I was wondering how (even if) farming rates had changed in Age of Empires II: Definitive Edition. Unfortunately, before this guide was made, I could not find any ‘experimental’ info on that, so I tried to figure out farming rates myself. Later, Spirit of the Law uploaded his estimates to YouTube which appeared to be slightly different from what I had found out myself.

This guide offers tables of farming rates in the game including impact from Heavy Plow, Wheelbarrow and Hand Cart technologies as well as various civilization stated or hidden-but-well-known bonuses (Aztecs, Berbers, Khmer, Slavs buffs + Mayans debuff). The results are compared with the rates from Age of Empires II: HD (2013). (Note that I used my own testing results as a reference for HD rates.)
Test setup used to record farming rates in this guide is simple: 8 farms surrounding a Town Center and 10 minutes of collection time while using triggers to enable techs and automate the test (refer to ‘Test Setup’ section for more info).

After obtaining farming rates, I also looked into fish traps. One particular 'quality of life improvement' in Definitive Edition is automatic fish trap rebuild mechanic which practically makes fishtraps far more viable than they used to be in HD; improved collection mechanic from fish traps buffs them further to the point where they may be a costly but a safer and more efficient counterpart of farm, and not only for the Malay or Japanese. For more info and results, see 'What About Fish Traps?' section.

If you have no interest in looking through the whole guide, refer to ‘Table’ and ‘Highlights’ sections; if you have any difficulties with all the figures in these tables, refer to 'Short Glossary' section, header 'How to read the tables'. If you just want the brief quintessence of presented info, refer to ‘Summary’ section instead. In 'See also' section, other sources of farming rates for different versions of the game are mentioned and briefly described.

• DE — Age of Empires II: Definitive Edition;
  
• HD — Age of Empires II: HD (or 2013);
  
• Userpatch — reference to Age of Empires II: The Conquerors version of the game used in Voobly client[www.voobly.com]. Farming Rates (see below) in Userpatch differ from Age of Empires II: HD (or 2013);
  
• TC — Town Center;
  
• F / min — food per minute: ammount of food collected per one villager-minute;
  
• Farming Rate — in this guide, by ‘farming rate’ I mean an average collection rate of food from farms dependent not only on the basic farming rate in game code (the rate of food collection from the farm itself while the villager is working it) but also on walking time and possible villager bugging out / freezing for several seconds on some occasions (i. e., it simulates a more realistic in-game scenario); where needed, it is specified that the coded farming rate is meant instead;
  
• Farmrate Cap — farming mechanic that puts a limit on farming rates: a given farm may produce only 0.4 F / sec (24 F / min) in HD version of the game (it could have slightly changed in DE, however). This is a reason for mostly identical farming rates after researching Hand Cart. For further details, see this link;
  
• Game Speed — the game speed multiplier on which the game is ran compared to coded values. It does affect base farming rates as well (you will bank lesser amounts of food in the same game (timer) time). You can change it via options or 'Num+' / 'Num-' in a singleplayer;
  
• Dark Age — here, it’s equal to no upgrades;
  
• Post-Imp — Post-Imperial age; here, it's equal to all upgrades obtained;
  
• Generic Civ[ilization] — here, by ‘generic civ’ I mean a civilization that is supposed not to have neither stated nor covert (like that of Mayans’) bonus affecting farming rate (by the definition used in this guide). Therefore, it is any civ except the Aztecs, Berbers, Khmer, Mayans, and Slavs. However, it is possible that some civs received a hidden (de-)buff to their farming rates in DE (I’m not familiar with the game code to check that out);

• Figures before the "/" symbol are rates without Heavy Plow researched, figures after the "/" are for Heavy Plow researched.
  
• Farming rate — total food collected / 8 (number of villagers) * 10 (minutes);
  
• Heavy Plow impact — impact of +1 carry capacity originating from Heavy Plow tech. In several cases, negative percentages are not a mistake;
  
• Farming rate, F / min in DE (SotL) — values presented in Spirit of the Law's video on farming rates in DE given for comparison (they are in Italic as they were not obtained by me). Row 'Compared vs SotL' shows the difference of the estimates presented in this guide and SotL values (100% = SotL values); positive values mean estimates presented here are higher than those of SotL.
  Note that the methodology Spirit of the Law used to obtain the farming rates is not explained neither in the video nor in its description, so all the differences may be explained by variations in test setup (e g.: 1 hour run with farm re-seeding). Also note that I assumed SotL didn't force Heavy Plow technology (assessments given in this guide are closer to SotL's values in case of no Heavy Plow researched);
  
• % Increase vs HD — this row indicates the growth (reduction) in farming rates in DE;
  
• % Civ Bonus impact in DE — this row means specific civ bonus effect on the farming rate compared with a generic civ with the same upgrades in DE only;
  
• % Increase vs no upgrades/Wheelbarrow — this row denotes the impact of a technology on the farming rate compared with a previous technology (or no tech in case of Wheelbarrow) in DE or in HD;
  
• Wheelbarrow / Hand Cart is less effective by this ammount in DE: simple substraction (not division!) of the technologies' impact on farming rates in HD and DE. '+' means the tech is less effective now.
  
