Quick Answers / 'Newbie' Questions

That makes perfect sense. Having the power lead has a major impact.

 
Does anyone have an idea of the criteria for when the AI will attack armies?

I note that the AI rarely if ever attacks full health cavalry armies although the army defense strength will be only 4 or 5. Mind you I have yet to pit such armies against tanks.

By way of a contrast, in a game I am still playing - my first at demigod level - a load of knights from a mixed stack hurled themselves at a city I had just captured where the main defender was a 7/17 health veteran 4 unit mech inf army with a nominal defense strength of 36. Ten veteran knights arrived in one turn and inflicted precisely one hit on the Mech inf army. A turn later the Medieval infantry and longbows (all veteran) from the stack had a go despite the Mech inf having healed up five health points to 11/18 over the previous interturn.

The other point of note was that the stack contained riflemen (plus pikes and spears) none of whom attacked. Would this be units marked as defensive not taking part in assaults?

Normal order was restored another turn later when the mech inf army was back to full health - the AI left it well alone
 
Does anyone have an idea of the criteria for when the AI will attack armies?

I note that the AI rarely if ever attacks full health cavalry armies although the army defense strength will be only 4 or 5. Mind you I have yet to pit such armies against tanks.

By way of a contrast, in a game I am still playing - my first at demigod level - a load of knights from a mixed stack hurled themselves at a city I had just captured where the main defender was a 7/17 health veteran 4 unit mech inf army with a nominal defense strength of 36. Ten veteran knights arrived in one turn and inflicted precisely one hit on the Mech inf army. A turn later the Medieval infantry and longbows (all veteran) from the stack had a go despite the Mech inf having healed up five health points to 11/18 over the previous interturn.

The other point of note was that the stack contained riflemen (plus pikes and spears) none of whom attacked. Would this be units marked as defensive not taking part in assaults?

Normal order was restored another turn later when the mech inf army was back to full health - the AI left it well alone
There is a threshold above which the AI will not attack an army. I think it is the odds of winning the whole engagement, rather than the absolute strength of the army. I do not know what it is.

The cases where the threshold does not apply include when the army is in a city, and if there is no other possible attack on the continent.
 
If the AI really 'wants/needs' to take a particular city, Armies will always be attacked if they are the top defender, regardless of stats.

Out in the field, though, "it depends" on relative A/D values and HP remaining.

Basically, the AI won't attack an Army (or any unit) directly, if it 'thinks' that its best attacking unit's probability of victory is insufficiently 'high'. While I have not conducted any systematic tests, just from having watched what the AI does under various circumstances, I would estimate that the threshold at which the AI decides that an attack is worthwhile, is ~10% probability of winning all possible combat-rounds.

That is, the AI generally avoids attacking Armies, because it is comparing each potential attacker's strength to that of the Army on an individual basis, and finding each potential attacker too 'weak' due to the Army's usually very high HP. It does not understand that it would likely kill the Army if it threw multiple units at the Army, one after the other.

(When a large stack of units also includes bombard-capable units, the AI will similarly also tend to send its attackers past the stack to go after a 'softer' target(s), rather than attacking the stack directly)

Conversely, if the AI can build and use its own Armies or other high-HP units (e.g. if running the Flintlock patch, or a mod containing individual units which have +6-8 bonus-HP), it will use those to attack enemy Armies.

And when an Army has been badly injured, it will also become fair game for attacks by individual (low-HP) units — and if those attackers have M=2 or more, the Army will also be unable to retreat.

Some examples from my past games:

AI-controlled ACavs (which have +1 HP by default, so an eACav has 6 HP) will attack reasonably healthy (green-yellow HP-bar) Berserk-Armies.

AI-controlled Cavalries will attack badly-damaged (red HP-bar) Infantry Armies — including one memorable instance where Khan used a stack of 20+ Ironclads to bomb my 3-vet-Inf-Army — fortified in a Barricaded Fortress, blocking a 1-tile land-bridge to his section of the Continent — down to 1 HP, before sending in his Cavs to kill it (which were then trapped in the Barricade, so became easy pickings for all my Arty + Cavs(?) during the following turn).

