AoE II DE Infosheet: Jineapple’s Work Updated

Overview

Some may recall there was quite a handy Infosheet made for an HD (2013) version of AoE II by Jineapple. There, on just 8 pages, all the civ, tech and unit stats were put together with a convenient styling. In this guide, its updated version for Definitive Edition is presented, alongside with some updated auxiliary info.

Download / Preview Links

PDF[drive.google.com] and Google Docs versions of the DE Infosheet are available @ Google Drive[drive.google.com]. It is updated to Hotfix 50700 version of the game (July 14, 2021).

Original Infosheet by Jineapple is presented here[www.aoe2sheet.info] and in a Steam guide below.
[link]

Intro

Some of you might remember the original brochure for AoK (AoE II: Age of Kings). I don’t – I believe I could have it torn apart when I was a kid 11; but I do remember AoE I: Rise of Rome tech chart and recently discovered one for AoE III.

There’s a theme that this chart is replicated as AoE II is updated frequently these days. An effort on re-designing the initial tech-chart was undertaken here by XTwo. A good collection of facts about AoE II: HD was performed here by Philippe le Bon. All this info and design ideas were later used in a masterpiece called ‘The best AoE2 Info Sheet’ made by Jineapple for AoC, HD and its DLCs where most of the stats and ‘technical’ facts about AoE II were put together in one piece and with a good styling. Used to refer to it a lot back in the day.

For DE, tech trees and a civ bonuses are available here[aoe2techtree.net] and here (by Cruzza) with in-game descriptions unmodified; note that in the presented Infosheet, I modified in-game descriptions heavily. An expanded tooltip mod for DE is available here[www.aoezone.net] (by harooooo).

Standing on the shoulders of the giants and with a permission granted, I decided to update Jineapple’s work for it to match the modern times of Definitive Edition that happens to be updated monthly. Questions about unit stats and/or some hidden mechanics are quite frequent, so I think HD Infosheet deserves to be updated.

Sources

To collect the data, I mostly used Advanced Genie Editor (currently it’s kept up-to-date by the devs in ‘Tools_Builds’ folder of your Steam AoE II: DE folder – usually it is ‘%Disk_Volume%:Program Files (x86)SteamsteamappscommonAoE2DE’). Still, lots of testing in Scenario Editor had to be conducted – either to confirm the values or to study how some of the game mechanics work.

There are other sources I used (or that inspired me to test something) – they are mentioned in the ‘External Links’ section of this guide. Here, let’s just list gentlemanly attributes of AoE players’ ‘technical’ sources (if one wants the answers, not tests):

https://www.reddit.com/r/aoe2/ – some ‘technical’ info was borrowed from there. Mechanics discussion arises daily there as well;
Wikia[ageofempires.fandom.com] – kept up-to-date; I find it a good quick source of changelog and up-to-date info;
AoE Zone[www.aoezone.net] – community hub with several technical discussions there;
AoKHeaven’s University[aok.heavengames.com] has some good technical articles of AoK/AoC days;
https://www.youtube.com/channel/UChzLZJo-SxuPHz-oYKAIC_g literally monopolized AoE II technical discussion there – and for a good reason.

Modifications of the Original Infosheet and Auxiliary Data

Plenty of them, aside from it being updated to DE balance state:

I decided to mention rate and time modifiers together where possible (civ and tech descriptions). There are sometimes comments stating that [for example] Japanese infantry actually attack 25% 33% 25% 33% faster. And it is actually hard to understand without testing or delving into AGE (and then testing anyway) what the bonus actually implies.
Explanation: most of the variables in game are coded in time, measured in seconds. (I believe it is a common practice for games which have to sync themselves with each game simulation frame.) Speed, or rate, is the inverse variable of time, and thus any modifier applied to time is inverted if we are talking about rates (when we say a car is ‘faster’, we imply it has a higher speed, AFAIK). (E. g.: 0.8 modifier for time is 1 / 0.8 = 1.25 modifier for rate.) There are a few exceptions to this rule –like costs, resource collection, and construction rates and production team bonuses that are rate modifiers– that can also be misleading (as, for example, you could expect Spanish villagers to finish a Castle in 30% less time while it’s actually ~23% in reality).
Also, see Spirit of the Law video on that topic. He can definitely explain it way better than me.
Added several ‘technical’ explanations (mostly speculations sadly) that may have an impact on the unit performance as footnotes in the units’ chart. E. g.: how does Letis attack work? why does Elite Chu Ko Nu have a reload time of ~3.6/3.85/4.1 sec and unupgraded has a set reload of 3.6 sec when the coded Reload Time for both is 3 sec?
Added ‘Attack Animation Duration’ variable to unit stats. It simply can’t be ignored as it affects attack delay (AD), (practical) reload time of melee units, and raiding potential of the units. See legend for unit chart and ‘Auxiliary Info: Specific Game Mechanics’ section here for more details.
Also in the mentioned legend, more info about reload time, accuracy and collision sizes of the units is added. Projectile speeds are mentioned in ‘Auxiliary Info: Hidden Effects’ section in this guide.
Some info is updated in auxiliary info; added some ‘Game Mechanics’ explanations. Unmodified paragraphs from the original Jineapple’s guide are highlighted in “quotes” (so that one could skip them if familiar with the original guide).
Updated gather rates (adding trade using a formula published on Wikia[ageofempires.fandom.com] confirmed via testing) and moved them to a separate table of the chart.
Added some external links as sources or ‘further reading’ materials. As I’m a relatively new arrival in the AoE II community, I could have missed some important articles/guides/videos, so make sure to pass them if you know any.