• In case of Khmer, 'stars' mean:
  – * — I haven't tested Khmer farming rates in HD and assumed them to be generic;
  – ** — In his video on Khmer farming update, Spirit of the Law mentions only % increase in farming rates compared with a generic civ without stating them explicitly. Thus, I'm quoting an increase instead (without adding this percentages to the values stated in his previoud video);
  – *** — Khmer had no farming bonus in HD, so there is no comparison in these columns.

• 2020-30-04: Update 36906 taken into account (Khmer farming rates slightly nerfed).
  
• 2020-06-04: Added coded farming rates and farming caps civ specifics from HD and DE.
  
• 2020-02-01: There is no gain in stacking fishing ships on top of fish traps in DE.
  
• 2020-01-29: HD farming rates updated. Farmrate Cap and game speed impact described and link to the post on this topic on reddit is added. Test setup description updated.
  
• 2020-01-21: Khmer bonus added. Partly tweaked the rates to take Heavy Plow upgrade into account (all the initial rates were for Heavy Plow researched). 'Highlights' and 'Summary' adjusted accordingly;
  
• 2020-01-07: added 'What About Fish Taps?' section and adjusted 'Highlights' and 'Summary' sections accordingly. Added comparison with Spirit of the Law's farming rates.

Let’s compare the numbers with those from HD. In HD, generic civ had an average farming rate of ~19.4 F / min in Dark Age. Thus, generic farming rate in DE has increased by 4.5%. This difference is in good agreement with Userpatch farming rates (see "See also" section, link 5).

After Slavs' nerf, the Aztecs are now superior Dark Age farmers, although they saw the lowest increase in their farming rates compared to HD baseline. Slavs' worst increase is definitely caused by their nerfed coded collection rate (lowered from ×1.28 to ×1.18 in DE); however, their Dark Age farming rates are almost identical to HD version. Mayans still have a nerfed collection rate in DE; however, its impact is slightly lower, despite the changes in coded rates were insignificant (from ×0.92 to ×0.93 (which a product of two separate multtipliers due to the waty Mayan civ bonus works); compare it with ×0.85 rates in HD with no DLCs). Khmer, despite having ×0.97 collection rate, still have 4% better farming rates in comparison with a generic civ.

Heavy Plow has diverse impact on farming rates depending on the civ. For the Aztecs, it only grants 1% increase; for generic civs and Mayans, the bonus is roughly 3%, for Berbers–4%, and for Slavs and Khmer it's slightly less then 5%. However, during Khmer and Slavs tests the results for separate runs were somewhat unstable (e. g. first Khmer run resulted in 23 F / min, while the others ended up with 22.1–22.2 F / min).

Prior to reconfiguring the test in DE, the disagreement between given and Spirit of the Law's estimates was pronounced, especially in case of Dark Age generic civ farming (my estimate was ~7% higher). However, after reconfiguring the test (automating it with triggers), my assessments moved closer to SotL's values. In most cases, the difference is ±1%, although the default scenario (Dark Age generic civ) in my case was more resultative, averaging on 3% higher farming rate. SotL's underestimation may be caused by fluctuations in farming rates in general (while playing with different framecaps prior to using triggers for testing, I managed to receive ±3% fluctuation in farming rates for Dark Age generic civ).

Compared with HD, most civs now experience slightly lower (mostly 1–2%) boost after researching Wheelbarrow in DE. The most pronounced differences are the instancies Aztecs (which in DE hit their Farmrate Cap with Wheelbarrow and no Heavy Plow) and Slavs with Heaby Plow.

In contrast with the case of no Wheelbarrow, the effect of Heavy Plow is a fairly modest addition for Wheelbarrow: for most civs, it's just +1.5–2% to their farming rates, while Aztecs and Khmer see no benefit from it.

With Wheelbarrow researched, farming rates are slowly getting closer to the levels of HD: the difference for Aztecs is only 1%; generic civs, Berbers and Mayans are now 3–4% faster than they used to be in HD, and Slavs' nerf is slightly more obvious with Heavy Plow now. Still, Slavs gain a lot from Wheelbarrow (+15% farming rates if no Heavy Plow) and already surpass generic civs with Hand Cart by a fair margin.

Compared with Spirit of the Law's values, the estimates given here are generally in a good agreement; the results given in this guide are 1% higher for generic civs, Berbers and Slavs with Wheelbarrow researched.

