Factorio: 8 advanced dragon's-teeth designs tested

Hello, I am back again with some more advanced dragon's-teeth tests!! This time with a death-world accurate mixture of biter sizes -- and flame turrets. :)
Each wave of enemies includes:
----------------------------------------------------
378 - Behemoth biters
255 - Big biters
129 - Medium biters
Design notes:
----------------------------------------------------
Wall #1 - Control group. Not very effective.
Wall #2 - Slightly more effective than control, but easily tilable in the X and Y axis.
Wall #3 - Very effective and easily tilable in both the X and Y axis -- creates corners very easily. Biters only make it about mid way through the design in this test with hundreds of biters. Very good.
Wall #4 - Very effective, but only tileable on the X axis (and when you hit a corner it turns to hell). Biters also have a rare chance to path quickly on the diagonal causing problems sometimes.
Wall #5 - By far the most effective design, but has the tendency to incur more damage than most when under large scale attack.
Wall #6 - Decently effective. An attempt by my to relieve some of the stress wall #5 was experiencing. It works somewhat.
Wall #7 - Not super effective. Catches biters decently, but disperses them a little too much for flame turrets to be fully utilized maybe?
Wall #8 - Very effective, but only tileable on the X axis, and corners might get a little complicated. This design plays into the flame turrets very nicely. Probably the most resilient under very very large scale attacks.

[@Trupen]

this is dope comparison :D

[@purpurthedragon]

Hello

[@olegmoki]

Hi

[@droid777]

Tell me why?

[@Mr_Tomato_Man]

love you bro

[@omaba]

Number 4 seems to be the best design in terms of upkeep since the biters didnt even try to damage the walls. (But at this stage you should have construction bots coverage to repair.) And in terms of stopping power number 5 seems to be the best since the biters only reached the 3rd layer walls. Edit: The fact that the biters are not attacking the walls in design 4 also means that construction bots wont die to flamethrower crossfire when attempting to repair the walls.

[@kowatello3793]

I also came up to the fourth desighn independently when experimenting with walls that don't get bitten by biters, but slows them enough for flamethrowers to be effective. But corners are pain.

[@F.M671]

It's a small sample but it seems like "5" makes the biter AI spread somewhat equally

[@Zinic_]

Finally a "dragon's teeth" design I've seen that doesn't include a final, unpassable wall. I've been experimenting myself (on my own world, so not much testing) for a design that inhibit biter movement, involves minimal wall damage, and one that I can freely leave and enter without having to temporarily remove walls. So I think my favorite designs are 2 and 7. Plus it looks like making corners is easy.

[@Hawk7886]

What's wrong with gates? It's pretty uncommon to frequently leave through random points, usually you'll have just a few you'll gravitate toward. You can also have a Spidertron follow you around and just hop in to cross otherwise inconvenient spots.

[@gfan83]

According to test 5 th version show the best. 4 less impressive but also not bad.

[@jeffreyfulkerson9684]

@@Hawk7886 Yall can just use the decostruction planner. Drag over, run through, then shift+drag over to undo the deconstruction

[@Coecoo]

It is still nowhere a proper realistic presentation. 50% of the mobs just hump the unpassable walls above (due to a confused constrained pathfinding) and to the side instead while you also have stuff like corner designs to consider.

[@idiokrat]

@@Hawk7886 If you close the walls, the spitters will attack them and you have to constantly repair them. When open, they only try to attack the weapons. Thus, an open construction needs much less maintenance.

[@user-sg6zh6vr7h]

5th version did the best, mostly thanks to the side protrusions preventing the biters from moving to the next line of wall, 6th was also good but the removal of some "thorns" made it less effective at containing biters

[@joezhou3639]

I guess the 5th reason is the best is that they also force the bugs to go back in order to get further, however their AI may not be able to do that calculation...Due to that 5th's wall may have a higher possibility to be damaged because they can't find the way.

[@nikelsad]

The 6th has the snake-like paths through the defence that are too straightened. Very similar paths to the 7th ones, just more sparse.

[@bugologist2]

Byfar the most effective one was the one that performed the best.

[@nevelis]

phd level 900 iq genius #1 most underrated comment on the internet right here

[@jozefuhasis1338]

This man could take a sip of liquid shit and conclude it's not mint mouthwash.

[@davisdf3064]

Truly, one of the comments of all time

[@ungabungalunga1002]

@@jozefuhasis1338 lmaooo

[@maelglorious]

Byfar the most informative comment about the effectiveness of the one that performed the best.