The Google Doc version is the one I was working on, so it’s at least somewhat readable. Other programs may butcher the formatting harshly, and some charts (namely the tech chart) just can’t fit into one page without resizing (hence its PDF-version can be pixelated badly). Aside from the units’ chart legend, the whole infosheet should be ‘Ctrl+F-able.’

If you manage to spot any mistake in stats, spelling, formatting or some unclear wording – make sure to ask and/or report it here in a comment section.

List of Abbreviations, Terms, and Clarifications

AD – attack delay, or fire delay: time between attack animation starts and damage is actually inflicted (for ranged units – start of the animation → the projectile is released). Not to be confused with ‘Frame Delay’ variable: the latter only partly determines the attack delay. (See this article[www.aoezone.net] for more info.)
AGE (Advanced Genie Editor) – modding tool that can access the game data files to look for most coded variables. It is currently updated by the devs and is situated in ‘Tools_Builds’ folder of your Steam AoE II: DE folder – usually it is ‘%Disk_Volume%:Program Files (x86)SteamsteamappscommonAoE2DE’.
Atk = attack, damage.
Attack animation duration – graphical representation of the attack. It does matter in damage timings in AoE II, see the legend for unit infosheet in PDF for more details.
B1 and B2 – separate armor classes of buildings. B1 armor class is upgraded via Masonry and Architecture, B2 is uncounterable. In AGE, B1 = 11 armor class (‘All buildings’), B2 = 21 armor class (‘Standard Buildings’).
BBC / BBT – Bombard Cannon / Bombard Tower.
BS – Blacksmith.
CA – cavalry archers. ‘Cavalry Archer’ capitalized refers to a particular unit from the Archery Range, ‘cavalry archers’ – to a class of units.
FA/CA/IA – Feudal, Castle, and Imperial Age.
FU – no, it means ‘fully upgraded’ (for a respective civ).
Gather rates – in this guide, these are the rates of resource generation. In most cases (except for farming), coded (i. e. theoretical) values are presented (they will be lower in practice due to reasons like upgrades, walking time, bumping, collision box size etc.).
Generic [civ] – a civ which has no specific civ bonuses in the given example.
HC – Hand Cannoneer.
Hill/Elevation bonus – units deal more and receive less damage while occupying an elevated position relative to the other units. In this guide, it applies both to negative and positive effects of the elevation, as well as to cliff bonus (bonus damage but no bonus defense).
LoS – line of sight.
[Attack] Reload time – coded timings which separates chunks of damage inflicted by the unit. Not to be confused with the attack delay and attack animation duration.
SotL = Spirit of the Law, AoE II YouTuber.
SW = Siege Workshop.
TC – Town Center.
UT = unique technology, UU = unique unit.
‘X-line’ (e. g. Militia-line) – means all the units within the same upgrade tree (e. g.: Militia → Man-at-Arms → Long Swordsman → Two-Handed Swordsman → Champion).

Additional abbreviations/clarifications are mentioned in the Legend for Unit Infosheet (see PDF or Google Docs). I may have missed some – can be probably found here on Wikia[ageofempires.fandom.com].

Screenshots of the Infosheet

Note that I was not initially formatting the Infosheet for it to be presented in screenshots, so they might look a bit off. Legend for Unit Infosheet is jpeg as it was transparent by default (which would look quite repelling in Steam preview).

Civilization Bonuses

Tech Chart

Legend for Unit Infosheet

Unit Infosheet

Gather Rates

Auxiliary Info: Monk Mechanics

As Steam has a limit for a guide’s section length, auxiliary info is divided into several groups. Refer to Google Docs version[docs.google.com] for better rubrication.

Basic monk facts:

Monks and Missionaries heal 2.5 HP/sec, or 1 HP/0.4 sec (used to be 1/0.8 HP / sec for Missionaries in HD). Garrisoning in buildings only heals 0.1 HP/sec for most of them (except for Khmer houses), or 0.2 HP/sec for Castles and Kreposts; Herbal Medicine increases this effect ×6 times (used to be ×4 in HD).
“Illumination actually increases faith regeneration by 87.5%, from 1.6 points per second to 3 points per second” (i. e. Monk will recharge in ~33 sec instead of 62.5 sec).
In DE, Missionaries are affected with Bloodlines, Fervor (used to be in HD as well) and Husbandry (new addition).

Monk Conversion

Here, Monk conversion mechanics are described briefly. For more info, see articles ‘How Monks Really Work’, part one[aok.heavengames.com] (AoC) and part two[www.aoezone.net] (HD). Quintessence info can also be found in Jineapple’s guide to original Infosheet.

Monks (as units) have a coded work rate of 1.25 and an accuracy of 25%. In aforementioned articles, the work rate is referred to as ‘Monk seconds’, as it simply translates into the conversion time one to one. Monks also have minimum and maximum conversion time for units set to:

{4;10} × 1.25 = {5;6.25;7.5;8.75;10;11.25;12.5} sec,

where “;” stands for ‘or’, and for buildings:

{15;25} × 1.25 = {18.75…31.25} sec.

Accuracy (25%) is applied as a conversion chance every ‘Monk second’ (starting at minimum conversion time). However, in both mentioned articles it is stated that the first conversion chance is actually 28%. From my own tests in DE (20 tests × 100 units conversions), I received a 26.75% initial chance for conversion. See the sources for more info if interested.