With Hand Cart, generic farming rates in DE are nearly identical to HD. Hence, this tech's effect is somewhat lower than it used to be in HD: with Heavy Plow researched, Hand Cart is 2% less effective than it used to be.

The Mayans are the one to benefit a lot from Hand Cart: their farming rates are increased by 11.5%; generic civs, Berbers and Slavs gain 4.5–6.5% boost to their farming (3–5.5% with Heavy Plow researched), while the Aztecs and Khmer boost is only 1–1.5% (which means Hand Cart can be easily delayed for these civs, unless speed of the villagers is of importance).

All the civs except for the Slavs ended up with 24.2–24.6 F / min farming rate, with Khmer and Mayans without Heavy Plow at the bottom with 24.2 F / min (1% lower than generic civs). The Slavs retain their declared 10% farming bonus in Post-Imp and thus are 3.5% worse they used to be in HD.

Another interesting thing is that Heavy Plow has no impact on farming rates in every case except for the Mayans where its bonus is +2%.
This plateau is possibly caused by reaching the hard-coded Farmrate Cap of 0.4 F / sec for a single farm (see the link for further info). An idle farm still produces this food that can be harvested later, and, as the test setup includes 5 seconds of 'warmup' (for the techs to apply and crutches to work properly), the results that hits this ceiling (achievable mostly with Hand Cart researched) will be 16 total food collected higher (0.4 [F / sec from a single farm generated by Farrate Cap] × 5 [sec] × 8 [quantity of farms]) than 1920 (24 F / min × 8 farms × 10 min). This is mostly true in HD (most of the rates hit slightly below 24.2 F / min); however, in DE this is exactly the case only for Khmer (which are already a representative of the farming mechanic); other civs, such as Berbers, Aztecs and Mayans with Heavy Plow, happen to outperform the suppossed cap by a slight margin (24.4–24.6 F / min as oppossed to anticipated 24.2 F / min) with good consistency. Note that these 16 'additional' food collected are not substracted from the rates in the tables (which means that real farming rates with Hand Cart are 0.2 F / min lower); they are taken into account in 'Highlights' and 'Summary' sections, however.

I could only speculate that the game engine fails to properly cap the farming rates (due to a different game speed, or differenet framerate, or a bug); and, if the Farmrate Cap is applied only after villagers have actually started to work the farm, the 'warmup' delay may be 1–2 seconds higher as villagers walk towards a chosen corner of the farm.

Slavs also manage to reach their ceiling, as it was reduced from ×1.15 to ×1.1 in DE (from 27.6 F / min to 26.4 F / min).

Compared with Spirit of the Law's values, estimates given here are on average 1% higher, except for the Mayans (the estimate in this guide is 1% lower). Here I'd just assume SotL used Post-Imp setting that gave the Mayans Heavy Plow in addition to Hand Cart (rates with Heavy Plow are comparable with SotL's values).

For this test, I decided to use generic civ with Heavy Plow. The result is shown on the screenshot below and is in agreement with the previous versions of the game (Ctrl + F for figure with a caption ‘1: very effective 6: least effective’ in the ref). The difference between farms 1 and 8 is 1.64 F / min (21.92 vs 20.28 F / min), or 7.8% if average food collection rate among all the farm placements is considered as a basis (100%).


In comparison with an average collecting rate (21.13 F / min obtained during the earlier testing) this test saw its average collection rate to be exactly 21 F / min (collection rate was based on the food remaining at (on? in?) farms). Surely generic civ Dark Age farming sees the most possible spread in its results.

Notes concerning farm placements:

• Farms 1 and 2 have +4–5% higher collection rates. In comparison with the Aztecs' test, they did switch places, though the difference is not relevant to prefer one placement to another.
  
• Farms 3 and 4 are just slightly above the average, with the difference being less than +1% (farm 4 nearly hit the average farming rate);
  
• Farm 5 is –1.5% worse than average;
  
• Farms 6–7 were nearly identical with their collection rate being –2.5% worse than average;
  
• Farm 8 remained the worst deviating by –3.5% from the average.

The result of these tests is simple and is in agreement with previous findings in HD: left-handed are discriminated. If defense is not of a concern, start placing your farms from top-right corner and end with bottom-left.

I then tried another farm layout presented on the screenshot below (referred to as 'alternative') to see if the layout used in the test is the most efficient one.

I thought it was logical to assume that this 'alternative' should be slightly more efficient as it favors the right-hand side of TC. I then ran three similar tests to see the differences. Total food collected was estimated based upon the food left on farms (which may not be fully representative).

In comparison with 20.95 F / min in a 'traditional' scenario, the 'alternative' resulted in 21.1 F / min, which is 0.7% higher. And the top-right farmer in 'alternative' scenario was the most efficient one, averaging exactly on 22 F / min.

