Best unit promotion

[Pangaea]

If you can be general, what is the best unit promotion based on how combat is calculated? Is Combat I and/or City Raider I the best you can get if you plan on taking cities, or will you be better off taking for example Shock or Cover depending on whether you face warriors/axes or archers?

 

[Zx Zero Zx]

Versus Axes and Archers City Raider will give you a bit more damage.

 

[diagonale]

Hi Pangaea, thank for opening this interesting question.

I heard all and the opposite about units promotions. There are three thing I consider important: don't trap yourself into dogma guidelines (always this, never that, et caetera), try use varied promotion on your troops and try to have units able to handle various situations (both defense and attack). Well, there's a fourth: promote according to your next moves. The last two could be somewhat antagonists so it's up to you to find a balance between them.

Practically, I'd advice against city raider promotion because those units are like unpromoted if on the field. 2 or 3 counter units in your city raider stack won't do it against an active defense. So to answer to your question, I think combat is better because of this and because its 10% bonus apply to the attacker while city raider apply to the defender that may be wounded and/or weak. Drill is good too particularly if you are protective; the question combat or drill is flawed because a mix of both is possible and often better balanced. Flanking and combat horses are good too (likely a mix of both).

That said, your attack stack should have a bit of counter units. shock arbalest, guerilla arbalest, formation lance or forest axe/mace are all welcome. Same logic later in the game. When you attack you don't choose which unit defend so it's likely those specialized promotion won't help that much to take cities However you could give them the easy battles to reach the level4.

It leads me to guerillaIII and forestIII. GuerillaIII gives +25% attack against hills (negate hill defense). It also gives +50% chances of withdrawal. It's as much as flankingII plus tactics together (but tactics is only for great generals). That is, this promotion worth three and is often underrated. Guerilla III troops are good in all situation (including city attack) which makes them very valuable. ForestIII. It's +50% attack against forest (negate forest defense). It's +15% healing on the square (as much as guerisonIII which's a great general promotion). And it's +2 first attacks (as much as drillIV). Worth three promotions too. Such unit, beside being a great defender (it can take a full stack alone by defending on a forest square) and a good healer can make a decent attacker against lower strength units (it barely get wounded thank to first attacks). In consequence you can promote it a bit faster thru combat line and you'll obtain a real killing machine particularly once you get the technology edge and upgrade it (angry shaman gambit).

 

[Zx Zero Zx]

-.- Why would you have a stack of 100% CR units? That's what the shock axe and the spears are for..... He asked for purely taking cities, not field wars.

 

[diagonale]

I answer just in case you honestly didn't got my point.

First, I don't know what is his/her knowledge about stack composition so I give him/her a full answer
Second, You have to get to those cities and may have to fight on the road; Possibly to clear the way thru those hills/forests strings. Or to kill an enemy stack which head toward the city you're going to attack to defend it. Or because you get attacked. Et caetera.
Third, you may have to defend those cities if you don't raze them.
Fourth, my point is that you're better served by units that you can use both for attack and defense so obviously my answer is not city taking only because some other things have to be considered.

Now, I don't mind you answer with only one sentence but if having me do complete answer bother you I don't care at all. Just let me breath, young one.

 

[coanda]

Generally...
City Raider for attacking cities (siege and regular).
Barrage for siege attacking in the field.
Usually combat for attacking stacks in the field, or attacking cities if you can't get CR... but there are exceptions. Specific promotions like Shock are sometimes preferable depending on the makeup of your stack and the enemy stack.
Promotions that further boost a unit's strength if it is primarily there to defend a stack - e.g., if you have some axes mainly to protect your swords from enemy axes... give your axes Shock.
Drill in extremely uneven fights (where one side is massively favored... I'm talking Rifle vs. collateral-crippled axe uneven).
CI Cover I units for AGG leaders instead of CI CRI if they are attacking strong archer/longbow top defenders and expect the attacker to die.
Flanking I and II sometimes for suicidal horse-archer, conquistador, or cavalry attacks against archers/longbows with several first strikes. But not always; sometimes it's better to just do a bit more damage by getting Combat II.
One unit with Sentry in a mounted stack can be useful.
Commando usually for units that get past Combat IV.
All stacks should at least have a Medic I unit, if they don't have a Great Medic.

Sometimes, promotions that boost a specific civ or unit - Guerilla III is great for Celts against hilltop cities, more Drill promos. for Oromos, and so forth.