There are a few things that can affect conversion chances:

“The Scout- and Eagle-line have a significantly lower chance of being converted. It’s somewhere around 4–5% per 1.25 sec.” Max conversion time can also exceed 1.25 sec for them (see HD article[www.aoezone.net] for a possible reason of that);
“The Teutons team bonus increases the min conversion time by 1.25 sec, max conversion time by 2.5 sec and also halves the chances of your units being converted, to around 13% per 1.25 sec”;
“Faith increases minimum conversion time by 2.5 sec, maximum conversion time by 5 sec and lowers the chances of conversion to around 9% per 1.25 sec, i. e. roughly a third. Buildings are also affected by the Teutons/Faith bonus;”
Faith and Teutons team bonus stack additively;
The Inquisition tech lowers min/max time by 1.25 sec for units and by 6.75 sec for buildings. It also increases conversion chance for buildings by about 2%.

Auxiliary Info: Hidden / Unlisted Effects

Hidden Effects

If the projectile/arrow hits the target the attacker was not aiming for, it deals 50% damage. If the ‘missed’ projectile still hits the intended target, damage is not modified (i. e. is fully inflicted). For a hit to be inflicted, the projectile (that has its own collision size, usually 0.1×0.1 tiles) must hit the collision box of the unit (see legend for unit infosheet for collision box sizes of most units).
If the unit misses, the projectile will be off-set by a parameter called ‘Attack Dispersion’. (See legend for unit infosheet for the exact values.) The attack dispersion variable allows to get the size of the circle where ‘missing’ projectiles will land which is equal to {range × attack dispersion} in diameter. It seems that if the attacker misses, it will usually miss by a lot and will tend to hit spots that are
Attack Dispersion × Range ÷ 2

tiles away from the initial target slightly more often. It should also be noted that this ‘randomness’ is quite unpredictable: for example, during one 4 hour test run Janissary had ~51.5% total accuracy (base accuracy is 50% meaning that 1.5% of the ‘missed’ shots still hit the intended target), while in the other run with test setup unmodified accuracy was close to 60%.

An example of 4 hour test run of 12 range, 80% accuracy, 0.33 attack dispersion Elite Longbow firing at Bombard Cannons (0.5×0.5 collision box radius).
Corpses indicate targets that were hit with ‘missed’ projectiles

From the screenshot above one can observe that a FU Elite Longbow firing from 12 tiles range does not hit all the intended targets within a 2 tile radius from the initial target

(12 × 0.33 / 2 = 2).

Note that ‘missed’ projectiles with a positive arc value (e. g.: gunpowder projectiles) will not ‘fall’ to the ground and may still collide with the targets their ‘missed’ shots were not planning to hit.
Also, see on this topic.
Projectiles (arrows, bullets, cannonballs) have different speeds (in tiles per second):
- 1.95 (Cannon Galleon);
- 3 (Bombard Tower);
- 3.5 (Onager-line, Trebuchet);
- 4 (Bombard Cannon);
- 5.5 (Slinger, bullets);
- 6 (Scorpions, Ballista Elephant; war ships, War Wagon, Harbor);
- 7 (most arrows and melee ranged units, Arambai; Spanish Cannon Galleon);
- 7.8 (Turtle Ship).
The projectile travel time is further affected by different projectile arcs.
Arquebus’ effect is relatively small as it grants +0.2 tiles per second to BBCs/BBTs and +0.5 tiles/sec to the other gunpowder units.
“Ballistics does not affect melee units with ranged attack (Mameluke, Throwing Axeman, Gbeto) or Gunpowder units (except the Bombard Tower; however the cannonball is slow enough to still be easily dodgeable)”. Fire Ship-line has Ballistics by default.
- Arquebus does not require Ballistics and enables tracking effect for any bullet projectile (Hand Cannoneer, main and secondary projectiles of Organ Gun, Bombard Cannon, Cannon Galleon, and Bombard Tower). BBT can get tracking effect either from Ballistics or Arquebus; combining the techs has no effect.
“Trebuchets have a fixed 80% accuracy against 1×1 and 2×2 buildings as opposed to the standard 15%.” Warwolf UT does not modify this value.
To be in range of a building/unit, an object must be separated by a range value stated in HUD (‘x’) which is calculated as the distance between the edge of the attacker’s collision box and the center edge of the victim’s collision box (diagonals are treated as hypotenuses). However, once engaged in the fight, the attacker may attack from the same position if the target moves away from the stated max range ‘x’ for a certain small distance ‘y’ (i.e., from a total range of x + y). This ‘y’ is hard to measure, as it seems to be calculated at some point in between the attacks; possibly it is in range of 0.5 tiles. This is quite important in case of Mangonel vs tower fight as it allows a few shots to land on a retreating Mangonel.
See

I was curious how the range of a tower works so fired up scenario editor. Also found a bug while testing: if you walk towards the tower you will be in range when stepping on the 8th tile as it should work, however when you are in range and walking away you will be still in range til mid 9th tile. from aoe2

.
It takes two swings for lumberjacks to cut down generic trees and start collecting wood (with a generic rate of 0.39 w/sec). Lumberjacks’ ‘attack’ animation occupies 1.5 sec and, as the tree is basically ‘killed’, the last ‘attack’ occupies only half of the animation (as in case of the units fighting each other). It means the chopping part of the wood gathering takes 2.25 sec in total. Sappers tech cuts down (pun intended) the swings requirement to one, or to just 0.75 sec.
Most multipliers stack together multiplicatively. E. g.: Lithuanian Halb after Squires moves at 1.21 tiles per second:
1 [base speed] × 1.1 [Lithuanian bonus] × 1.1 [Squires] = 1.21.

