10:27, I literally "oofed" out loud. To those who are arguing that it's not worth the time to add a 4th villager, you might just be missing the point. Is it worth the time to increase efficiency at all? Probably not. Is it worth the time to build a portal-based iron farm? Probably not. Is it even worth the time to build an iron farm with more than a handful of villagers? Honestly, probably not. BUT (and I stress "but") in Minecraft, worth is relative. If you *want* to build a 100% efficiency iron farm, then it's worth it. If you *want* to build a 1:1 replica of the overworld in the void, the it's worth it. Worth (in Minecraft) is far more about what makes you happy. With that said, I believe the main takeaway here is some insight into the engineering process: chiefly, *Never* assume something is right just because it's always been done that way, and *always* be willing to reevaluate your design. Excellent video, Nico!

Joked about how the last video felt like an academic journal, surprised to see we already got it peer-reviewed lol

If for the question why iron farm guides with more than 3 villagers aren't really popular, my take is we are going as far as collecting more than three times the amount of villagers for one iron farm cluster, to increase 5-10% efficiency towards 100%. Isn't it more beneficial to use those extra villagers to create more clusters instead? In my experience the villagers are the most time consuming resource to build this farm. Even with multiple breeders, you still have to wait a while before you can use adult villagers.

6:00 Something I always wondered about is that most people don't hook up their farms storages to their chunkloaders. You can just use a comparator to detect when the last chest is full and then wire that to your on/off switch of the farm itself (if it has one) and the on/off switch of the chunkloader, automatically stopping and unloading the farm. If people did this, chunkloaders were much less of an issue for serveradmins. I think the coolest part about this is that you can also use it to automatically restart the farm. As soon as you empty the storage, your comparator detects that and restarts the chunkloader and farm.

Remember to put twisted vines in the villager cells to minimize entity cramming lag

Couple of things I'll point out, while yes making small scale iron farms more efficient per cell it kind of misses the point of practicality. Iron farms are (mostly) designed with a practical use in mind where things like ease of building is a big concern. Getting one extra or even 7 extra villagers just to push the rates per cell to its limit while fun as an idea will never be used as it is much more effective to build multiple less efficient farms with the same lag impact while getting more rates. I'm sure you know this already but I don't think writing the comment is a bad idea. Thanks for the shoutout!

Breeding more villagers and moving them into the farm honestly doesn't take THAT long, adding a 4th villager (or more) is most likely going to be overkill for most people, especially those playing in singleplayer, but on servers where efficiency is key and they build huge mechanisms it's definitely worth it to consider adding more villagers to reduce AFK time to get the required large quantities of iron. Great content.

Daylight sensor changes on the same tick if the signal is changing from 0 to 1, which we can use for world tick syncing to sync spawning. Which, I'm using in my iron farm (DFRIF soon tm) that cylces villager groups of 3 down a tube, through nether after spawning a golem, let them sleep in the nether, and back to the overworld. It uses the portals so that the villagers lose authorities for their bed and more villagers can sleep, while having a small amount of beds placed. After I watched this video, I was surpirsed that I haven't even thought about using villager groups of 4. Well, for DFRIF, it may be better to aim for lag friendliness. However, it is fascinating to see the efficiency when used groups of 4 instead of 3. How could I not think of tha- Anyways, as far as I know, after 1.16 update the lag from villager panicking significantly reduced. But there are still a considerable amount of lag in the game, so in my design it only lets the villagers to see the scare source (pillager) right before when an attempt happens. Also, one cycle is 100gt, making the farm itself a chunkloader. So villagers constantly entering and leaving nether wont affect that much mspt. Also, I couldn't find a better way to push the villagers out of the portals, so I used Cow-Boats. Any simmilar sized animal would do it, but Cow-Boats push villagers out of the portal using its hitbox, the boats would not let the villagers to go back in again resetting their cooldown for portal traveling (300gt). 3 villager groups in overworld, 3 in the nether. The farm just looks like a tube. Sadly, I don't think I can protect our computers from villager and Cow-Boats hitbox touching each other.. just laggg Tho the estimated rate for this farm is 17280 iron/h and the total size (over and nether) is still smaller that VUAIF so I consider taking the title of Densest.. Just joking... AnyAnyways this video was very interesting to watch. Keep this up I hope you revive the iron farm field :)

These videos are so refreshing. I have watched so many iron farm videos and just couldn't find one that fit my situation. Built one and couldn't figure out why it didn't work. The video only gave step by step instructions on how to build but didn't explain anything at all. I kept saying I wish I could find a video that explains the mechanics so I can build my own farm. Than your first video popped up in my recommendations. Thank you. Well done.

