attack X units always works.

No, I disagree. The real issue is "ranged combat carries no risk." Using 2X units to attack X units always works, but you generally lose 0.5X units in the process, which you have to replace. With ranged combat, so long as you can arrange a first strike, you can bombard away then clean up at 95%+ odds. That way, your 2X units eliminate X units with virtually zero losses. You can proceed to bombard another X units next turn; and, soon enough, your 2X units have chewed through 10X worth of AI units. That's why ranged combat is fundamentally unworkable.

I'll have more comments and hopefully some concrete suggestions tomorrow.

P.S. I agree that suicide-Catapults are silly. I'm certainly not suggesting we return to that model.

 

[Xyth]

The real key, I think, is "ranged combat is nonlethal."

Yes, this seems a very good place to start.

May I correct this? The reason why ranged combat works in PAE and the AI knows perfectly well how to use it is a small change to the AI's UnitAIs. It's that sequence (related to archers in this example):

My bad, I meant "no AI scripting", i.e no extra coding in Python. I'd hadn't looked at the UnitAIs and I should have.[/QUOTE]

No, I disagree. The real issue is "ranged combat carries no risk." Using 2X units to attack X units always works, but you generally lose 0.5X units in the process, which you have to replace. With ranged combat, so long as you can arrange a first strike, you can bombard away then clean up at 95%+ odds. That way, your 2X units eliminate X units with virtually zero losses. You can proceed to bombard another X units next turn; and, soon enough, your 2X units have chewed through 10X worth of AI units. That's why ranged combat is fundamentally unworkable.

I think there's a fair bit we can do to overcome this. PAE seems to address it by making siege units very expensive, giving ranged attacks + first strikes + collateral damage to archery units (and the appropriate UnitAI for both defending and attacking cities). I haven't looked at other unit types in PAE either, they may well be more things done.

More generally, there are flanking attacks, there are 'target first' options (like the Ballista Elephant) and as Keinpferd wisely mentions there are all the UnitAIs that we can assign to units to guide the AI. Here's a useful link with more information on those.



EDIT: I should mention, before we get too carried away with redesigning and tweaking units, that I don't wish to add more units that require human art. So no Slingers, Heavy Skirmishers, Arquebusiers or such. This is due to art availability; while we could probably find almost any unit art we may want for Classical Rome or endless variations of European Musket/Riflemen we're not very likely to find what we need for Polynesia or Angkor. Indeed sometimes it's a struggle to find appropriate art for the standard units of some civilizations. Back when I first created HR I looked through all the available art and decided that there was (almost) enough to add Skirmishers to every civ and to have distinct Horsemen and Horse Archers. Even nowadays I don't think there's enough unit art for anything else to be added. For me, cultural diversity is an integral aspect of HR and I don't want to dilute that by using generic or repeated art too often. It's a different story for siege and vehicle units of course which can get away with it much easier.

 

[Highwayhoss]

A citizen specialist? I like it! This gives Workers a unique ability that doesn't compete with Great Engineers. At the same time, it eliminates the stockpiling problem; a free citizen specialist just isn't worth the trouble. I have only one concern: after your 15 or so Workers finish building Railroads in the Industrial Era, they could all come together and join a single city. That would give the city an immediate 15 hammer/turn boost; 30 hammers/turn if it had a Forge, Factory, and power plant; and even more with an Ironworks or Heroic Epic. I don't think it matters so late in the game, but I thought I should point it out anyway.

OK that is a good point: to offset this I suggest that in addition to the +1 , add also a -1 per citizen specialist.
Also as an aside does anyone else think that we need ways like buildings for cities to boost their food output similar to the klin boosting ?

 

[Simon_Jester]

No, I disagree. The real issue is "ranged combat carries no risk." Using 2X units to attack X units always works, but you generally lose 0.5X units in the process, which you have to replace. With ranged combat, so long as you can arrange a first strike, you can bombard away then clean up at 95%+ odds. That way, your 2X units eliminate X units with virtually zero losses. You can proceed to bombard another X units next turn; and, soon enough, your 2X units have chewed through 10X worth of AI units. That's why ranged combat is fundamentally unworkable.

I'll have more comments and hopefully some concrete suggestions tomorrow.

P.S. I agree that suicide-Catapults are silly. I'm certainly not suggesting we return to that model.

What if we limit the ranged attack max damage further, at least for most units capable of doing it?

For practical purposes, no amount of artillery fire can dislodge a unit that is prepared to receive the attack- witness the Somme bombardment, which dropped literally kilotons of shells onto the German trenches over a period of a week with negligible effect. At best, you can use artillery superiority to "soften up" a stack, to make it marginally safer to engage with normal units or to fight the battle with less damage.

But if the max damage that can be inflicted by ranged units is, say, 25% of a unit's total hit points, you can't just shell a bunch of City Garrison infantry into submission and then easily brush them aside with your own troops- they're still tough opponents, and your units will be badly damaged fighting the battle.