If not defended by mobile AA-capable units (and sometimes even when they are!), Armies will also be preferentially targeted by AI-Bombers (and by bombard-capable land-units, if running the Flintlock patch)

And I played one game (Player1 Fanatic's tweak-mod, Large 70% Archi, as the Celts, IIRC) where, during a Modern-Age war towards Domination, Cathy abandoned the AI's usual "No First Use" policy to nuke my stacked MechInf-Armies, apparently because she was losing so badly.
 
Not sure, but I wonder this for any unit.
A vet or elite Rifleman on a mountain will be left alone for a loooong time.
Think armies are the same as other units, its just that they have waaay more HP so the odds of winning drop, so the AI will not engage. Quite intelligent, really.
I think its just the pure odds with perhaps a RNG of +- some percent.
Some turns they do, some turns they dont.
 
It also has to do with what the AI's best options are. It will gladly send units along a road for great efficiency™ instead of attacking a unit that is just out of sight because it knows that your unit cannot attack it that turn.
 
My experience is similar to @tjs282 in that I have had knight armies or cavalry armies attacked by the AI when they are yellow-lined or red-lined. My armies have been bombed by AI bombers, especially if they are in a city.
 
I think that the only time in recent years that an army of mine has been attacked was when it was an army of Keshiks at half health. Even so, it survived the lone Spanish longbowman and one or two turns later it was again fighting at full capability.
 
Keshiks, are they def 2? Def 3 last most of the game, but 2 has a shelf of attack 6 units in my experience. Of course in town or down to red are always at risk.

edit corrected spelling
 
Last edited:
Keshiks, are they def 2? Def 3 last most of the game, but 2 has a shelf of attack 6 units in my experience. Of course in town or down to red are allows at risk.
Yes, A.D.M=4.2.2. Their UU-"superpowers" are a 10s price-discount and to ignore the move-cost of Mountains (but not Hills), because riding across Mountains (but not Hills) is of course what the Mongols were famous for... :rolleyes:
 
Weren't the Mongols a steppe people?!
Well yes, exactly: high, dry, cold, plains.

(Wiki sez that) the Mongols adopted the habit of attacking their targets during winter, by following the courses of and crossing frozen-over rivers. Which presumably means that they went along valleys: around the mountains, not over them. And that they gathered information on those targets by extensive scouting before they attacked.

(Wiki also sez that) their primary weapon was the composite bow, loosed from horseback during a charge; and that they could travel unprecedented distances by bringing multiple spare horses along on every campaign, and not relying on logistical supply-chains, because they lived off the land (and also the spare horses, in a pinch!).

So if anything, the Keshik (rather than the Ansar) should have had A.D.M=4.2.3, and maybe also the "Radar" flag and/or B.R.F-stats (even if only def-bomb, 2.0.1)...
 
In my current game (Monarch/China/Pangaea), I started wondering: how does one judge whether a city needs a courthouse or not? I'm basically eyeballing it ... if a city looks like it could benefit from a courthouse, I'll typically build it (and rush it, especially if a captured city far from the capital). Obviously the core cities do not need one, but how to calculate where the core "ends" is something I haven't thought much about yet.
 
I'm generally inclined to build a Courthouse in any town where corruption/waste is already >30% (and the easiest way for me to eyeball that is to look at shields harvested vs. wasted). Might also build one in a town closer than that to the Palace, to get a neater shield-output number, e.g. if I'm getting 8-9 spt usable plus 2-3 shields wasted, a Courthouse might bring that up to 10-12 spt for more efficient builds.