Great explanation! I really appreciate the scientific approach to understanding and applying the game mechanics. Plus bonus points for following up when you realized that you made a mistake.

You know, I'm glad I watched both these videos. I initially thought it was silly, cause I thought it was just going to be a tutorial for something no one would need. But the way you did these videos and the goal you had is really interesting. I also learned I could pretty easily modify my farm to make it much better. I'm still using an old gossip-based farm. I think it's basically still Amatulic's design? Just with a drop to get the golems out of the detection box. Not sure if I'll end up doing it, cause I already have more iron than I know what to do with, but who knows.

As an student studying engineering, I was honestly excited when I also realized how the 4th villager would effect the probability calculation. I was genuinely thinking “omg he gonna multiply the probability - yes let’s see that significantly smaller percentage 😤”. Btw I really like how you present this and the previous video because it’s more than just a Minecraft video. You present it like a experimental design. As someone who doesn’t play Minecraft anymore, I still find entertainment in the way you communicated it so you earned a sub from me! Keep up the good work!

This may have been touched on in earlier videos, but there are multiple ways to measure efficiency. Space efficiency, lag efficiency, & supplies efficiency. These 3 need to be balanced based on what kind of farm is being made. Supplies efficiency is important for early game. Lag and space efficency is the other optimization point. Space efficiency can be measured in full 3d space or in 2d footprint. How many items do you get per increase in mspt? How many items do you get per block (footprint or 3d)? How does lag efficiency change with different physical hardware and java versions and launch parameters?

Hey Nico, thank you for making these videos. They’re truly a great dive into the mechanics of this game. I don’t know if you’d be willing but I’d love to see similar videos for different types of farms. That being said the probabilities here 9:52 are a little bit off. The probabilities you’re talking about are the probability that anything besides “all four villagers don’t detect the golem” happens and the probability that anything besides “three villagers don’t detect the golem” happens. These events have opportunity for overlap and therefore aren’t independent. I’ve done the calculations myself, if anyone cares to see them just ask. I’m happy to say, assuming the events of individual villagers detecting a golem are independent, that the actual probability here is closer to .015% this being the probability that 0 or 1 villagers detect a golem. The same statistic for 2 ticks comes out to .059% Again, wonderful videos and I’m happy to see you addressing any questions to your assumptions. Keep up the good work!

Really nice video. What an insane guy this slly, I read the first part of his answer and it hit me immediately, but coming up with this... Also I really enjoy the way you make your vids, keep it up. Can't wait to add this new updated version to my server.

I love your farm it is really good, keep it up!

Thank you for making these videos, I really appreciate your methodology

14:40 saving this for myself, seems like a simple enough yet very efficient farm.

Such a simple idea for boosting rates. Your iron farm vids continue to push the meta forward

Great video, good on you for testing and not assuming.

Great video, the best iron farm just became even better!!

I just found your channel, an I'm excited to see your growth.

Awesome video! I'm so glad redstoners make follow up videos like this, it shows that learning is more important than being right :_ That being said, I wanted to ask about the math at 9:00, because I'm getting different numbers for the 4-villager scenario. You multiplied (probability at least 1/4 vills detect)*(probability at least 1/3 vills detect), which I suspect is double-counting some scenarios but I can't put my finger on how. I calculated it this way: the probability of not spawning a golem is the sum of the probabilities of 2, 3, and 4 villagers detecting the golem. When exactly 2 villagers detect the golem, there are (4 choose 2) = 6 ways of selecting which villagers detected the golem. Similarly, there are 4 ways exactly three villagers detect the golem and there is one way all villagers detect the golem. For each villager that detects the golem I multiply by (1/200) and for each villager that doesn't detect the golem I multiply by (199/200). This produces 6 * (199/200)^2 (1/200)^2 + 4 * (199/200)^1 (1/200)^3 + 6 * (1/200)^4 = 0.000149001875 = 0.015%, which is about half of the 0.03% you calculated. If you find a mistake in my calculations please let me know!