 

[noto2]

I find in the early game I specialize my promotions a bit, I'll use CR and shock quite a lot, and give my mounted units flanking. Later on in the game, though, I find the specialized promotions just almost always disappoint me, so I end up going very general. I give practically everything combat, except siege gets CR and if I have a tech lead I might go drill. That's about it.

 

[Pangaea]

I usually put accuracy on some siege and then barrage on the rest for collateral damage. So it's better with city raider on siege?

I don't know if this site does everything correct, but I've seen in the game too that the first combat ups the win change a great deal, while the second combat does little. I find that odd. For example, going by that site, axeman vs archer in a non-hill city with 0% cultural defence.

Axeman v Archer 25% fortify: c. 28% win chance
Axeman Combat I v Archer 25% fortify: c. 58% win chance
Axeman Combat II v Archer 25% fortify: c. 64% win chance

So maybe Combat I is a good idea for the first promotion (for both attack and defence) and then City Raider I for attacking cities. That gets the odds up quite a bit. At least at my level (Noble) most defenders are in cities, so I tend to fight few battles out in the field (mostly active defence).

What is arbalast btw? Can't see that name in the unit list nor the promotion list. Not lance either for that matter.

Unfortunately Shock doesn't seem to do anything on that site I linked.

In any case, it makes sense to mix and match the promotions a bit, at least when you get past the very early rush-ish phase of the game. Having a medic is always beneficial too. The problem I have with Guerrilla promotions past the early game is that most of the forest and even jungle tends to be cut down, so their usefulness diminishes. If you can get them to level 3 I see there is a good benefit, but that can take a while - especially if you put one promotion into combat, as I've done recently.

 

[coanda]

The deal with siege and city raider is that it helps you get the top defender(s) down to a manageable range more quickly... which is what in turn really determines your losses (since once the best defenders are whittled down, and everyone else has taken at least some collateral damage, you can often just run off all your siege units in a row with wins). If you were up against a large stack of identical defenders, collateral might be preferable, but you never are.

The way combat works is that it's the attackers strength relative to the defenders, but that determines both the odds of winning each individual round of combat, and the amount of damage done if they win the round. A small boost to strength will usually make a small change in the odds, unless it pushes the damage over a threshold which changes how many rounds need to be won. Since combat continues until one of the units drops below 0 HP... if both units start at full health (100 HP), hitting for 20 damage is no more or less useful than hitting for 24 damage (either way, you have to win 5 rounds to end the combat). But hitting for 19 damage is significantly less useful, as suddenly you need to win 6 rounds.
Evenly matched units need to win 5 rounds. But if one unit is ever so infinitesimally stronger than the other (1.000001 strength vs. 1.0 strength), the weaker unit will do only 19 damage and need to win 6 rounds instead, which makes a significant change in the combat odds.
The deal with the axe vs. archer issue you considered is that the Combat I promotion just happens to push it over that threshold - it goes from 5.0 strength vs. 5.25, to 5.5 vs 5.25... which means it goes from axe has to win 6 and archer only 5, to axe has to win 5 and archer 6. You just picked a lucky example for Combat I.
If, for example, you add a 20% city defense bonus to the archer as well...
Suddenly it is 5.0 vs. 5.85 with no promotions. You'll see a small boost in combat odds with Combat I, where it becomes 5.5 vs. 5.85. Then you'll see a big boost in odds with Combat II, which makes it 6.0 vs. 5.85.

I'm guessing arbalest is crossbow and lance is knight, for some other country's release of the game.

Guerilla is a bonus on hills; Woodsman is the one that is less useful when forest/jungle has been cleared. Guerilla III is quite nice for cracking open well-defended hilltop cities (+25% hill attack and +50% withdrawal); while generally not nice enough to be worth burning 3 promotions on, it may be worth 2 promotions as Celts when your unique Walls building gives a free Guerilla I.

 

[VoiceOfUnreason]

I usually put accuracy on some siege and then barrage on the rest for collateral damage. So it's better with city raider on siege?

For catapults, the difference between unpromoted collateral damage and promoted collateral damage is pretty insignificant. For details, review the infinite collateral damage thread of doom; just be sure to read all of it - the plot twist at the end matters.

For front line catapults attacking cities, I'll promote along the city raider line. For backline catapults - the guys who I expect to survive because the top defenders have been taken out of play - it's been suggested that running these guys up the drill line is a favorable play. Not sure I've seen a good analysis.

It's probably simpler to assume that the back line catapults are just waiting their turn to be front line catapults, in which case CR promotions make sense there too.