[@Vastin]

In terms of absolute effectiveness, #5 seemed the best - but I'd probably go with #4 in most cases as it was nearly as effective, but far simpler to arrange and the biters made no effort to damage it, so both placement and upkeep are minimal effort affairs.

[@chainsawplayin]

It's like a pachinko machine but with flamethrowers and lasers

[@krinkrin5982]

Didn't expect the brick pattern to be so effective.

[@Coolguydudeness1234]

Loving the pilotredsun soundtrack

[@benjamincasatimcintosh2918]

I think the best design would be one that forces the largest amount of biters trough the same hole so that the flamers do the maximum damage. There's definitely a way to maximize unit superposition, would like to see a few different desings

[@narancs5]

An important factor can be which of them requires less walls. Thank you for this demonstration. I was using basic designs so far - the cross pattern and a compressed version of the no.1

[@reda29100]

You can't believe how many great ideas this wall stuff gave me! I did design a hole-in-wall design in 0.16 but that was a while ago.

[@hansdampf640]

my favourite was 4 for quite a while now,guess i try 5 from now on. amazing work!

[@dadbear5316]

5 definitely gonna be my favorite, thank you for this

[@prolbononsence6602]

I like samples 4, 5 and 8. 4th looks super simple and very efficient, 5th shows unparalleled stopping power, 8 seems to be able to compress all the pests in tight spots which allows flamethrowers to roast them all in one place.

[@benjytom7371]

I'm very new to the idea of dragon's teeth walls. I saw a complicated design a while back, and while playing a few days ago, I remembered that, and tried to do a very simple design. As it turns out, it is exactly the same as design 4. I noticed that it took a bit more damage than I would care for, and instead of doing the 7 long wall lines, I did 5 long wall lines. I haven't used them long enough to know if they are overly superior, but so far, they have seemed to take less damage. That being said, they aren't being hit with attacks as large as you're simulating. I really found this video interesting, and hope you make more, and I'd love to see which is better, the 5-wall lines or the 7-wall lines. Outside of that, I hope you could make a diagonal dragon's teeth wall comparison. I don't know if many designs exist for diagonal walls, but a simple wall design I made from scratch consists of a 9-wall right angle (so it looks like _| and is made of 9 walls) and then you take that and put another one diagonally from it. Then, you rotate it to flip (upside down and backwards) and put it so it's one wall away from the others, and nested within. It's hard to describe with words, but I hope you understand. In practice, it looks like nested zig-zags, without walls in the very corners, so there's just enough space to walk through, but only some of the corners (the innermost corners of one zig-zag, then the outermost of the next, and repeat) so that you can't just walk through them, but you are funneled into a hole, then walk backwards kind of, then you walk into a hole, and then you are funneled, and so on.

[@harrygenderson6847]

It's worth noting that 2-wide gaps horizontally (but keeping 1-wide gaps between vertical layers) can reduce the chance of biters attacking the walls. However this may require further testing with 6- and 8- wall segments.

[@lordilluminati5836]

#5 seems to be the best

[@egerlachca]

I don't agree. 5 and 6 both took damage, which I think means the biters were lost enough to attack. The ideal is to have them slowed down but not to attack. 2, 3, 4, 7 and 8 did that. 3 & 4 also seemed to do the best at preventing biter advances.

[@knoop7102]

@@egerlachca Since I was only using biters and not spitters, this test isnt entirely accurate to a real scenario. Spitters seem to damage any design no matter what if they really feel like it. So a little damage is expected to matter the design. The upkeep of a few extra stone walls can definitely be worth the protection it provides, but its entirely dependant on your scenario and situation. I feel like every design offers something a little different.

[@irek1394]

@@knoop7102 what is the name of the song used in the video?

[@migueeeelet]

@@egerlachca True. 5 in a way is one of the best since it screws up with the pathfinding as they have to move away in order to go further ahead. But if that screw up ends up making them attack the walls, then it's not as effective as one that simply slows them down

[@samuels1123]

@@knoop7102 The purpose of dragons teeth is to make the biters slow down before attacking, if walls are broken then the design failed or spans too far

[@elliejohnson2786]

Music is dope!

[@katiehesse6578]

It's by pilotredsun I think

[@jayst]

@@katiehesse6578 Warsaw by Pilotredsun. That whole album is amazing

[@pafnutiytheartist]

My problem with dragon teeth designs is that biters will deal minor damage to outer walls and robots will try to repair right away, ignoring the fact that the whole area is still very much on fire. And replacing bots is more expensive than walls.