Odd, I have been using a modified 3 villager generator by adding in a 4th because I was having problems getting them to sleep. When I added in a slight change, at night, one can sleep while the others were panicked and could not fully sleep. I really had not checked to see if the iron rates had improved, I only make them so I have enough iron to keep doing things that consume my iron. I usually have more than enough. I do like to see what others come up with for these super dropper farms that make insane levels of loot.

LOL I'm just imagining if this information was in the wiki. good video :)

Great video, I think would be great have a video that recopiles this methods, and use the next metric, iron generated/time needed to construct the farm. In some way, that would maximize the iron we get in the time.

Very amazing community and very amazing video, good job.

I really love the detail in explaining how and why rather than "Just do it like this."

Current bee farms are debatably also underdeveloped, it would (hypothetically) be better (for hive rates, not hopper use or simplicity) to have 4 hives (or nests?) around a single flowering azalea leaf block, and have 3 extra bees per such unit (15 bees per 4 hives), this prevents bees from being pushed off their flowers, from choosing a flower further away, and most of all ensures that hives are always making honey (instead of idling) in addition to a few other benefits/differences. Although there may be a few drawbacks admittedly, such as bees needing to move into whatever nest (err, hive) is empty at the end of pollen gathering instead of the nest that they last came out of. Construction is also complicated, mostly because of the final 3 bees, timing the bee shifts might also be an issue.

That's a great follow up video right there, I tell you what.

the reason that three villager farms became the standard after 1.14 is because of cat and their spawning mechanics, for a good while as developers were working out bugs and technical mine crafters were figuring out what was need to create efficient iron farm, trading halls, and village breeders that had to deal with cat spawning. it was hover quickly found that you needed at least 4 villagers to enable cat spawning and only three to spawn an iron golem, so three villagers became the norm as far as I can figure out as of 1.18+, you need to have five villagers with both job sites and beds you create a cat spawning village

now i would love to see a collab with you and Mumbo Jumbo

the thing id like to see is a farm that has the least amount of lag for the most amount of iron produced per hour offset by the difficulty of building the farm and its size, so if adding a 4th villager increases the iron output more per lag metric than other design choices, like nether loading, then that is the design choice made, whichever one wins out.

got here trying to get some iron for my hardcore server, but now i’ve grown incredibly invested in the iron farm lore.

well presented , great vid . thx

Monstrous work was put into this video! It is great. But I have one question: will there be worldmap with tripwire scaring system?

Not boring and very very very very informative.

Bro straight up said, “rookie numbers”

Just built the 99% efficiency farm and it works really great, although it was kinda hard to build

Very nice. Though I think the reason you only see 3's is because most of those designs, like the gnembon rollercoaster one i'm fond of, have each group separated enough that they only see the golem they spawn, so there won't be that issue you are dealing with.

I built the design on 12:20 on MC 1.20.1, but for whatever reason, sometimes villagers would fail to sleep if I used a bed too early at night in the vicinity of the farm(which was pretty often). To solve this I added an observer observing an inverted daylight sensor, connected to the triphooks' redstone, so it would hide the zombie exactly at sunset, giving the villagers a chance for a quick nap. A repeater was also needed since the observer signal was too short to trigger the circuit. I know that daylight sensors are affected by weather too, but so far it has solved my problem!

This channel will definitely blow up if he keeps this up.

Great vid again👌🏽

Incredible video!