Because I have come to loathe ICS city-pox, I have also started spacing out my boonie-towns a little more, and building 'Ducts (where needed) to get them all to ~Pop12. If my own Lux-supplies are limited and/or those coming in from overseas Civs may get cut off by wars, and/or WW is likely to be an issue, and I feel the boonie-town(s) 'need' a Market to maintain happiness with the Luxes I control, then I'll also (first) put up a Courthouse. A fully-roaded Republic-town at Pop12 with half the citizens as Specialists should give a minimum of 12 commerce per turn (and more if on a river), so even at the 90% corruption-cap, a Courthouse (brings corruption down to 80%, i.e. 2-3 uncorrupted Commerce per turn) should still at least 'pay' for all 3 improvements — though that may not be worthwhile in the late game...

But I will not be doing much if any cash-rushing in those boonie-towns: first I'll work the high-food tiles to get the town to maximum-pop (whether hard-capped by freshwater limits or soft-capped by discontent), then convert excess pop-points to CivEngs to get the buildings up, then switch them back to Geeks afterwards.

e.g. in a Pop6 town, 3 Civ Engs = 6 (+1) spt, so will build a non-Agri Duct in (just over) 14T, or a Court in (just over) 11T; and at Pop12, 6 CEs = 12 (+1) spt, which will build a Court in (just over) 6T, or a Market in (just under) 8T — less with Forest-chops and/or unit-disbands.
 
In core cities, by the time you're building those massive 1000-shield game-deciding wonders such as the UN or the Internet -or spaceship parts- then yes, you do need the extra shields. Everything counts, especially if by that time you have built a few factories and power plants and multipliers get wild.
 
In my current game (Monarch/China/Pangaea), I started wondering: how does one judge whether a city needs a courthouse or not? I'm basically eyeballing it ... if a city looks like it could benefit from a courthouse, I'll typically build it (and rush it, especially if a captured city far from the capital). Obviously the core cities do not need one, but how to calculate where the core "ends" is something I haven't thought much about yet.
Agree with the others. I would add that sometimes an AI capital city, even if it's far from your core, can be a potentially strong contributor to your empire. A Courthouse might make sense there to help it contribute more. Overall, I tend to build them where a) 4 or fewer red corrupted shields or b) NOT where I've built my Forbidden Palace or Palace. Most of these are located within 10 or 15 tiles of the capital.
 
FP + Courthouse caps corruption at 10% (or 15%? FP alone caps it at 30%, and adding a Police Station later will reduce corruption to zero). So if you've already decided where you want your FP, it might actually be worth building a Court first (assuming you're not getting dogpiled and so have the shields to spare), to get the FP built faster later.

In my current DG game as Carthage (Large Random Continents), my coastal capital has only about 10 land-tiles in its BFC. My FP town is fully landlocked, ~7 tiles NW, and was getting about twice the shields for most of the game: even after building an Offshore Platform (I've got the UN, but I'm going for Space), Carthage is still lagging behind Leptis Minor.

[Red text = edit added later]
 
Last edited:
I started wondering: how does one judge whether a city needs a courthouse or not?
As a very simplifying rule of thumb you can estimate that a courthouse will reduce corruption by about a third of what it was without it. So starting at about 30% corruption courthouses tend to pay of their costs reasonable quick. In the long run however even courthouses, that take longer to pay for their costs, are reasonable.

 
It would appear I made a huge mistake in: 1) not prebuilding in preparation for Hoover Dam; and 2) having traded Electronics to the Iroquois in the first place. A couple turns ago, I got Shanghai down to 17 turns remaining to match Salamanca. Now, 2 turns later, I'm at 15 and Salamanca is at 11. I've tried everything I can think of to boost shield production (building a factory at this point would not improve matters), but I just can't get it down far enough. Mines; joined a worker to bring population to 12; no room for specialists, though it would only get turns down to 14, anyway. Would it be prudent to drop it now & build something else useful (notwithstanding a waste of shields), and then plan to prebuild for the UN (a race I definitely do not want to lose)? I'll have Steel in a couple turns, so I'm thinking I start the palace prebuild for the UN after that. Thoughts?

(P.S.: Monarch/China/Standard Pangaea.)
 
Back
Top Bottom