[@bluedrache]

Have a switch operating on the level in an accumulator that powers the laser. As soon as it drops to 25 percent on the "A" signal, connect power to the main grid to fill the accumulators. Once your accumulators are full, send another signal to something to release the bots for repair.

[@matchley500]

From the concept to the way the unit testing was done ...to how it was presented. I think I enjoy the people that tend to play Factorio...more than I love Factorio

[@followingtheroe1952]

The game really is a Rorschach test to see how people approach design. Like I have noticed through general comparison that I am very heuristic and intuition based, and that I like to build my factory from the outside in, to allow more complex iterations to proceed closer to a central location

[@giin97]

Dude! 5 is so similar to what I came with and use, nice to see some validation 😁

[@themenagerie5622]

I'm a big fan of 5, looks really cool and I'll probably use it for the walls on my ribbon world I've been playing lately

[@abndmt]

Warsaw by Pilotredsun

[@anxstxsia9184]

5 is smart, forces them to move backwards else they cannot get further in

[@benhawk9936]

That's good science.

[@jnzsblzs1]

Not really, without repetition it's not. But it's still useful, and it's a great representation of some ideas, and since the video maker didn't try to make a scientific assessment, it more than accomplishes its goals.

[@wesleysmith5275]

@@jnzsblzs1 Buzzkillington has entered the chat

[@davisdf3064]

*"Reconstructing Science" starts playing*

[@Jaqbasd]

Music used is Pilotredsun - Warsaw

[@mathiusj74]

I feel really silly. Dragons teeth as a wall defense never occurred to me. Thanks !

[@ScouperGaming]

Only the spittars will be split up and come to their range but its nice with his enemies don't attack in one place and are more spread out you can do in the middle flamethrowers and on the outsides machine turrets or energy it safes way more resources

[@nospamallowed4890]

I tried pattern 6. Worked great... until I was attacked by a wave of large spitters that stayed out of range of lasers and flamethrowers while it attacked the outer layer of the wall. Is there a minimum opening size to make the large spitters ignore the wall?

[@Spoodertron]

yoo pilotredsun music epic

[@wyhiobcarlile4879]

you should do this again with spitters.

[@G33KST4R]

the bitters that are going the "wrong way" just know what's going on and they don't want to be burned alive 😢

[@Wilson_Kelly]

Nice music, love pilotredsun!

[@Oliepolie]

take your time :3

[@knoop7102]

Music: Pilotredsun - Warsaw kzbin.info/www/bejne/q2LQhpx3n62td6s

[@jayst]

Yo that pilotredsun

[@atomspalter2090]

thanks

[@edomeindertsma6669]

The flamethrowers were not in the same positions for each section, so sometimes an extra flamethrower may help. You should also run more tests to decrease randomness, as some biters ran away pretty soon.

[@followingtheroe1952]

I agree, but there is a caveat that the whole point of these designs is meant to exploit pathfinding. Essentially each design is introducing "noise" to the biters behavior which is exploited by the turrets.

[@patrickgoode4230]

I agree, sample size was low and margin of error seemed really high. It was insightful, but not conclusive. A cost comparison of walls per unit length also would have been nice.

[@dentingwind3148]

Take your time :3

[@MrHics]

Imagine the watch time % on this video. I've watched it through 3 times straight and added to my blueprints bookmarks folder for many later references

[@awetails]

I know that I am asking for a lot of work... but did you do repeat testing to make sure these situations were on a fluke?

[@kalacaptain4818]

good taste in music

[@TheHuntermj]

What if you used big rectangular blocks of wall to form large corridors with a flame turret at the end so the rest of the hoard would have to pass through the residual flames.

[@keonisan]

I prefer to make a setup where there's no more than one opening in the wall segment so that all offensive power is directed to one spot.

[@ragequitgaming8937]

@Knoop... Great video! May I ask, if there is a solid wall on the back side of these dragons teeth, does that make the biters start to attack the dragons teeth more? That is an experiment I would love to see. Oh... Is it possible you double clicked on a test or two and put more biters down than intended? #7 looked that way, but maybe you did that on each one?

[@josephshreeves8192]

I don't think a solid wall behind makes much of a difference. What does is when there isn't enough space for a bug to walk through. They start attacking walls then. Mostly doesn't happen until hordes of large or groups of behemoth bugs

[@Vastin]

@@josephshreeves8192 A solid wall in the back might actually cause them to attack things further forward. Basically it makes a pathfinding check to the thing it WANTS to attack (guns & buildings), and if the path gets too long, it changes tactics and just straight-lines toward the target, attacking anything in that line. Because the solid wall in the back would block all reasonable paths, the AI might well just immediately go to straight line mode before it even enters the dragon's teeth.

