Understanding Firaxis Unit Values and Costs

[tom2050]

This thread is primarily a study I have been doing here and there to understand how Firaxis created unit values to their costs. This is not necessarily the same thing as Oz' thread on what the AI makes, but they both may be helpful to each other.

The reason for this is primarily: The game obviously is hardcoded to some extent on what the AI builds. Sometimes it is difficult to get the AI to build certain units; which ultimately has to do with a units' values, costs, and special abilities. The game designers most likely IMO coded the game around the values and cost system they setup for their units. For the most part, the AI seems to make units pretty well in the normal game. Although different unit values can work great, this is here to help see the 'patterns' that Firaxis had in mind, so we can use to to help emulate cost and strength balance in the game if we so desire!

 

[tom2050]

General Observations

Foot Soldiers
Defensive bombard values of foot soldiers is half of the foot soldiers' attack value.
For Defensive Units, the Attack value is usually very close to 2/3rd of that units' Defense Value. Only the Pikeman does not fall into this case, and probably for balance reasons.

Defensive
Defenders generally have the same value of defense as their corresponding fast attackers' Attack value. Later in the game this changes, whereas defenders have 2/3rd the value of defense compared to fast attackers' Attack value. Also likely to adjust for civil defense and metro-sized cities.

Artillery
Bombard values of Artillery generally keep in step & slightly above of defender defense values.

Artillery appears to get it's cost based on the "(Bombard * 10) / 2" with additional cost added in based on increased Rate of Fire.
Artillery: +1 RoF and +1 Range added an additional 20 shields to the cost.
Radar Artillery: +2 RoF and +1 Range added an additional 40 shields to the cost.

Other
The "Cost to Unit Worth" ratio (a units' statistics compared to it's shield cost) generally gets alot better in the 3rd and 4th ages; whereas it is pretty consistent in the 1st and 2nd ages.

For most units, 120 shields is the maximum a unit will cost in most unit lines. Exceptions are the 2 Nukes, a few late naval ships, nuclear sub, and stealth bomber. This is probably a balance issue to keep build times according to avg city production factors. This is a reason for 'more bang for your buck' when it comes to later game units. * Find average late game city production

 

[tom2050]

Warrior used as the "Base Unit" with a base cost of 10 shields for 1A, 1D, 1M, no special abilities.
+ Unit Value = the # of additional values a unit has over the 'Base Warrior Unit'. This includes sum of all A.D.M and Spec. Abilities.
Relative Unit Cost = The cost based on how many additional stat increases a unit has compared to the 'Base Warrior Unit'. This tends to be consistent until the 3rd and 4th ages, in which the actual cost drops compared to this relative unit cost (strength of the unit).

Cost column = Cost in shields
A.D.M = Attack / Defense / Movement
Spec. Abilities = The # of Special Abilities a unit has

Higher % is better for Artillery; RoF vastly increases effectiveness from stat-perspective.

 

[tom2050]

What this means:

Relative Shield Cost
100% means that the cost of each additional 'unit value' over and above the base 1.1.1 was equal to a 10 shield cost.

A lower % (60%) means that you are getting many more 'unit values' at a much cheaper shield cost. In this case, each 'value' is only costing 6 shields instead of the normal 10.

A higher % (120%) means that you are paying more shields for each 'unit value'. This is the case with the Musketman. Higher % for artillery doesn't mean it is worse; in fact it's probably much better bang for your buck; it's just that I wasn't sure how to include higher RoF in the calculation at this point.

This is a purely statistical calculation only; and not really a algorithmic cost formula!

 

[T-mun]

Very nice study, which confirms what I thought about the AI not building certain units because it considers they aren't worth it. I need to read this more in details, but be sure I will add it to my personnal reference links.

Idea: it could be interesting (but would need more time) to include Unique Units and some special units from scenarios like the Sengoku Conquest (I'm thinking Yamabushi Monk here). Sometimes the AI produces one unit like crazy and uses it for pillaging almost exclusively (I've seen waves of Yamabushi coming just to cut roads and ressources down).

 

[tom2050]

Notes:
So far there is a consistent pattern here in regards to build times and unit costs.

I will try to get the naval ships added in also. My cities were not built according to any 'pattern'. I most often irrigate plains, mine grassland, and will irrigate grassland if there are hills and mountains to use. So no super-technique was used of any sort to alter production levels from what might be considered a 'norm'.

NOTE: Core Cities are cities built by my civilization. All Cities include those built by me and those I have captured.
For Build times, the number of turns generally has to be rounded up for a city to actually complete the unit (this I did not do, since they are averages).
=================================

End of 1st Era (Despotism) (before 1st switch of govt)
Average 7.58 shields / city (12 core cities)
Average 6.17 shields / city after corruption

Near end of 1st era build-times:
Unit Shields - Avg build time - after corruption
Swordsman 40 - 5.28 turns - 6.48 turns
Horseman 30 - 3.96 turns - 4.86 turns
Archer 20 - 2.64 turns - 3.24 turns
Spearman 20 - 2.64 turns - 3.24 turns
Catapult 20 - 2.64 turns - 3.24 turns

Shields per Coastal City: Average 4.6 shields / coastal city (3.4 after corruption)
Galley 30 - 6.52 turns - 8.82 turns
Curragh 15 - 3.26 turns - 4.41 turns


=================================


End of 2nd Era (Monarchy)

Average 14.85 shields / city (13 core cities)
Average 10.92 shields / city after corruption

Average 10.76 shields / city (all 34 cities)
Average 7.21 shields / city after corruption

Near end of 2nd era build-times:
Unit Shields - Avg build time - after corruption
Calvary 80 - 5.39 turns
Musketman 60 - 4.04 turns
Longbowman 40 - 2.69 turns
Medieval Infantry 40 - 2.69 turns
Cannon 40 - 2.69 turns

Shields per Coastal City: Average 11.8 shields / coastal city
Frigate 60 - 5.08 turns
Galleon 50 - 4.24 turns
Privateer 50 - 4.24 turns
Caravel 40 - 3.93 turns

 

[Ozymandias]

Bravo!

,

Oz

 

[tom2050]

Idea: it could be interesting (but would need more time) to include Unique Units and some special units from scenarios like the Sengoku Conquest (I'm thinking Yamabushi Monk here). Sometimes the AI produces one unit like crazy and uses it for pillaging almost exclusively (I've seen waves of Yamabushi coming just to cut roads and ressources down).

Good idea. I haven't had a chance to do those yet, but I will get them done as well. Naval and air is done also, but haven't added yet... primarily to try to see their relationship of stats / cost in regards to everything else.

Bravo!

Whats up Oz!

 

[Ozymandias]

Whats up Oz!

Lots of travel while my g/f handles kitchen reno. Also, new book contract in hand ... crazed on the road ... Oh, and it'll take awhile, but I've decided which of the mods I've been tinkering with to release first: "Warlords Of The Air" (redux)

- And you?

All Best Wishes,

Oz