This is such a great discussion about essentially one of the biggest conflicts in creating technology. Pushing the tech to its absolute limit, fine tuning every last detail for optimization vs. "Yeah, but what for?" These were two fun videos showing the mechanics behind iron farming, figuring out what is theoretically possible and how to get as close as possible to it. Some people are gonna have a blast using these designs just to show off. For 99% of other players just chilling in a normal survival world? Eh, throw three villagers in a hole, more than good enough. And that's what sells in the end. Great content

In my village I build an accidental iron golem farm. There is just plenty of them running around and I'm now going to investigate why that happens and how it can be abused. It's on the hill part of my coast village with only two houses(one empty) and a farm (that were 2 farmers, which I regrettably reduced to 1). There is a fence keeping mobs out, but visible. My guess is they see mobs and go to sleep , wishing for golems. Or the farmers reported something to the village down under which has fewer iron golems around.

Wow so interesting how just a few minor tweakes can change the percentage by miles

My ultimate farm is a village with 20 villagers, and a fence a few blocks above a + shaped berry bush. I call it the Iron Gallows and use a lead to hang the golems from the fence at just the right height that they float in the bush

Great analysis btw. Here are my thought on the most optimal farm: Firstly i have to take note that your so called farm efficiency refers to iron production rate per villager cluster. And thats an interesting topic, since (as I will later talk about, this contributes to maximizing space efficiency), but there is other factors to consider: In my opinion theres no reason for more than 3 villagers in real scenario application since there is most often no need for high efficiency. Heres why I think so: Consider your farm a black box, then only three things matter: -Work to build (cost to build is probably neglectable) -Space Efficiency (Iron production rate / space) -Lag Efficiency (Iron production rate / produced lag) Work to build: considering a fixed villager count (that means constant work), lets take the convenient example of 12 villagers, you can build 4 three-villager clusters or 3 four-villager clusters. Relative to one max farm, 4 three villager clusters produce 4 * threeVillagerEfficiency, which is greater than 3 * fourVillagerEfficiency. This is true because the efficiency per villager in one cluster is flattening as shown in your graph at 13:20 Therefore I follow up, that production per villager is most efficient for lower villager per cluster ratio, therefore 3 villager clusters are the most efficient to build. Space Efficiency: This has to be calculated explicitly, it mostly favors more villagers per cluster. But since there is quite a lot of space in the world, space efficiency probably unimportant. There could be special cases tho. Lag Efficiency: In case your not using portals, the lag is approximately proportional to villager count. That means that lag efficiency is maximized by using three-villager clusters (as shown earlier). Lag efficiency is probably the most important aspect in here, since if you are willing to build big farms, lag is all you care about. concluding, 3 villager clusters are more efficient when talking about general purposes.

Technical minecrafter here: In Scicraft, we have built recently a 100k/h iron farm. We did consider the possibility of adding a fourth villager to increase the efficiency exactly as you're mentioning here. We also ran several different ways to place the villagers to find the most optimal pattern. We went with a 3 villager setup instead of a 4. The efficiency? 99.97%. Why 3 and not 4? As you can see, the efficiency it's already at almost 100%. There' simply NO need for it. There are different ways to achieve this. Of course we're talking about TMC farms, pushing the limits. This farm runs at 30ms in a 7950x. This means we have 20ms to spare to not start lagging the game. The moment we run 1-2 more farms, we get to lag issues. If we ran it with 4 villagers, the lag would increase to around 35 to 40ms. That's just not ideal. We would not be able to run a single farm with this on. Conclusion: The 4th villager is only viable for small farms. Those that are less than 1k/h.

Even if the sound and particles don't appear, the weather still affects the daylight sensor, because it is a global thing. How it is displayed is biome-specific, sure, but not the redstone functionality.

I think it depends on scale, the small amount of efficiency u get from adding a forth villagers is nothing compared to just making another module.

I built the iron farm in your tutorial in 1.19.2 and it’s getting 4,688 iron per hour on average, personally I won’t be putting in a 4th villager, as I afked once and received 24 double chests of iron, very impressive farm. But I am extremely interested in the smaller farm, the one at the end of the video looks very promising! Keep up the good work and the great videos!