[@lackofsadness5146]

i'll take my time :3

[@brozzy9423]

At 1:55 if you look at the burning mass on the left it’s almost in the shape of the earth… what are these biters telling us?

[@KingThrillgore]

Five looks like the best design since they only got about 3 meters in.

[@Alexander_Sannikov]

i want to try a design made of H blocks interlocked in a staggered fashion

[@giin97]

That's like what I do, quite similar to design 5, but without the middle spikes.

[@slowpnir]

What about "Peano curve" one? Will they pathfind through?

[@tristen9736]

The only issues I have with dragons teeth is the random single spitter that is just our of range of everything and will constantly attack 1 or 2 walls

[@glovely4119]

song?

[@namshimaru]

yoooooo the pilotredsun

[@PluvioZA]

Firstly, love this video - Really cool comparison. But even more importantly, what is this song? It's a jam and I wanna listen to it! :D

[@PluvioZA]

Found it! Artist: PilotRedSun Album: Achievement Song: Warsaw

[@youforgotthelinkinthedescr6798]

They should have linked it in the description, thank

[@mo3g608]

yay,pilotredsun

[@hyperfox0934]

Honestly? I'd go for five lines of the fourth option and finish the rest with five. That way it cuts down on the amount of bots lost to fire damage while still offering the maximum protection. Plus I think it'd look nice :)

[@kerbodynamicx472]

After this testing, let’s invite everyone to a meal of deep-fried biters.

[@Mr_Tomato_Man]

could you technically replace the walls with pipes and they wont get destroyed?

[@MrDoggo23]

Huh, neat!

[@haikuheroism6495]

tkae your time :3

[@internallyinteral]

What is the song

[@user-kb5qr2ns7p]

None of the options are suitable for real gameplay. The first priority is to ensure that Behemoth and Big Spitters cannot destroy the wall with impunity while staying out of range of turrets. Then, an optimal balance needs to be found between the time and materials required for construction and minimizing bot casualties during wall repairs during an attack. All of the proposed options fail to meet even the first criterion.

[@knoop7102]

It's true I tried this design in game and the biters just attack the wall

[@Euler271]

Why are lasers used alongside flamethrowers?

[@BORN753]

What is the song?

[@slenderpaul]

kzbin.info/www/bejne/lWGzgpyZmbmZhZo

[@movax20h]

Wait. Flamethrowers do not destroy stone walls? My mind is blown. I was playing defense game wrong for so many years. Still I just use solid wall with enormous amount of lasers. Good enough for late game, but in the mid game, flamethrowers are op.

[@charlescourtwright2229]

flame throwers will constantly hold their own, due to the dot burn

[@matsv201]

The issue is that flamethrower will kill construction bots. And if you have a close wall (that most poeple do) spitters will attack the wall from range, and damage them. When the construction boot come around to repair the wall. If the flaimthrower try to kill biters at the wall, it will also kill the boot.

[@PhuongNguyen-zb2en]

What about awhole matrix? :D It will be really fun seeing them strugling trying to find the way

[@totallynotminx4682]

Unfortunately, if the biters can't find the way fast enough they'll just blow up the wall to get through faster; it's why sometimes biters will attack stuff they don't normally, like pipes and poles.

[@xGOKOPx]

The point of a wall like this is to let the biters pass so that the wall doesn't get damaged, but slow them down greatly so that your turrets can deal with them. If you block the passage completely then they will just attack the wall

[@alexanderlove53]

The problem of this spaced walls is that your rate of construction workers loss increasing abnormally when they try to repair it.

[@raizors1331]

That shouldn't be a problem unless you use bad design that make the path inaccessible for the biter (spacing is too small that huge/behemoth can't fit through, leading to them just chew through the wall) or adding a solid line of wall behind the teeth which trigger the splitter into attacking walls. In general, bots won't die any more than your regular double wall (wall damage rate is 100%), and single wall, of course, do nothing against splitter so there's no point in comparing them since wall is way cheaper than turrets anyway.

[@alexanderlove53]

@@raizors1331 its more about bots dying from friendly fire

[@FurryEskimo]

So what’s the advantage here, co pared to a solid wall? My wall’s just three thick but bugs have never made it through. Is this so fewer turrets can be used?