Why isn't it that popular then? Well, I'd assume these 0.7% to be negligible enough to even consider, and there are more important factors that will neglect this small 'profit' such as better double-layer farm fitness (as there is no empty space inbetween the initial 8 farms) and therefore higher survival chances for the villagers for a 'traditional' farm placement method.

Test setup is available here[drive.google.com].

• Test setup was built around 8 farmers surrounding a TC as shown on a screenshot below. The same farm layout can be found in an old Spirit of the Law’s video on that topic; I also noticed that most pros tend to place their first farms using the same pattern given there are no obstacles.
  
• I’ve deviated from SotL’s methodology as I used 10 minute samples instead of 5 (for more consistency in the results).
  
• To prevent farms from expiring (yes, I decided to exclude farm re-seeding from the equation), research of Horse Collar and Crop Rotation was forced through triggers, as both of them do not affect villager farming rates in any way (and because I'm a noob in a scenario editor and don't know how to change initial farm food count).
  
• Game speed was set to 2.0× both in HD and DE. Farming Rates are affected by game speed, especially in HD. The reason I chose 2.0× is, firstly, it's available in both HD and DE, and secondly, it is faster that way (nobody can take laziness out of the equation). Another reason is that HD and DE are running at different game speeds (1.5× and 1.7×, respectively).
  
• Britons were used as an example of a 'generic civ'. Starting age was always Dark Age. Mill was added far away from a TC (just in case farming upgrades would not be forced without it).
  
• Test setup was automated via triggers both in HD and DE; however, the way it was done was different in each version of the game, as different crutches were used to force farming upgrades.
  – A required tech (Heavy Plow, Wheelbarrow, Hand Cart) was forced via triggers on par with Horse Collar and Crop Rotation which were always present;
  – Both in HD and DE, after a 5 second delay ('Triggers' → Condition & Effects → 'Condition' → 'Timer'), villagers were forced to work their farm via 'Task Object' trigger ('Effects' → 'Task Object'). Clarification: as spawning farms through 'Units' sections while forcing techs through triggers (not the starting age!) does not force them to have farming upgrades, another way of enabling them (read: crutch) must be used instead. To ensure it would work properly, 5 second delay was added. Yes, I know it made farming rates reaching Farmrate Caps slightly higher than they should be; it is specified in Hand Cart table, and the results in the other sections of the guide are modified to exclude these "overfarming" (16 food for most civs, 18–19 food for the Slavs simply subtracted from the results);
  – Both in HD and DE, villagers were tasked to drop their food off at a TC after 605 seconds (10:05 minutes, so that the test ran 10 plain minutes) via 'Task Object' trigger;
  – In HD, a blunt way of enabling farming upgrades was used: first, P1 farms were spawned and villagers were tasked to work them via 'Task Object' trigger; then, these farms were deleted and replaced (via spawning, not triggers) with P2 (enemy) farms (for most civs, Imperial Franks P2 'donor' was used; for Slavs, as enemy farms remember their Farmrate Cap based on the original owner, Imperial Slavs P2 was used instead). To ensure farms would not expire (as they tend to change their food content to a lower number after being 'converted'), Chinese allies for both P1 and P2 were added. Hopefully this rather barbaric approach did not influence the results (they are mostly in agreement with the manual test I used before);
  – In DE, newly added 'Replace Object' trigger was used as a crutch instead. Initially, P1 rice farms were spawned and villagers were forced to work them via triggers (the same way as in HD). Food drop-off was forced the same way as well. However, initial rice farms were replaced with usual farms (with no timer). Villagers still started to work them after 5 seconds.
  – In HD, 3 runs were aggregated (5 for a Dark Age generic civ). In DE, 4 runs were averaged (5 for a Dark Age generic civ and 7 for the civ + upgrades that saw the highest discrepancies in the results).

This test is of course an ideal setup. In a real game, you will have farms surrounding a Mill, double layer farms, t90 Khmer or other branded versions of farm placement (how many times has this joke been used already?) and more bumping with the obstacles on villagers’ side which will lead to a worse overall farming rate and give an upper hand to the Aztecs, Berbers, and Khmer in comparison with other civs.

• It will take 1–2 sec for the villagers to actually start farming at the start of the test.
  
• Villagers sometimes bug out for several seconds during the transition between different points of a farm. On rare occasion, villagers do some weird zigzagging around the farm;
  
• As villagers randomly choose a point of the farm from a bottom-left 2×2 corner to collect from, this part of discrepancies may be considered 'natural'–that is why several runs were used and is partly a reason for different amounts of food left for the some farm positions throughout several runs.
  
• Farm expiration is excluded–otherwise the rates would be lower, especially if the test was to be prolonged.
  