 

[diagonale]

Iknow about the calculations, graphs, simulations et caetera. The thing is despite giving some rough hint they don't handle the complexity of the game at all. They won't do the strategics choice for you; Else robots would play well. Example combatI is a great deal against equivalent unpromoted units, but combatII is a great deal against equivalent combatI units and so on. City defense may play a role here too. No math will tell you what would do better for the next 100 turns of a given board. It's good to have your unit survive now, it's better to have it still be appropriate in the next situation you encounter.

About leveling the units. Personally I like to use stables/vassalage/theocracy to produce the level3 units. Then I use battles and attach great generals to reach the level4. In 3.19 vanilla bts a great general dispatch twenty experience on the stack, it may be:
- one experience to twenty units
- two experience to ten units
- three experience to four units and four to two other including the attached one.
- four experience to five units
- five experience to four units

Some prefer to settle generals to produce level3 units. Some use it on only one unit to promote it more; often to make a guerisonIII unit. I prefer forestIII plus guerisonI for that matter - not that harder to reach thank to the first strikes. And I prefer few level4 units now that more level3 units later. Of course here again possibilities can be mixed; and a given board may call for one or the other.

For a guerilla/woodsman unit I suggest to stick on this line until you reach the III one. If you feel you need those combats promotion then simply do less woodsman (in case of axes). However you'll probably have few wounded unit to finish so a couple of woodsman can make it here.

Also I tend to use few siege and to prefer spys plus withdrawal units to take out first defenders.

P.S: by arbalest Imeant crossbow;by lance I meant pike. Sorry I'm not used to the english names.

 

[vicawoo]

City raider 1 = combat 1 at 120% a total defensive bonus (pretty simple reasoning, defender gets about 100% bonus, so combat being applied to the attackers strength counts double).

Take city raider if it significantly increases your odds of winning (which it often does).

I don't like the barrage is always useless argument, because it's assuming a bunch of small bonuses boosts sums up to a small boost. If I can get (almost) full collateral off of 4 units instead of 5, that's a big boost, even if 10 collateral damage looks the same as 12.

 

[VoiceOfUnreason]

If I can get (almost) full collateral off of 4 units instead of 5, that's a big boost, even if 10 collateral damage looks the same as 12.

What scenario are you considering that offers 10 vs 12 collateral damage?

What's your evidence that the is a big difference between 60 hp defenders and 52hp defenders?

What's your evidence that the difference is larger than the advantage given to you by the city raider promotion?

 

[vicawoo]

5 strength vs 5 strength should be 10% damage (pretty sure collateral is 10 x weighted ratio).

If there's no big difference between 60 hp defenders and 52 hp defenders (8 hp), maybe the whole idea of collateral is overrated. But I assume it is, and small differences can add up to reasonable utility.

That's why it's worth checking how it changes the odds (which you can then use to estimate damage). I've seen city raider give a 10% increase in chance to win very often, which probably is more extra damage than barrage can ever hope to inflict, I've also seen it go from 5% to 7%, pre-longbows.

 

[VoiceOfUnreason]

5 strength vs 5 strength should be 10% damage (pretty sure collateral is 10 x weighted ratio).

Yeah, that's fine. I normally think about catapults vs longbows, where collateral damage is 9/10/13/18; and then further screwed up by checking against a 3.13 collateral calculator, instead of using formulas matching the 3.19 patch.

That was my bad.

If there's no big difference between 60 hp defenders and 52 hp defenders (8 hp), maybe the whole idea of collateral is overrated. But I assume it is, and small differences can add up to reasonable utility.

Yes, but not really for that reason.

Which is to say - 8hp of difference are interesting because that's a really wide range, and it's reasonably likely that there's a jump point in it somewhere, which improves your follow up odds quite a bit.

Whether or not there is a jump in that range is (a) predictable but (b) depends a lot on the specific circumstances (defensive bonuses, strengths of the attackers, and so on).


That all said, I'll stand by "insignificant". I'm usually more concerned with defending longbows than I am with defending axes, and by then the weaker units are pretty irrelevant.

 

[Seraiel]

Iknow about the calculations, graphs, simulations et caetera. The thing is despite giving some rough hint they don't handle the complexity of the game at all. They won't do the strategics choice for you; Else robots would play well. Example combatI is a great deal against equivalent unpromoted units, but combatII is a great deal against equivalent combatI units and so on. City defense may play a role here too. No math will tell you what would do better for the next 100 turns of a given board. It's good to have your unit survive now, it's better to have it still be appropriate in the next situation you encounter.