[@raz3465]

Since the biters can actually pathfind to the threats (turrets) they will ignore the walls, meaning upkeep is vastly reduced in comparison to solid walls

[@FurryEskimo]

@@raz3465 Thanks. I seem to recall fire burning walls, perhaps I was mistaken.

[@josephshreeves8192]

3 layer thick walls can also start to fail when you increase default size of swarms. Its not usually an issue until you start clearing stuff out with artillery

[@FurryEskimo]

@@josephshreeves8192 If I edit the game then yes, the effectiveness of my defenses would change, but are these walls just, unbeatable? The bugs seemed to really struggle against some. My only concern would be the size of these walls. They have a wide footprint.

[@josephshreeves8192]

@@FurryEskimo defenses are always resource intensive. The bonus of this design is less turrets (and power) and easier movement. If you don't have the space and plenty of power, solid walls and turret lines still work excellent on default settings

[@derchomat]

Why are there no Spitters? Do they matter?

[@hansdampf640]

not really,spitters attack when way is blocked,they would be just some fuel in this equation

[@wikidude445]

Not enough factories, 0/10

[@knoop7102]

scrimbo

[@cavejohnson982]

Hey, whats the name of the music?

[@katiehesse6578]

Warsaw by Pilotredsun

[@ShadaOfAllThings]

alternative title: Playing plinko with bugs

[@UyeGaming]

Music?

[@knoop7102]

Pilotredsun - Warsaw kzbin.info/www/bejne/q2LQhpx3n62td6s

[@evgenkonyshock4913]

omg PilotRedSun song

[@berzerker4815]

It's possible to make a blueprint ?

[@oguzhanparadoks]

music?

[@turtlememes2334]

what if you use belts to slow the biters down?

[@knoop7102]

Flamethrower fore destroys anything that isn't a stone wall

[@turtlememes2334]

@@knoop7102 oh yeah totally forgot about that, probably isnt worth giving those up

[@Cypekeh]

@Hanako IsMyWaifu You could try but it would be expensive

[@X1ongu]

Biters will destroy belts if they move over them.

[@xGOKOPx]

I'm not too sure but I think belts don't move biters anymore because people were using blue belts to make it impossible for them to get through

[@vs3347]

what's the song?

[@snakeliquid34]

and here I was just using Plasma cannons like a simpleton

[@person-cf3di]

why do the biters go up?

[@slipperyjohn3144]

What is the name of the music?

[@katiehesse6578]

Warsaw by Pilotredsun

[@slipperyjohn3144]

@@katiehesse6578 Cheers!

[@hansdampf640]

whats the name of the music? sounds familiar somehow

[@kirkoboyo5276]

pilotredsun - Achievement track 4: Warsaw

[@hansdampf640]

@@kirkoboyo5276 Thank you :)

[@PluvioZA]

@@kirkoboyo5276 You're such an MVP mate, thank you for this :D

[@user-jr2ue9nu6y]

3 and 5 seem the best. Now combine. Add mines in key choke points and presto

[@massimo946]

Name of mod for placing aliens pls?

[@scrings1164]

Stupid question, but song name?

[@Johnyknowhow]

Warsaw by the talented PilotRedSun AKA PilotRedSky AKA Bittertooth

[@kawgit]

Song name?

[@SaltyRamen.]

What music? So addictive

[@slenderpaul]

kzbin.info/www/bejne/lWGzgpyZmbmZhZo

[@hiteshshivkumar7235]

Warsaw - pilotredsun

[@tagnati5585]

The 5th design is superior in combat

[@Lolinatorishere]

Mfw every factorio player put on the 4 and a half hour prs mix while playing

[@POCKOCMOCVladimir]

5 the best

[@mika_iran]

DUDE!!!! WARSAW BY PILOTREDSUN!?

[@vehicleboi5598]

just a bit of wall talk

[@INGANSHIN]

5!

[@ruslan_yefimov]

Wont work with rampant :(

[@keenandlittle5889]

I don't play Factorio but... what is that song?

[@freshtauwaka7958]

Warsaw by Pilotredsun

[@akyle96]

This is what factrio is about.

[@c2V2ZW54]

5 and 8 are best imho

[@Nyyaaan]

music name?

[@slenderpaul]

kzbin.info/www/bejne/lWGzgpyZmbmZhZo

[@mdr721]

What is that song?

[@knoop7102]

Pilotredsun - Warsaw kzbin.info/www/bejne/q2LQhpx3n62td6s

[@mdr721]

@@knoop7102 thanks! :)

[@CrabQueen]

*many biters were hurt during the filming of this video