• The rates were recorded from Scenario Editor using triggers for technologies. This may result in corrupted results as sometimes Scenario Editor uses different version of the game, and forced technologies could probably overwrite buffs/debuffs of the civs (maybe that’s what happened with the Mayans w/ Hand Cart).
  
• Game speed affects the gathering rates–and it's known that it is indeed true for HD version: tests with ×1.5 speed would end up with 1.5% higher rates as opposed to ×2.0 speed.
  I ran several (five for each) tests with generic civ Dark Age farming for ×1.5, ×1.7 and ×2.0 game speeds, and the results are quite interesting: compared with ×2.0 speed, ×1.7 speed tests ended up with 1.6% (±0.8%) higher farming rates while ×1.5 tests were still higher but only by 0.5% (±0.4%). This means that the actual farming rates in games with normal speed in DE should be 1.5% higher than those stated in this guide, unless the values hit the ceiling of 0.4 F / sec set by Farmrate Cap mechanic.
  
• And who knows, maybe the collection rates depend on your FPS? I can call out other game titles where your framerate would affect your in-game performance (TES and CoD, for example) so I wouldn't be that surprised this may be the case in single- and, in the worst case, multiplayer of the game with an engine originating from the previous century.
  To look on the FPS impact, framerate was capped via RivaTuner. Five generic civ with Heavy Plow runs were used for testing. I started at 10 FPS and obtained 20.45–20.75 F / min farming rates at 4 out of 5 runs but then managed to max out (21.64 F / min) even at this framecap. For 30 FPS cap, I managed to max out at ~21.64 F / min in 3 out of 5 runs, while one run still ended up with 20.75 F / min. Interestingly, it was without the framecap (80–90 FPS on average during the tests according Steam frame counter–I refrained from using the in-game framecounter which refreshes too often) when the values had the smallest possible deviation (20.86–21.06 F / min). Unfortunately, my CPU is too weak to afford a constant 60 FPS lock via RivaTuner without VSync (VSync will result in 30–60 FPS steps which is unacceptable), so i could not check out the most common framecap.
  In short, I could not find an answer, but lower FPS tended to end up with lower farming rates on average, while high FPS showed more consistency without hitting max or min farming rates in process.

In Definitve Edition, fish traps can pretend to be a costly but efficient substitude for the farms because of two changes since HD: firstly, fishing ships now automatically (not manually, which is micro-intensive!) fish from any corner of the fish trap (used to be only left side in HD) and secondly, an auto rebuild option was introduced. These two factors finally made fish traps a worthy alternative for farming not only for the Malay and Japanese.

I decided to assess the rates of fishing with fish traps using the same methodology as in case of farming: 1 dock, 8 fishing ships and fish traps, 10 minutes. The fish trap layout is presented on a screenshot below.


I used this layout (1 tile inbetween both fish traps and docks) based on several pro matches availabe on YouTube; I believe this pattern is logical as it allows to place the fishing ships effectively (no travel time and bumping, easy to place with a single click) while not blocking off the docks and wasting several seconds on microing.

At first I thought several runs would be required to obtain a representative average fishing rate; however, after several runs it was obvious there is almost no descrepancies (±1–2 food collected) in the results (even with various frame caps), probably due to travel time excluded from the equation; so just two runs were used for instancies other than a generic civ with no upgrades.

It is possible, however, to stack fishing ships on top of fish traps (you can manually force them to collect from the fish traps built right next to a dock sitting right on top of them, effectively eliminating the need for ships to travel; see the screenshot below).


I used 'Task Object' triggers to simulate these round of tests (tasking a fishing ship to move to a dock, then to work a fish trap, then, after 10 minutes, to drop off food). After several runs, it was clear that the difference between this micro-intensive method and a method shown earlier is almost non-existent (it was +2 / +4 total food without and with Gillnets, respectively) and could be caused just by automating the test setup.

The results are presented in two tables below. In case of Japanese, FA, CA and IA stand for Feudal, Castle, and Imperial age, respectively.


* — Gillnets is not available in Feudal Age. Column not removed for easier comparison between two tables.

Here we can see that the collection rates for a generic civ with no upgrades are almost identical for farms with Heavy Plow and fish traps (~20.9 vs ~21 F / min), and in case of Dark Age, fish traps are actually 3.5% faster in terms of collection rates. Surely Wheelbarrow and Hand Cart will increase the farming rates (up to ~24.2 F / min for a generic civ and most other civs), but atfer Gillnets, fish traps will prove to be a superior food source income averaging at ~26.3 F /min, which is 8.5% higher than average Post-Imp farming rate. As such, fish traps after Gillnets are only outmached by the Slavs' farmers (~26.6 F / min); but do not forget that the last ones' bonus diminishes the further the farm is from a drop-off point. And Japanese fish traps in Post-Imp will be one of the fastest realistic food income source with 31.5 F / min from a single fishing ship (if we agree that there would not be any practical instancies of Mongol hunt / Indian villagers fishing and possibly no close deep fish in Post-Imp battle).