Several staggered (per Age) multipliers are usually coded as multipliers [of each other] so that the outcome would match a stated bonus. This results in some odd situations where Vikings Berserk has 61 HP in Castle Age instead of expected 62 HP:

54 [base Berserk HP] × 1.15 [expected bonus HP in Castle Age] = 62.1 HP.

Unlisted Bonuses of Ages

There are several bonuses for buildings and some units when advancing.

“Scout Cavalry-line LoS increases by 2 with every age.” Eagle-line LoS is increased by 1 in Feudal Age only, but, like other infantry, it is also affected with free tracking (+2 LoS).
Scouts get +2 attack and +0.35 tiles/sec speed in Feudal Age.
Eagle Scouts get +3 attack, +2 bonus dmg vs cavalry, +1 bonus dmg vs. camels and ships in Castle Age; Eagle Scouts are also produced 25 sec faster starting in Castle Age.”
All infantry receive Tracking technology immediately upon reaching Feudal Age (Barracks not required).
“Outpost LoS increases by 2 with every age.”
“All Buildings except Outposts, Towers, Castles and Walls will gain +1/+1 armor with every Age. (Most start with 0/7 and end up with 3/10, excluding other techs).” Newly unlocked buildings (e. g. Monastery in Castle Age) already have their armor upgraded.
“Most Buildings will gain HP through the Ages. The usual number is 1200/1500/1800/2100 through the ages.” Stone Walls and Gates only have 900/1375 HP in Feudal but double the amount in Castle Age (to the original value in the base game). Watch Towers have 700 HP in Feudal Age, 1020 in Caste and Imperial Age.
“Drop-off buildings have 600/800/1000/1000 HP. Houses have 550/750/900/900 HP.”

Auxiliary Info: Specific Game Mechanics

Note that this section contains my own speculations. Prepare your grains of salt prior to reading it.

If a melee unit (‘attacker’) is chasing another (‘victim’), Attack Reload Time usually does not matter, as the time difference between the attacks is higher than it. This is because whenever the attacker starts its Attack Animation (and stays in it until it’s completed), the victim has a spare time to run for its life at full speed while the attacker is stationary; this gap must then be closed again by the attacker to finish the loop. As you can see in the Unit Infosheet, Attack Animation Durations are always >1 sec and ~1.4 sec for cavalry (for which the time it takes to close the gap is the shortest). If microed properly, this mechanic can be sort of abused as the damage is dealt precisely at 50% of Attack Animation Duration (you can see that pros sometimes do it with Scouts).
(Self-advertisement) I made a slightly bigger post on that topic @ aoezone[www.aoezone.net].
Something of a more speculative (yet backed by some testing) nature is that Attack Reload Time is actually resampled to match Attack Animation Duration: melee units can only deal damage at the start or at the half of Attack Animation Duration. From what I have been able to observe, the first hit is usually inflicted at 0.5x (where ‘x’ is Attack Animation Duration), the second hit is inflicted at 1.5x, and then the damage inflicting frames are resampled so that in a far future (the point of time is obscure to me, however) the unit will reach its coded Attack Reload Time.
E. g.: (Elite) Keshik has a Reload Time of 1.9 sec, Attack Animation Time of 1.4 sec. The first hit will be inflicted at 0.7 sec, the second at 2.1 sec, and later hits will fluctuate between 1.4 and 2.1 sec time frames so that Keshik would reach its coded 1.9 sec Reload Time at some point of the future.
Playing around with Fire Ship reload time, I found out it
Comment from discussion akimbofmg9’s comment from discussion "Fire Ships don’t attack at 0.25; Byz Fire Ships only attack 15.6% faster; Greek Fire causes Fire Ships to miss more often".

(inversely, 60 FPS/Hz) at ×1.0 game speed and >200 FPS. My speculation is that this is basic game refresh/update rate (tickrate, game simulation speed or however else it is called in the other games), at least for combat mechanics. It is backed up by testing of vanilla 0.25 sec reload time, though ×8.0 game speed deviates from predicted values.
Indirect justification: in the post by Mark Terrano here[news.ycombinator.com], he commented:

…There were actually two RNG’s in the game, one was synchronized with the same start seed on all machines (basically the same random pool) for combat and whatnot – the other was unsynchronized and used for animation variance, etc -things that weren’t gameplay related.

It means that 1) there is probably a set refresh rate (it could be 20 FPS/Hz in AoK days, according to ‘turns’ length in this old article[www.gamasutra.com] by Mark Terrano and Paul Bettner) and 2) there are two separate refresh rates for something like rate of fire and something like animations (I think I’ve encountered the difference between them while looking at the

How does Keshik gold generation ability work? (An attempt to explore it) from aoe2

, though of course I might be wrong).