I had to comment on this one. If your units survive 100 turns, you're either not enough in war or you have some insane battleing skills I cannot imagine. In my games, when I have enough units, I conquer someone. Units die, and new ones get built and reach the survivors in form of reinforcements. After a war, my military is either used up, so I have to play peacefully for some time, or I instantly declare war on someone else. This results in very early domination victories and just to make that sure, I have had units that survived from the beginning till the end, but those are the absolute exception to the rule. A normal unit in my games has a lifespan of 1-2 wars as a maximum.

About leveling the units. Personally I like to use stables/vassalage/theocracy to produce the level3 units. Then I use battles and attach great generals to reach the level4. In 3.19 vanilla bts a great general dispatch twenty experience on the stack, it may be:
- one experience to twenty units
- two experience to ten units
- three experience to four units and four to two other including the attached one.
- four experience to five units
- five experience to four units

Splitting up the XP of a GG can be a good move, no argueing about that. Anyhow, how many GGs do you get in your games? In mine, I maybe get 3 which I'll make Superhealers / Superdefenders and after that, I'll build military Academies in some cities, I never get more than 5 GGs unless I'm IMP, before the game is won.

Some prefer to settle generals to produce level3 units. Some use it on only one unit to promote it more; often to make a guerisonIII unit. I prefer forestIII plus guerisonI for that matter - not that harder to reach thank to the first strikes. And I prefer few level4 units now that more level3 units later. Of course here again possibilities can be mixed; and a given board may call for one or the other.

Forrest 3 + Garrison? What would be the function of such a unit? Garrison 3 gives greater chances when defending cities, and Forrests get chopped over the game, this is against your argument of a unit being suitable for 100+ turns.

I agree that settling a GG is often a poor choice as AI does that, and one can get their GGs by conquering their cities.

For a guerilla/woodsman unit I suggest to stick on this line until you reach the III one. If you feel you need those combats promotion then simply do less woodsman (in case of axes). However you'll probably have few wounded unit to finish so a couple of woodsman can make it here.

What you say is, that regarding the WM / Guerilla line, everything but the 3rd lvl is weak, and that is true. That is also the biggest argument against these promotions, or like Coanda says, "Gurrilla makes sense for Gallic Warriors because they get lvl 1 for free" .

Also I tend to use few siege and to prefer spys plus withdrawal units to take out first defenders.

P.S: by arbalest Imeant crossbow;by lance I meant pike. Sorry I'm not used to the english names.

If you use XBows and Pikes, you're using 1 move units, you should (!) be using siege, even if you use Spies and Withdrawl units, because Siege > all.

Not using siege is something for wars with mounted units, and it results in serious losses if one isn't already in the Cavs + Airships era.

--------------

Back to the OP:

You'll want CR for most of your units.
You'll want some anti-units specialized for field-defense, for the case of a counter-attack (Drill-units to not be weakened by the collateral, Shock-Axes, Formation-Spears, etc.)
You'll maybe want a GG-Super-Defender (you'd be astonished how funny it is, when AI suicides its whole stack on a single CG3 + X + X + X XBow / Greandier / MG. )
You'll definately want a GG-Supermedic.

Difference between Combat / Shock / Cover and CR is, that the one gets applied to the defender, the other gets substracted from the defender, but this is a minor thing that only makes a difference when the gap is very large, i. e. Praetorians against Archers, in this case the bonus on the Praetorian is greater then the malus for the Archer (meaning, Combat / Shock / Cover > CR in this special case) .

In general, CR is the way to go, Combat is the way to go for few (or many if having mounted only where CR isn't available) and Drill btw is neither bugged nor calculated wrongly. To understand Drill, you simply have to learn that "Drill is for units where the difference in power isn't big or in your favor" and "Drill units take less dmg than normal units" , "they aren't stronger though" so "they die if they're too weak" . I. e. a Drill-XBow is a good idea against a Mace, because this is a close call (9 vs 8) , it's even a better call when fighting a LB (6 vs 6) but it's a terrible call when fighting a Knight (6 vs 10) even if the Knight would not be ignoring Firststrikes. The more you lack in STR, the better CR and Combat are, if it's close, give Drill a chance, and always have some Drill-promoted anti-siege-specialists in your Stack if you use 1-move-units.

 

[coanda]

Collateral damage uses almost the same formula as regular damage does -
floor(10*(3*A+D)/(3*D+A)). There are two differences; first, the base damage is 10 instead of 20, and all combat modifiers other than barrage are ignored (some collateral-causing units don't use their whole strength for the purpose of calculating damage).