To compare with, fishing with fish traps was re-tested in HD (only a single run was used). As by default fishing ships use the left side of the fish trap to collect from in HD, the resulting fishing rate should surely be lower than in DE, as several boats will spend some time travelling to a dock (especially the top-right one which spent almost 4 in-game seconds for every drop-off).

That is why resulting fishing rate was only 19 F / min without Gillnets compared with 20.5 F / min rate given in Infosheet (look for 'See also' section, soucre #4). Such a pronounced difference possibly originates from a fact that in Infosheet's case, a trick of stacking fishing ships on top of fish traps was used (the pattern is shown on a screenshot below).


Using this case (which made me spend some time to micro the fishing ships properly), I was able to reach 20.3 F / min fishing rate. A noticeable increase, but with too much babysitting required to make it viable, especially due to the fact that fish traps cannot be rebuilt right underneath the fishing ships.

With Gillnets in HD, fishing rate increased to 23.1 / 25.2 F / min (without and with a micro-intensive approach, respectivelly).

Compared with a more realistic non micro-intensive scenario in HD, collection rate from the fish traps in DE increased by 10.8% (even more in case of Gillnets–13.9%). In case of micro-intensive scenario, the increase is more modest: 3.5% / 4.3% (Dark Age / Gillnets); though in this case a real improvement is not in collection rates but in offloading the player from micro-intensive job of managing the fishing ships and rebuilding the fish traps.

• There is no need in stacking fishing ships onto fish traps: the collection rates are identical with one-tile-away scenario;
  
• Compared with HD, fish traps are way more (>10%) effective in DE; compared with micro-intensive approach in HD (fishing on top of the fishtraps), the increase in collection rates is modest (~4%) but this approach was already impractical for any civ but Malay in HD due to huge time investments and a requirment to rebuild the fish traps manually;
  
• Compared with farming rates in DE, collection rates for fish traps are on par with farming rates with Heavy Plow and no Wheelbarrow (~21 F / min) and are 3.5% faster than Dark Age generic civ farming rate;
  
• After Gillnets, however, fish traps are superior to farms worked by villagers with Hand Cart: you will average on ~26.3 F / min from fishing ships and ~24.2 F / min from farmers close to their drop-off building (the difference is ~8.5%); however, Slavs' farmers are still slightly better averaging at 26.6 F / min (still, only in case of farms that are close to a drop-off building);
  
• Post-Imp Japanise fish traps are now probably the best realistic Post-Imp food income source allowing to collect 31.5 F / min with a single fishing ship; practically, it may only be outmatched by Japanise deep fish collection given it is relatively close to the dock (which is highly unlikely in a Post-Imp scenario).

Please note that these highlights and the mentioned figures are based upon the ideal setup used in the test. In real game scenario, other factors will increase or reduce the effectiveness of the bonuses mentioned below.

• In DE, generic farming rate with no upgrades or bonuses is increased by 4.6% compared with HD and is ~20.3 F / min now (used to be ~19.4 F / min in HD).
  
• In comparison with HD, Wheelbarrow and Hand Cart impact on farming rates is slightly lower: in case of a generic civ, Wheelbarrow (no Heavy Plow) effect is lowered from ~13% to ~12%, Hand Cart (with Heavy Plow) — ~7.5% to ~5.5% (for a generic civ). In several specific cases, the reduction is slightly more visible, and only the Aztecs benefit slightly more from Hand Cart in DE. With Wheelbarrow, generic farming rate is ~22.8 / 23.2 F / min (here and onwards: no Heavy Plow / with Heavy Plow).
  
• With Hand Cart, all the civs except the Slavs have farming rate of ~24 F / min which is nearly identical to HD farming rates. This is caused by 0.4 F / sec (24 F / min) threshold applied via Farmrate Cap mechanic (however, in some cases it is slightly surpassed).
  
• Heavy Plow effect is most noticeable for Dark Age farming rates (~+3% for a generic civ, up to plus ~5% for the Slavs and down to just ~1% for the Aztecs). Its effect is mild for Wheelbarrow (~+2% for most civs except for the Aztecs) and is negligible after Hand Cart for every civ except for the Mayans (the last civ still gets a +2% buff).
  
• Rule of thumb: the better the civ’s collection rate is in Dark Age, the less effective Hand Cart is: with Hand Cart, Aztecs and Khmer receive a mild 1–2% farming boost, while Mayans gain an 11.5% increase.
  