I believe these timings to (partly) be the cause of known issues like aforementioned fire ships erratic Reload Time, other Reload Times usually mentioned as ‘2.03 sec’ instead of coded 2 sec (stems out from this article on AoC reload times[aok.heavengames.com], I presume),

Some updated & completed Farming Workrate Values for the tail end of 2017 from aoe2

, and monk conversion timings offsets. (The latter used to be more pronounced in AoC/HD, according to ‘How Monks Really Work’ articles, but in DE, I found these effects to be negligible as every ‘monk second’ is only off-set by 1 / 60 sec thus building up to 0.16 sec at max unit conversion time.) There could be more applications of this factor.
Integers. Resource requirements, creation time (but not production rate of the buildings), and some damage calculations seem to be rounded to the closest integer, residuals not remembered. Several implications include:
- Mayans Stone Walls cost 3 stone / 1 tile (same with other cost reduction bonuses) even while constructing 2+ sections;
- Incas team bonus (faster farm re-seed) is slightly worse as 7.5 sec construction time is rounded to 8 sec (other construction bonuses modify work rate of villagers instead).
  This is also the case for unit creation time modified by Aztecs civ bonus, Turks team bonus, Ethiopian Royal Heirs and Cumans Steppe Husbandry UT (other production modifiers affect rates instead).
- Wheelbarrow and Hand Cart. Their multipliers for carry capacity are 1.25 and 1.5, respectively (so, up to 1.875 when stacked), but the result is 2× carry instead. My guess is that carry capacity must be an integer and thus is rounded after every research following these calculations:
  Wheelbarrow: 10 × 1.25 = 12.5 → round to 13 carry;
  Hand Cart = 13 × 1.5 = 19.5 → round to 20 carry.
- Extra projectiles having no dmg applied to them (additional Mangonel-line projectiles) or just 1 dmg (Longboat, Harbor) inflict 1 dmg only if they hit the initial target and deal no damage at all if they hit an unintended target.
- Tenths of damage are remembered and are inflicted as soon as 1 additional damage is “banked” (i. e.: if a unit attacks downhill and should deal 1.25 damage per hit, it will be graphically represented as a victim unit receives 2 damage once per every 4 hits, starting from the first one. This means that the first attack will kill the victim if it has just 2 HP).

Auxiliary Info: Elevation and Cliffs

“If units/buildings are uphill relative to an enemy target, they will deal 25% more damage and take 25% less. Increasing the height difference won’t change anything. In case both are on the same floor level, no one gets bonus damage/defense.”
Hill bonus is applied if the center of the building/unit occupies an elevation or its slope (see the screenshot below) relative to its attacker/victim. In other words, as long as one of the units’ center is higher relative to the other’s, it will get the hill bonus.
(Do not forget about isometry!)
Here, second and sixth Kreposts partly occupy the slope of the hill but receive no benefit from it. Third and fifth Kreposts’ center is on the slope of the hill, but hill bonus is still applied.
“If units/buildings are on what looks like the top of a cliff, they deal 25% more damage but still take the same amount of damage. Number difference of several cliffs won’t change anything. Hill + cliff do not stack, in that case you only get +25% damage (cliff) and +25% defense (hill). In case both are on top of a cliff, no one gets bonus damage/defense.”
Tatars receive +25% bonus damage (additively, not multiplicatively; i.e. +50% in total) firing from the elevated position or/and behind the cliffs, with their buildings affected as well. Their bonus grants no defensive advantage compared with the other civs. (See .)
“The one standing on a hill deals +25% damage against another one on top of a cliff, but the latter also gets +25% damage (cliff bonus damage not only annulates hill bonus defense, but overwrites it completely).”
Hill/cliff bonus has no effect –neither negative, nor positive– if the initial attack was to deal 0 dmg but instead was rounded up to 1. However, if the attack was meant to deal 1 dmg with no rounding going on, hill/cliff bonus is still applied. E. g.: FU Arbalest (10 pierce dmg) will deal 0 → [rounded up to] 1 dmg to Elite Huskarl regardless of elevation; FU Elite Longbowman, on the other hand, will deal 2 dmg to Elite Huskarl every fourth attack if occupying a higher elevation / being on top of the cliff.
“Bonus damage is calculated when the target gets hit. So if your Arbalests stands on a hill but dies before his arrow hits, it won’t deal 125% damage.” And vice versa: if the unit dies prior to hitting a target occupying a higher elevation, it will deliver full damage. This trick is sometimes used in a Mangonel vs Mangonel fight: the one that is firing uphill is deleted and thus can deliever its full damage to the one occupying the elevation.
- This is very important in Mangonel vs Crossbow or Mangonel vs Mangonel fights: if firing uphill, Mangonel is not able to one-shot a single Crossbow or an enemy Mangonel. However, if it dies before the shot has landed, the elevation factor is “forgotten” about and the full damage is dealt resulting in a killshot within a small blast radius. (And extra Mangonel projectiles will deal 1 damage if they hit the intended target.) This is sometimes exploited as a ‘delete trick’ in high level games (see by TWest).
- The Blast Damage drop off nature is also altered if the Mangonel is deleted: instead of an immediate linear drop, it deals full damage in a certain radius (0.5 tiles?). This is important when fighting against a clumped up group of ranged units as it allows to one-shot more at once. (See by theHellRazor and
  
  Mangonel delete (left) and Mangonel alive
  
  [imgur.com] on this screenshot.)
- Unlike in HD, Trebuchets are exempt to this rule: if Trebuchet is able to pack while its projectile is in the air (i.e. with Kataparuto UT), it still deals full damage (including anti-building damage). However, if it is destroyed, the elevation bonus is still ignored.
- As damage is calculated upon the hit, not upon the shot, hill bonus may be achieved if the attacker moved to an elevation after the initial shot. (See
  
  T-West Showed How Deleting your Unit Makes Onager Shots Stronger, I Now Confirm that Moving up a Hill while the Projectile is Still in the Air also gives the Projectile the Hill Bonus. from aoe2
  
  by Tudsmfa.) Also, if an upgrade has been researched while the projectile is in the air, upgraded unit damage will be applied (e.g.: Archer shoots an arrow and is upgraded into a Crossbowman a split second later, before the arrow hits the target: Crossbowman’s damage will be used for damage calculations. See a https://www.reddit.com/user/Tudsamfa/comments/iacs1h/interesting_mechanic_no_idea_what_happens/?utm_source=share&utm_medium=web2x).
I believe that hill bonus is applied after all other damage calculations. Justification: Tatars’ BBTs deal 1 damage to an unpacked Trebuchet if firing downhill.