Someone should make a list that rates the farms based on various factors like one segment for no restrictions, chunk loading constant, time to build from scratch, and so forth

good one

This and the 100% eficincy video were amazing

I have been using 4 villager farms for a long time. Actually when it comes to small builds with 1 set 3 worked better. When it comes to larger builds it works far far better to have 4. You did cover why, but also on the server I played through it was important to keep lag down but I was the iron supply for the entire server. My base was the super farm of the server with all the farms that where lag intensive. Because I was on the most they could be small, tiny even, and run huge collection systems and not lag the server out. I had the creeper farm, exp farm, zombie skelly farm, the spider farm, iron farm, prismarine farm, sugar cane and bamboo with bamboo split in half to feed my blast furnaces in the prismarine farm.

I love your humility here, you realize you're not perfect, that there's the chance for improvement and you tested the idea to increase reliability resulting in a better design. You didn't fall to your pride and instead you learned and published it for all to see that you had made a mistake. Bravo, not sure I could do the same in all honesty, but I hope that I could.

10:15 aka survivorship bias :P

Really great series :) What mod did you use to display the areas (e.g. where the villagers can see the golem & which villagers belong to their cluster) of the villagers?

Awesome!!

The man made a full fledged video about the best possible way to make a iron farm.. Knew 100% it was right.. get's challenged in discord and admitted he was not.. That's worth a sub to me because I know the information is well thought out. (I'm 8 months late but still wanted to say this :) )

I know a lot of this video wasn't using clocks and just using tripwires but, for the problem with players sleeping messing with the clocks, assuming there's no difference between what a bed and a daylight sensor consider day and night, you could have a daylight sensor connected to a clock that counts a minecraft day. If the daylight sensor detects day when it should be night, it resets. if it's night when it would be day it is ignored. This allows the timer to ignore weather false nights and recognise when the player sleeps and creates an accelerated reset to day. This would be done with having the sensor only reset the timer when it turns on and ONLY when its night according to the timer. This avoids weather clearing up midday breaking the timer. I think by ensuring the piston is triggered at the same set of numbers after every rest (e.g 2, 4, 6, 8 ticks or something like that. Never 2, 5, 7, 8) you can get it to work. Please let me know if theres something wrong with my idea I'd like to hear feedback :)

Underrated

for a type of farm like this, i think an off switch of some kind would likely be a good idea just in case, since it would probably produce more iron than you likely would ever need depending on how big it is.

Many low-end iron farms (the super simple ones that does not optimise anything at all and are rather common amongst non-technical minecrafters) often use more than 3 villagers. Typically 4 or 5. So they do indeed exist but not in the area you frequent :) Something a bit different: I've had some thoughts on making Iron Farms in Peaceful Gamemode and would like to hear your feedback on this. Most peaceful ironfarms rely on random/gossiping spawning of iron golems and are thus rather inefficient and slow. But it is possible to make villagers panic in peacful and thus make more reliable iron golem spawning. When villagers take any form of damage they will panic, even if the damage is 0, such as from a snowball or egg thrown at them. This of course requires a bit more elaborate setups to accomplish this. For dealing 0 damage to the villagers I can think of 3 methods, snowballs or eggs shot from dispensers or snowballs shot by snowmen. (I don't know if there are other methods of dealing 0 damage than these?) The two first ones will require a source of these items and both have their advantages and disadvantages. The snowballs will require manual farming from a player and loading the snowballs into the farm. The egg method can be automated but has a problem with requireing both a lot of chickens but also the eggs shot at the villagers will sometimes produce a baby chicken which needs to be dealt with somehow to not cause problems. The baby chicken could perhaps be lured away by having an adult chicken nearby it will try and pathfind to and then fall down and away. For the third method, using a snowgolem that shoots the villagers might be better in terms of not requiring an external source of snowballs/eggs since it produces these itself. I do however not know if it's possible to make a snowgolem shoot at a villager in peaceful, also I would imainge it's more difficult to get the timings correct with a snowgolem shooting at villagers rather than using redstone controlled dispensers to do so. So there are definetly advantages and disadvantages to all three methods. One thing I have found from playing around with this is that when hitting the villagers with snowballs/eggs while they sleep, they need to be hit twice in a row to ensure they actually panic correctly and don't just wake up and goes back to sleep without panicking. (Granted, it's been a couple of versions since I've last tested this so it might have changed since then.)

2 am. videos getting better and better, lol.