• Compared with Spirit of the Law's video on this topic, farming rates obtained for this guide are mostly in a good agreement (after I managed to realize Heavy Plow increases carry capacity by +1, oh well…): mostly the difference is ±1% exscept the generic civ Dark Age farming where the difference is +3% (the figure given in this guide is 3% higher).
  
• Position of the farm, at least near a TC, still matters: farm in a top right corner has a ~7% higher collection rate than a farm on the opposite side. For other farm placements, farming rates are not that different from each other.
  
• Game speed does matter in any title: according to a thorough farming rate assessments for HD and Userpatch, HD ×1.5 speed farming rates are 1.7% higher than HD ×2.0 speed, and Userpatch farming rates are 4.5% higher (thus, they are almost identical to DE ×1.7 speed farming rates mentioned in this guide). According to my own findings, in DE, ×2.0 and ×1.5 speed farming rates are both slower (1.5% and 0.5%, respectively) than ×1.7 farming rates.
  However, I could not get a reasonable proof (or disproof) that the framerate in DE affects the farming rates as well. All I found was that with lower FPS farming rates tend to be lower on average (most minimal results originate from lower FPS tests), and with unrestrained framerate, farming rates tend to be more averaged.

• The Aztecs are now superior Dark Age farmers (they are on par with a generic civ with both Wheelbarrow and Heavy Plow); however, Wheelbarrow has even smaller effect on their collection rate now (farming rates increased only by ~4%), and with Hand Cart, the Aztecs are only merely faster than a generic civ (the same was true in HD as well). Long story short, as the Aztecs, feel free to delay Wheelbarrow and Hand Cart unless villager speed is of importance;
  
• Berbers' speed bonus is as effective in Dark Age as it used to be (+3% farming rate compared with a generic civ), but with Hand Cart researched, they almost have no farming bonus at all;
  
• Despite a lightly nerfed workrate in Update 39606, Khmer farmers are still confidently outperforming generic colleagues in Dark Age and after Wheelbarrow by ~4%.
  [**] Nerf mostly impacted Dark Age farming lowering it almost by 3%. Wheelbarrow farming rates were nerfed by ~1%. Dark Age farming with Heavy Plow seems to have retained the same effectiveness.
  
• Mayans' farming rate reduction still exists in Dark Age and with Wheelbarrow but is slightly (~1%) less severe than it used to be in HD (it's still roughly –4% in Dark Age and –5% with Wheelbarrow in comparison with a generic civ). (Coded farming rate values were also increased by 1%.) With Hand Cart, however, their collection rate becomes nearly (–0.5%) generic, as this tech increases their farming rate by 11.5%, and with Heavy Plow researched, they lose any penalty to their farming (and even get a slight edge on the other civs except the Slavs).
  Tl;dr: as Mayans, your farming is behind by ~5% unless you have researched Hand Cart.
  
• Slavs' actual farming bonus observed in the test is 10% / 12.5% / 10% (no upgrades / Wheelbarrow / Hand Cart + Heavy Plow) compared with ~ 15.5% / 18% / 15% in HD. Wheelbarrow is still of a high priority as it boosts farming by 15%, and Heavy Plow has a noticeable (+5%) impact if Wheelbarrow is not researched. With overall farmrate increase in DE, Dark Age Slavs farmers are actually on par with their HD counterparts, but with Hand Cart researched, they have ~3.5% worse farming rate (26.6 F / min vs 27.6 F/ min) as they successfully hit their modified Farmrate Cap (it was reduced from ×1.15 to ×1.1 in DE).
  
• Initially I forgot to add this, but surely the Vikings should be mentioned here as well as their free Wheelbarrow and Hand Cart bonus is slightly less (roughly –2%) useful in DE. The Franks instantly receive ~3% / 1.5% (Dark Age / Wheelbarrow) farming rate boost upon reaching Castle Age thanks to free Heavy Plow.

• Compared with farming rates, collection rates for fish traps are on par with farming with Heavy Plow (both average at ~21 F / min) and are ~3.5% faster compared with Dark Age farming. Do not forget, however, that fish traps are more expensive than farms and are less efficient than deep fish (unless the travel distance for the later is too long);
  
• After Gillnets, food collection rate from fish traps is superior to farming rates with Hand Cart (in case of a generic civ, the difference is ~8.5%, or ~26.25 F / min for fish traps and ~24–24.2 F / min for farms close to their drop-off buildings); however, Slavs' farmers are still slightly better averaging at ~26.6 F / min (Gillnets fish traps are actually just slightly better than Slavs' farmers with Wheelbarrow and Heavy Plow);
  
• With Gillnets, Japanese fish traps built close to a dock will provide 31.5 F / min income in Imperial Age.