Auxiliary Info: Multiple Arrows

Only the first projectile of a volley from buildings and units is affected by a hill/cliff attack bonus (or penalty), which means Castles on a hill take less damage, but hardly deal more. Also, as Town Centers technically only fire extra projectiles, they gain no elevation bonus/negative damage at all.
“The first arrow of a building will actually just transmit the attack value of said Tower/Castle. The other arrows are extra “units” that have different attack values.” (E. g.: Krepost has 9 pierce atk for the first arrow but the remaining arrows have 11 pierce atk as they are similar to Castle additional arrows.)
All the upgrades except for Heated Shot affect all the arrows of buildings (with an exception of Harbor). Heated shot affects the first arrow only.
“Only the first arrow of a Chu Ko Nu/Kipchak is affected by any tech.” (They use the same projectile.) Ballistics does affect all their arrows.
If a unit/building fires multiple projectiles, the reload speed may get slowed down and exceed the coded value.
Projectiles spawn in groups of 1 or 2 (randomly; possibly checking if there are no other projectiles in the spawning tile) in bursts separated by ~0.05–0.07 sec (depending on the unit/building). Reload time is applied after the last arrow with a small delay. Chu Ko Nu / Scorpion and Ballista Elephant after Double Crossbow upgrade are affected by their attack delay multiple times.
In the case of Chu Ko Nu, practical Reload Time is:

{3 sec [coded Reload Time] + ~2 × 0.28 sec / arrow [i. e. circa AD + ~0.05 sec] ≈ 3.56 sec.

Elite CKN has Reload Time of ~3.58/3.85/4.13 sec as arrows are (randomly) launched in 3–5 bursts. Ballista Elephant and Scorpion with Double Crossbow may –or may not– be affected once with their attack delay in between the two projectiles.

This effect is taken into account in the Infosheet. Organ Guns and Mangonel-line are not affected.

Extra arrows through garrisoned units

Formula with an example:
1. “Calculate the damage per second (dps) of the building and of the unit. This is the current attack value divided by Reload Time (see the Infosheet for values).
  Thus, a Castle has 11/2 = 5.5 dps (assuming it only has a single arrow) and an Arbalests has 6/2 = 3 dps (both without blacksmith upgrades).”
  The attack value of extra projectiles of the building is applied, however. It means that for Krepost, 5.5 dps is used as a baseline (instead of 4.5 based on the first arrow’s 9 atk).
2. “Divide unit dps by building dps to get the impact of one unit. Multiply by the number of garrisoned units, round it down and you get the total number of extra arrows.
  In said example you get 3/5.5 = 0.5454… This value is rounded down, so 1 Arbalests does not add anything, 2 Arbalests will add 1 arrow, 4 Arbalests 2 arrows etc.”
Upgrades affect dps of both garrisoned units and projectiles of the building in DE. “The stronger the unit’s dps (and the weaker the building’s dps), the higher your number of extra arrows can get.”
“Only archers, cav archers and gunpowder units will add extra arrows. Ranged units dealing Melee damage (Mameluke, Throwing Axeman, Gbeto) won’t add anything.”
“Note that Chu Ko Nu and Kipchaks are really bad when garrisoned, because the formula takes the dps of their secondary arrows instead of their main one.”
“Villagers are an exception, as they are factored in with a fixed dps value of 2.5 (unaffected by Supremacy / Inca villager upgrades) which is also the case for any Teuton infantry with Crenellations researched.
Garrisoned Villagers are a huge damage boost in the beginning, but their influence decreases with more upgrades researched.”
“There is a special calculation for Bombard Towers: they will always get one less extra projectile than they should according to the formula.” (Possibly due to initially having only 1 max total projectile.) ”If you pack a Bombard Tower with five Hand Cannoneers, you should get one extra shot (theoretically, because it’s not the case). Teutons though can garrison up to 10 units, so they have enough space for several cannonballs: 9 Hand Cannoneers or 7 ally Elite Janissary for 2nd, 10 ally E. Janissary for 3rd cannonball.”