I love a good iron farm

Great video. I think it would have been better with a reduced introduction at the beginning, however. Nice work though!

Interesting, I’ve always used 5 villagers. No zombie either. Looking at this has given me some ideas

It be like that sometimes

I think we can find a sweet spot where for a certain number of villagers (a range maybe) more villagers per cluster is better than more clusters. For example with n=12, do you want 3 clusters of 4 or 4 clusters of 3 villagers for higher throughput?

Looking at the design at 14:40, there are some interesting calcs you can do. For example, for a 3 pod design, you need 9 villagers for an effective 273% spawn rate (compared to a single pod) if you add a villager to each pod. You use 12 villagers and go up to a spawn rate of 285%. Alternatively for the same number of villagers you could instead increase it to 4 3-man pods, increasing your generation to 364%. Id go as far as saying. With villagers being the primary souce of lag in these farms, that 3 man pods having a higher villager to iron ratio would likely be the main reason 3 man farms are most common.

0:01 reminds me of that Elon video where the dude asks him but only for the boosters right? lol

You’re beginning to make me fall back in love with maths

Instead of water try slime block piston pushers and tripwires A slime block hitting a mob sends them flying so push them at a fence post centered above the hole that falls farther than the vilifies detection range

Bro waited until 7:35 to ask the question, my man is a full fledged KZfaqr now :)

you just made a switch that turns off spawns if you dont need iron anymore and thats awesome

I always just build my iron farm with 60 to 100 villagers put them in a ring around the spawn area and use it as the trading hall typically spawns 4 to 6 a minute

While I do think in your last example of a 4 cluster iron farm, you would get better rates if you take the bonus villager in each cluster to make another cluster. I still prefer the 4 villager version because it’s technically more efficient, and 4 clusters will produce more than enough iron anyways. So this way I will know I made a really good farm, and not a lot of farms where I can easily identify a flaw. It satisfies the perfectionist in me.

makes sense

Most Efficient Per Pod is an interesting concept, though on large scale it is definately the case of iron per lag

Dont know why but near top left at 4:27 looks like there is just a distant golem chilling

10:00 "Question, prod, and push back" I love doing this but people really do find it annoying when you question every assumption and underlying concept. It makes sense to do so, though.

Im not sure if its because im on bedrock edition or not but i didnt have any lucky getting them to spawn in the water with my makeshift farm. I did however make a ring around the water where they spawn, and a bunch of pistons to push them in lol its not at all efficient but the pistons certainly sped things up, think my other issue is i only dropped them down 20 blocks something i think i will have to try adjusting

It seems the efficiency increase starts to taper off before not working, and villager 4 and 5 give the biggest increase. One might try to add 2 more villagers to each module.

So I'm trying to build a simple iron farm that uses slabs and a trap door to scare the villagers and some angles to get them to sleep and it just quit on me completely. The villagers are spooked but for some reason they don't make golems.

i usually get pretty big desert near spawn

Your voice seems similar to ant venom’s and it’s coool :p

2:15 I literally laughed out loud 😅

Atleast you had the balls to say they were right

great video

I like the idea of making a huge perimeter just to put some dumb farm in the middle, like a chicken farm

I think you can compensate for the lag by doing away with redstone dust and using leaves instead

I think the issue with a 4 villager setup is that if you have 3 pods going from 3 to 4 villagers, you'd be better off and result in higher iron rates of instead of going to 3 pods of 4 villagers, just going to 4 pods of 3 villagers. So if your goal is iron rate per lag, adding more pods increases the rate for basically the same lag impact

How badly I'd love to min-max my minecraft efficiency by getting really good with Full-house modpacks mods like all the technical, scientific, and magical mods and applying the deep understanding of redstone, Minecraft spawning, Minecraft mob farms, and general Minecraft architecture together to make a massive super factory world that produces insane quantities of basically infinite everything to obtain effectively creative mode ins survival.

Vidéo très qualitative !!!!

Really need the last farm build instructions

Classic example of "Betriebsblindheit" (German for operational blindness)

Can you make it with a pillager in the middle instead of 2 zombies on the side if so does the crossbow need to be broken or can I add slabs around the sides of him