OK that is a good point: to offset this I suggest that in addition to the +1 , add also a -1 per citizen specialist.
Also as an aside does anyone else think that we need ways like buildings for cities to boost their food output similar to the klin boosting ?

At the moment, HR still seems to suffer from "too much to build," at least in certain cities (my experience may be warped from the behavior of capitals under Monarchy). Adding more stuff might be interesting. On the other hand, more food really isn't urgently needed- cities are quite capable of feeding themselves satisfactorily in vanilla. About the only thing you can't do is build cities in barren terrain and have them grow large, which is totally appropriate.

 

[Howard Mahler]

A few thoughts on ranged combat.

I would suggest that if you are going to have units that just bombard such as the Battering Ram, then early units such as Catapults not be able to both bombard and do ranged attacks.
Also I suggest that if Skimishers have a chance to do collateral damage, Catapults and other early units should not be able to do collateral damage.

I see no problem with Bombards being able to both bombard and do ranged attacks.
Should bombards be able to do collateral damage, maybe not or maybe only a limited amount.
I see no problem with Cannons being able to do collateral damage (as well as being able to both bombard and do ranged attacks.)

While it is not a perfect fix, perhaps every time you fire or use a battering ram, catapult, etc., they take a fixed amount of damage (wear and tear that has to be repaired.) This would slow down their use somewhat.

 

[Azoth]

I thought it would be useful to republish these posts.

My original argument against Siege units, some two months ago:

Siege Units: I don't think the current implementation of siege units works very well; the AI certainly can't handle it. I understand why you made the change from BtS: using suicide-catapults to soften up enemy garrisons was very silly. So now siege units can (1) lower city defenses and (2) make ranged attacks. The problem is that (1) is too weak and (2) is ridiculously strong. Let me explain. (1) Since Walls and Castles are now available early, cities are very resistant to bombardment. I often see the poor AI bring 5 Swordsmen, 5 Horsemen, and 10 Battering Rams to a city, then wait for 10 turns for the castle to crumble, letting the defender stack units in the city and easily resist the attack. For this reason, I don't bother building the early siege units at all. Which leads me to... (2) Making ranged attacks at no risk to your own units is too strong. My standard HR military strategy is to build a lot of Catapults, defend them with whatever units are available, then march to every enemy city, ignore the walls/castles, and simply use ranged attacks to weaken the defenders, who can then be taken out at 80-100% odds. It's possible to do the same thing in BtS but at least you lose a lot of Catapults in the process, which you have to replace. In HR, you risk nothing. A smart human player might try to flank all those Catapults away with mounted units, but the AI isn't clever enough, so it's doomed. (As a side note, disallowing regular attacks also makes it almost impossible to earn XP on siege units, and unlock something like the Accuracy promotion. Speaking of which, I don't even know if City Raider promotions do anything for siege units anymore, even though the AI always takes them.)

I'm not sure how best to fix the problem. Here are some ideas I came up with.
(A) Limit the number of siege units with ranged attacks that an empire can have at one time. This is probably the easiest fix. A hard cap of 4-10, adjusted for map size, would do the trick. Siege units would still be powerful but they couldn't be everywhere at once.
(B) Remove the ranged attack abilities of siege units. Create a new class of units with collateral damage on attack, say Skirmishers. (I know Skirmishers are already supposed to have this ability, but it doesn't seem to work for me. Could it be bugged? Anyway, I'm talking about a whole class of units.) Maybe Skirmisher --> Crossbowman --> Grenadier --> whatever. These units would all have some withdrawal chance and do collateral damage to enemy stacks (like catapults in BtS.) This is the hardest fix with a lot of major changes. Skirmisher units would initiate most battles and have a high casualty rate.
(C) Allow some ranged attacks to miss, or give defenders a change to intercept ranged attacks (and deal some damage back to siege units.) This might be feel weird: it's a middle-of-the-road fix. At least this way, there would always be some uncertainly about ranged attacks.

And the first detailed outline of possible Skirmisher units, by Simon Jester:

Hmmm. Maaaybe. A 'skirmisher tree,' maybe even running up into the modern era with a dedicated "guerilla" or "partisan" unit (remember Civ II Partisans?), would be interesting.

Say... Skirmisher -> Slinger -> Sharpshooter -> Partisan

You might give them bonuses against the bugaboo units of their era, or might not. The strong unifying feature should be relatively inferior strength to pretty much any other unit of the era, one or two first strikes, but also a very good withdrawal chance, good enough that on the attack they will withdraw from battle more often than not. At 50% it's still not really worth maintaining them because you'll lose them half the time and they don't soften an enemy unit up enough to justify that.

Between those posts and this thread, we've come up with some pretty good ideas.
We can't implement all of them; so I guess the next step is to put forward concrete proposals for revised Siege and Skirmisher units.
As I see it, we are limited chiefly by:

• the ability of the AI to understand and use new units and mechanics; and
• the limited availability of unit artwork for all civilizations.

I'll try to have an Option A, B, and C for Siege and Skirmisher units up over the weekend, so we can debate the merits of various approaches to unit balance.