• Average (generic) farming rate is ~20.3 / 20.9 F / min (no Heavy Plow / Heavy Plow) which is 4.5% higher than it used to be in HD;
  
• With Wheelbarrow, generic farming rate is ~22.8 / 23.2 F / min; its impact on farming rate is lower than it used to be in HD (+12% vs +13%, respectively, for a generic civ);
  
• With Hand Cart, all civs but the Slavs have farming rate of ~24–24.2 F / min which is nearly identical to HD levels;
  
• Heavy Plow +1 carry capacity for the farmers is a ~+3% buff for Dark Age farmers and ~+2% after Wheelbarrow. However, it has nearly no impact in Post-Imp;
  
• The Aztecs are now superior Dark Age farmers (23.1 F / min with no upgrades), but they are the civ that is least affected by upgrades: their farming rate is increased only by 6% after all of them;
  
• The Mayans are still the worst farmers (~–4% lower rates compared to a generic civ), but after Hand Cart, they are now on par with all the civs but the Slavs;
  
• Compared to HD ~+15% farming rates, the Slavs now see only +10–12% increase to their farming rates. They are the only civ to have ~26.6 F / min farming rate after both Wheelbarrow and Hand Cart (in HD, they used to have 27.6 F / min);
  
• After Update 39606, Khmer have 4% better farming rates in Dark Age and after Wheelbarrow compared with a generic civ; however, they have generic rates in Post-Imp. After Hand Cart, it's almost safe to multiply the number of villagers you have on food by 0.4 to obtain your food per second income (after Wheelbarrow, the result will overestimate your income by 1%);
  
• As game speed affects farm rates, in Deathmatch (×2.0 game speed) the rates will be ~1.5% lower than on a normal speed (×1.7 game speed). Note that the values in this guide are given for ×2.0 game speed (as it's faster to get and can be compared with HD ×2.0 farm rates);
  
• Without upgrades, generic civs' fish traps are almost identical to Dark Age farms with Heavy Plow in terms of collection rate (~21 F / min) and are 3.5% faster than generic civ Dark Age farming. After Gillnets, fish traps are 8.5% more effective than farming with Hand Cart (~26.2 F / min vs ~ 24–24.2 F / min, respectively) but are still inferior to Slavs' farmers (~26.6 F / min). Japanese end up with 31.5 F / min.

Here is the list for other posts, guides and videos where farming rates for HD, Userpatch and DE versions of AoE II are assessed:

1. https://www.youtube.com/watch?v=GRutI6IMjkY
   
2. https://www.youtube.com/watch?v=0P-KbjPWFYY
   
3. https://www.youtube.com/watch?v=2r6R-atjlDs
   
4. https://psteamcommunity.yuanyoumao.com/sharedfiles/filedetails/?id=548698803
   
5. Farming rates for HD and Userpatch versions and different game speeds (reddit)
   
6. Farming Cap (reddit)

1. The first video of Spirit of the Law (SotL) is rather old (uploaded in late 2015) and refers to HD farming rates. Its estimates are ~6% lower than those of other sources, presumably due to high game speed and 5 minute samples used. Neither Hand Cart nor Heavy Plow farming rates are mentioned; Mayans used to have lower farming rates back then, and Slavs hadn't received a buff to their Farmrate Cap. I used the same farm placement as shown in this video.
   
2. The second SotL's video is quite fresh and is regarding the change of the farming rates in DE. I used it as a point of reference in my tables and found that the estimates are quite close to each other. No Heavy Plow or Khmer buff are mentioned (the second one came out later); the test setup is not specified.
   
3. The third video of SotL is a comment on a Khmer buff. No specific farming rates are given, only percentage difference with the other civs.
   
4. Mentioned Steam guide collects most of the unit stats, civ bonuses, available techs, and collection rates, including farming rates, from HD version of the game. The last ones are updated according to the latest patches (Mayans' and Slavs' buff). The test setup is not described and the farming rates are higher than observed during testing for this guide (they are mostly lower than those mentioned in the next source). Heavy Plow not included.
   
5. The most scrupulous testing was conducted in this mentioned reddit post: 13 (!) ten minute runs for every civ (including Mayans both w/ and w/o expansions) and upgrade in both HD and Userpatch version of the game and for 3 different game speeds. In the given reddit thread, farming and Farmrate Cap mechanics are explained, as well as the impact of game speed in different (HD and Userpatch) versions of the game: ×1.5 game speed in HD farming rates are 1.7% higher compared with ×2.0 game speed; Userpatch version of the game has 4.5% higher farming rates compared with HD ×2.0 game speed.
   In this source, however, only a single most optimal farm (presumably at top-right corner) was used, which possibly resulted in higher-than-average rates mentioned in the source. The other downside is that it is not updated to take Slavs' buffed Farmrate Cap into account.
   
6. The last reddit post refers to Farmrate Cap mechanic (originally it was used to clarify the reason of Slavs' 'bugged' farming rates).