Auxiliary Info: Blast Damage

Do not forget that area of effect is coded as Blast Radius. E.g.: Mangonel has 1 Blast Radius, Onager – 1.25; the latter has its blast damage covering 56% more area compared with the former, as it’s squared.
“Battering Rams have 0 Blast Radius (so no blast damage). But Capped Rams have 1.5 and Siege Rams 2 Blast Radius, which means they can destroy two layers of wall at once.”
“Capped/Siege Rams deal the same amount of damage in the whole radius, but it only applies to buildings.” Same applies to Petards, but their blast impacts all the units within the area of effect.
Mangonels have 1, Onagers 1.25 and Siege Onagers 1.5 Blast Radius with damage dropping linearly. Onagers and Siege Onagers can fell trees. Additional projectiles deal 1 damage only if they hit the intended target, including buildings (and no damage with attack ground).
Logistica (Byzantines) and Druzhina (Slavs) grant Cataphracts/infantry 5 uncounterable damage per attack (not affected by armor) with 0.5 Blast Radius (this is a 1 tile box centered at the center of the attacking unit).
“Elite War Elephants/Battle Elephants have 0.5/0.4 Blast Radius, respectively.” War Elephants deal half the damage they would do in a direct attack, Battle Elephants deal 25% splash damage.
Bombard Cannons, Warwolf Trebuchets, and Turtle Ships have 0.5 Blast Radius, Flaming Camels 2, and Demolition Ship line 2.5/3/3.5. Their damage drops linearly.
- Note that for Flaming Camels and Demolition Ship-line, blast damage is calculated in relation to the center of the demo unit; however, as they are separated by their collision box radius away from the target (0.25 tiles for Flaming Camel, 0.5 tile for demos), they will never inflict their full damage (with an exception of passable objects like Town Centers). Their blast damage (incl. bonus damages) drops linearly, so Flaming Camels usually do not inflict higher than 87.5% of theoretical damage, demo ships – 80–86%.
- Warwolf Trebuchets’ and Bombard Cannons’ splash damage also drops linearly. (Speculation) However, the Blast Radius is so small that it seems to matter which part of the collision box of the target (imagine it being a parallelepiped with width, length, and height) is hit by the projectile; usually the closest part of said parallelepiped is hit by the treb/BBC, and therefore units behind the target tend to receive less damage. It could also be a reason of BBC/Warwolf Trebs not inflicting full damage to their target: the damage could be projected to the ‘floor’ tiles and thus be perceived as ‘deviating’ from the center of the target. (My other speculation is that even with 100% accuracy, the unit still does not hit that very center of the target with a possible miss by ~0.05 tiles.)
“The intuition is misleading in case of Bombard Towers and (Elite) Cannon Galleons which have no Blast Radius at all, so even slow moving targets can be missed easily.
Thus Trebuchets and (Elite) Cannon Galleons only deal half the damage when using Attack Ground, but Bombard Cannons/Trebuchets with Warwolf do full damage, without a reduction described above”.
Torsion Engines grants +0.45 Blast Radius to siege units (except Trebuchets) and +0.5 Blast Radius to rams, including Battering Ram.

Auxiliary Info: Construction and Repairing

Building with Multiple Villagers

If t(coded) is the build time for a building (with a single villager), and n is the number of villagers, the formula to calculate how long they need to construct a building t(constr) is
t(constr) = 3 × t(coded) ÷ (n+2).

An easier representation is:

t(constr) = t(coded) ÷ (0.67 + 0.33n).

“You can think of it like this: Each villager after the first one works only 1/3 as fast (or each additional 3 villagers will do the same work as the first one). First one works full-time, two will need 75% of the time (3/4), three will take 60% (3/5), four 50% (3/6) etc.”
Note that Treadmill Crane and Spanish civ bonus both affect the construction rate and thus lower the t(constr) by an inverse value:
(1 – (1 ÷ 1.2) ≈ 16.7 %

for Treadmill Crane,

1 – (1 ÷ 1.3) ≈ 23%

for Spanish civ bonus).
“As you can see in the table below, Treadmill Crane compensates about one builder for small numbers of builders.”

Time to Build,
% of t(coded)

Number of
Villagers, n

W/ Treadmill
Crane

W/ Spanish
Bonus

75%

2

1.33

1.07

50%

4

3

2.5

33%

7

5.5

4.9

25%

10

8

7.2
“Unless you need to get up your building as soon as possible (e. g. Tower, Castle, Town Center), it is better to split up builders.”
Knowing the construction time + average walking distance, you can calculate the actual lost villager-seconds (hence gathered resources) while constructing buildings and walling off. See

The cost of using multiple villagers to speed up construction from aoe2

and for implications.

Repairing

“Buildings: A single villager repairs ~750 HP/min, every additional one will add ~375 HP/min to that.”
“Siege/Ships: A single villager repairs ~187 HP/min, every additional one will add ~94 HP/min to that (¼ of building repair speed).”
“Fully repairing a building/ship/siege weapon costs half of its resources. Civ bonuses are applied e.g. Incas pay 15% less stone.
Exception of ‘fully repairing = half the construction prize’ rule is Town Center where you lose double their wood cost, but no stone (civ bonus does not work for Britons). Furthermore, you need to have at least 1 stone to be able to start repairing the TC.”
“Techs like Masonry, Architecture and Hoardings increase the total number of Health Points. As it still costs the same resources to fully repair the building, you will effectively get more HP/cost when repairing. The repair rate stays the same.”
Spanish civ bonus and Treadmill Crane have no impact on repair rate.

Auxiliary Info: Farming and Trading

Farms and Fish Traps

“When farming, villagers collect from the left 2×2 square, which makes Farms placed on the right side of a TC or Mill slightly more efficient, so it’s usually a good idea to start placing them on the right first (unless this area is more vulnerable to attack). It’s still worth it to fully surround Mills/TCs with farms.” Although on a rare occasion, villagers may gather from a center of a tile or outside the 2×2 box.
I found that in the case of farm layout as presented on a screenshot below and 10 minute test runs, farms at the top right corner ended up with 3% more food harvested compared with an averaged value, while farms at the bottom left ended up with –3.5% food harvested compared with an averaged farming rate. One can assume that the difference is ±3% (compared with an averaged value) in farming rates.