What the Barrage promotion does is replace A with 1.2*A.
Ignoring the "floor" round-off, the ratio of damage of barrage to non-barrage is then...
[(3.6A + D)*(3D+A)]/[(3D + 1.2A)*(3A+D)].
Expressing that all in terms of R = A/D, the ratio of attacker base strength to collateral-damaged defender base strength,
Damage ratio = [(3.6R + 1)*(3 + R)]/[(3 + 1.2R)*(3R + 1)]
= (3.6R^2 + 11.8R + 3)/(3.6R^2 + 10.2R + 3).

That's a straightforward one-variable function you can plot if you want. But in case you're lazy, here's what you'll find:
The % damage boost from barrage is negligible when the defender is much stronger (i.e., R << 1). Catapults vs. machine guns, barrage just doesn't matter.
The % damage boost rises fairly quickly to a peak of 9.5% when the attacker is ~80-100% of the strength of the defender (which happens to line up roughly with catapult vs. longbow, or cannon vs. Rifle).
The % damage boost then falls off slowly and asymptotically towards 0 as A >> D - cannon vs. archer, you're down to just a 6% boost in damage.
It should be noted that tanks have no collateral without Barrage, so it has a bit more of an impact in that situation...

That all represents the average damage boost from Barrage promotions to siege - anywhere from 5-10%, depending on how even the fight is. Because of the floor round-off in the damage, it can actually vary a bit more.

End conclusion: for same-era matches, expect Barrage to mean doing 1 more point of damage. Barrage II is another point of damage on top of that, and Barrage III is 1-2 more damage again.
For example, cat. vs. longbow starts out at 9 damage. With Barrage 1 it's 10, with Barrage II it's 11, with Barrage III it's 12.

You can debate whether it's worth giving up City Raider promotions to get that; personally, I ignore Barrage.

 

[VoiceOfUnreason]

For example, cat. vs. longbow starts out at 9 damage. With Barrage 1 it's 10, with Barrage II it's 11, with Barrage III it's 12.

Umm....

That result doesn't match the code I was reading today, and it doesn't match the experiments I did with world builder today.

 

[coanda]

Umm....

That result doesn't match the code I was reading today, and it doesn't match the experiments I did with world builder today.

If http://forums.civfanatics.com/showthread.php?t=137615 is wrong, I'd be delighted to know it. I was working entirely off of that (detailed work below). If it is wrong, please correct me and let me know what the actual mechanics of collateral damage are.

Catapult: 5 strength. Longbow: 6 strength.
floor(10*(3*5 + 6)/(3*6 + 5)) = floor(10*(21/23)) = 9.
floor(10*(3*5*1.2 + 6)/(3*6 + 5*1.2)) = floor(10*(24/24)) = 10.
floor(10*(3*5*1.5 + 6)/(3*6 + 5*1.5)) = floor(10*(28.5/25.5)) = 11.
floor(10*(3*5*2.0 + 6)/(3*6 + 5*2.0)) = floor(10*(36/28)) = 12.

Edit: So this anomaly got me curious enough to WB test myself; you're quite right. It's 9-10-13-18, not 9-10-11-12.
Which in turn got me curious enough to go dive into the code, which found the resolution.
Rather than, as the Combat Explained article suggests, Barrage promotions boosting the unit's effective combat strength for collateral damage... they just multiply the pre-promotion collateral damage after that has been calculated.
Barrage I means you do 20% more collateral damage (and it's just bad luck that it rounds just under 11 damage, so you only see a very small boost).
Barrage II means you do 50% more collateral damage (hence 13, 9*1.5).
Barrage III means you do 100% more collateral damage (18, 9*2.0).
Which means Barrage may be more worthwhile than I've been giving it credit for. Barrage III catapults do about twice as much collateral damage as regular.

 

[happyturtle]

City Raider siege tends to survive more battles with higher hit points, making it easier to keep the war train moving. I like a few accuracy siege units for quickly bombarding a city down to 0% defence.

Flanking is wonderful for taking out enemy SODs in the field. Seeing all those cats and trebs get wounded makes me smile.

For attacking/defending units, I'm just not sure. I'm not good enough at war yet.

I prefer settling GGs in my HE city, because I find if I send them out in the field, I get all precious about trying to keep my special unit safe, which leads to making poor war choices. The AI is stupid about settling GGs because it can't specialize cities. Three cities with one settled GG each is definitely far inferior to one city with three GGs.

I'm still very noob at war, but getting better. My last game, I actually fought some wars where I was at tech parity and did okay. Improvement!