Counted food left on farms after 10 minute runs at ×1.7 game speed in 5 test runs with a test start delay of 1 sec. Generic civ, no upgrades, no farm re-seed. Sample size is quite low to exclude the randomness of villagers freezing.

As seen on a picture below, in 5 averaged 1-hour test runs (same conditions, no farm re-seed – farms modified to have 2000 food storage), one can see the difference becoming even more pronounced.

Long story short, if safety of farmers is not of a huge concern, start off by dropping farms in the top right corner and finish with the farm in the bottom left corner. Same implies to the second layer of the farms.
While farming, villagers tend to gather from 4 different spots of the farm collecting ~25% of their total carry capacity each time and then dropping the food off after hitting their carry limit. (With some upgrades, villagers may occasionally perform an extra fifth walk.) This mechanic is still in effect in the case of Khmer, so they are still affected with Wheelbarrow/Hand Cart/Heavy Plow.
Farmers have a remarkably high gather rate of 0.53 food / second. However, farms themselves also produce food at a set rate of 0.4 food / second (0.44 food/sec for the Slavs). Therefore, after a certain point (usually after Hand Cart for generic civs), farming rate is capped at 0.4 food / second, or 24 food / minute. Basically it is reached when time villagers spend walking around the farm and towards the drop-off site is small enough for villagers’ superior gather rate to outmatch the farm production rate. Taking the previous paragraph into consideration, speed and carry capacity bonuses (Wheelbarrow, Hand Cart, Heavy Plow) both allow to reduce time walking and therefore increase productivity while not directly affecting gather rates.
In DE, Fish Traps and Farming are about equally fast food income. Wheelbarrow and Hand Cart put Farms ahead of Fish Traps, but with Gillnets researched, fish traps are ~9% more productive in case of generic civs and are just slightly behind Slavs FU farmers.
In DE, fishing ships collect from any part of fish traps and can drop off food instantly if adjacent to the Dock. This means that coded fish trap gather rate can be realistically achieved.

Trading

“Markets or Docks generate significantly more gold the further they are apart. If the distance is doubled, the gold per trip more than doubles.” (See gather rates in the infosheet.)
“Trading does not depend on diplomatic stance. You can trade with neutral markets, or even enemy markets (if they’d let you). If one of your teammembers resigns with his market still intact, you can delete his market to avoid the enemies using it.”
Gold per trip is the same for Trade Carts and Trade Cogs if the distance is even; however, Trade Cogs are faster (base speed 1.32 tiles/sec vs 1 t/sec of trade Cart) and thus generate more gold per time. Berbers’ bonus and Dry Dock also increase gold generation rates in DE. (See gather rates in the infosheet.)
Trade Carts and Trade Cogs will choose the longest trade route possible from the allied market you ordered them to trade with ( is outdated).
If one of your markets gets destroyed, the trade carts will instead go to the one of your markets that is the closest to the destroyed one (without changing the amount of gold they’re carrying accordingly). If you do not have any markets left, they will stop where they are and lose any gold they were carrying. This is in contrast to HD, where they choose the closest Market to them but modify the amount of gold they’re carrying accordingly.
“If the allied market your carts trade with gets destroyed, they’ll return to your market (and drop off gold there if they were carrying it), then stop.”
In HD, Trade Cogs had a big collision box and had difficulties with pathfinding which meant significant losses caused by bumping. In DE, their collision size has been updated so that they glide alongside each other instead of zigzagging. If there is a safe and long water trade route available, it will generate more gold compared with a land one. (See .)

External Links

Mechanics, (Hidden) Effects, Military Stats, Changelogs

Jineapple’s comments on the original infosheet (unmodified sections from there are presented within “quotes” in this guide)
‘How Monks Really Work’, part 1[web.archive.org] by MrFixitOnline Staff, Neilkaz and AndyHRE [archived] (Monk mechanics in AoC)
‘How monks really work v.2 – all the details’[www.aoezone.net] by Jineapple (Monk mechanics, extended and updated for HD version)
by T-West (Monk mechanics in DE, extended and updated for DE version)
‘My research on how “Frame Delay” works’[www.aoezone.net] by Icewind (how to calculate the attack delay of archer units in DE and prior)
‘Chasing time-to-kill: Calculation for melee units’[www.aoezone.net] (self-advertisement) (time-to-kill while raiding & why attack animation duration matters)
How does Keshik gold generation ability work? (An attempt to explore it) from aoe2

(self-advertisement) (+look into ‘attack reload speed vs attack animation time duration’ problem)
by Spirit of the Law (to answer possible questions ‘why 33% if 25% is stated in tech tree in-game?’)
Fire Ships don’t attack at 0.25; Byz Fire Ships only attack 15.6% faster; Greek Fire causes Fire Ships to miss more often from aoe2

by slothismysin (Fire Ship attack speed, including Byzantines’ bonus. Tests in HD; DE has a better performance)
by T-West and ‘Armor’ article[ageofempires.fandom.com] @ ageofempires wikia
by Spirit of the Law (theoretical impact of hill bonus on the fights, as well as implications of Lanchester’s linear and square law – see also )
Balance Changelog in DE[www.forgottenempires.net] at Forgotten Empires website (compared with HD; with a changelog, last updated february, 2020). More detailed changelog is presented @ Age of Empires Wikia[ageofempires.fandom.com]