Gamedev here with some experience with Procedural Generation (specifically procedurally generating terrain/combat environments) . This is 100% caused by using a pseudo-random number generator (or PRNG) to apply offsets to a set of points. The reason why there are consistent patterns is because the instance of the PRNG being used for orbital strikes is always the same on the client for a given session. PRNG's have some interesting (and sometimes annoying) properties: They are deterministic: PRNG's ALWAYS return the same sequence of numbers if they are using the same seed. This is why a minecraft world (or helldivers map) will look the same with the same seed. Access Order Matters: Since the sequence of numbers is the same, if you change the order of requests the output *will* be different. This can be incredibly annoying with multiplayer synchronization and is probably why this is most likely handled server side (which is why the 'randomness' is the same for each player) This How I think that barrages are handled/how I would do it (this is slightly simplified/not going into technical implementations): - Each barrage is a file/asset that can be changed by designers that effectively defines a set of points that determine the 'perfect' impact offsets. This asset file would also contain the damage/explosion radius, visual effect settings, etc. But most importantly, it would include minimum and maximum spread values. - When a stratagem is triggered, the system would grab the location of the stratagem and then add that to each impact offset value. This creates a 'perfect' impact, from there randomness would be added to get the final impact points. - This is where the PRNG comes in! For each impact the system will query the PRNG and ask it to generate a number between the min and max spread values and add that to the 'perfect' impact point. From there, it's just spawning the effects + explosions in order. - For the walking barrage, this method needs a slight modification. In addition to adding the stratagem coordinates to the impact offsets, you would need to shift/rotate the coordinates. If the impact points are rotated BEFORE applying the randomness, the min/max values will end up having different effects depending on your angle with the highest differences around multiples of 45 degrees and little differences at multiples of 90 degrees. Note that these rotations are relative to the WORLD not the player, so 0/360 would be north. Using this or a similar method to generate impacts would explain the consistency for the shape of the first stratagem and explain why the patterns dramatically change based on order. Another fun fact that can be used to 'prove' this hypothesis: If you fire two barrages WITHOUT the extra salvo upgrade and then WITH the extra salvo upgrade, the first pattern will be identical until the extra shot, but the second pattern will be completely different when done with the salvo upgrade (because the order of access changed :D)! The reason why the order of applying randomness to the walking barrage matters is that if it's applied AFTER rotating, then the offsets are done relative to the compass direction instead of relative to the direction that the stratgem was facing.
Another (former) game dev chiming in here, this is the correct answer. If AH made the initial seed also random then you would get "true" randomness every time. If I was coding this and needed a consistent spread, having the perfect spread template + randomness would be a good way to do it. I would also randomize the order of the salvos as well. Upgrades to the spread could simply shrink the template the randomness works off of
So do you mean that each time the orbital is thrown it has like a "default" pattern (for example, first throw has one default pattern, second throw has another default pattern, and so on) that changes between the min and max spread values generated by the PRGN?
Software engineer here, this is what I thought once he got to the degree part. Slightly randomize a set pattern. They might also use an existing number such as your x,y,z for example
Huge respect to people in geography, geometry and patterns, this is black magic to me but as a mechanical engineer I'm super grateful to those that can do this, definitely vital analysis for any related study. If legitimate democracy is the body of an advanced civilization, research is the food that will keep it going through eons. And space exploration, both robotic and manned, is the physical training that makes it stronger and more capable in the long term.
Man it would be so cool if the devs could give out unique titles for this kind of stuff. That way when you see one in the wild you know you are dealing with a legend. Props to you @cashcrop_ and a well done video
Former Artillerymen. Props to arrowhead for the attention to detail. There are many different kinds of Fire missions. This one most resembles a sweep and zone fire mission. You take the deflection and the quadrant (angles of your cannon) and you adjust them a certain degree every time you fire. Its meant to cover a large area but still be controlled to minimize any collateral and fraticide. The paperwork was a nightmare but always fun and strenuous work
I enjoyed Sweep and Zone missions from the Fire Direction Center side. It was always fun to send the fire commands to the gunline and let them race each other to fire all of their aim points first.
@@StrngeI don't understand a word you said, but I am guessing it translates to "own position" Because if you aren't willing to shell your own position, you aren't willing to win.
@@kyousey there are likely an infinite number of patterns, but you're unlikely to see more than a couple dozen per mission because they start over at the beginning of each mission. That's just how pseudo-random number generators work. Helldivers 2 doesn't need 256 bit encryption to randomize their orbital barrage patterns, so they just use a pseudo-random number generator with a fixed seed value for algorithmically-efficient calculations (aka better optimization)
Automaton 1: What are you trying to tell me? That I can dodge barrages? Automaton 2: No, I’m trying to tell you that when you’re ready, you won’t have to.
Being able to pack this much info into just 20 mins is almost more impressive than the shear amount of work that went into gathering and analyzing the data in the first place.
I just started watching the video, but when I saw 20 minutes I was wondering how it could possibly be that long on such a simple topic. Now I am extra curious. Hahaha.
You know what this sounds like to me: Seeded Psuedorandom distributions. Unfortunately, the conditions and information available to you where you'd want to rely on an orbital bombardment makes it unfeasible to rely on any more than select basic features like "Pattern x is safe in the middle"
I mean signal noise could be number of orbitals and the pattern generation is a matrix of polar coordinates. You might be able to work it backwards with some transforms. Already you have distance to polygon based on throw direction, set up a matrix that represents the orbital of each blast as a function of compass angle then compare the difference in polygonal functions to see what the constant offsets would be from the noise input.
If it was a seeded distribution, then there wouldn't be as much consistency between throw patrerns. Chaos theory would start kicking in at like the 15th iteration or something. Me being lazy would make it a single transfer function that I could throw another variable at to offset the pattern in different directions. That offset function could be another transformation matrix internally changing with orbital throws and externally applying to the vectors that represent hit positions afterward.
I'll look into seeded pseudorandom distributions. Never heard of it. Sounds like my brain would like an understanding. One thing is for certain...Orbitals certainly have a perception of randomness to them, and I like that.
@@cashcrop_ It's the technical term for a type of random number generation where a seeded number is used as the starting point but from there it randomizes. If you use the same seed every time then the same pattern will show up every time but each number in the sequence will appear random at first. If you took this exact same system and simply had it randomize the seed number then it would be for all intents and purposes truly random. Likely some sort of simple number generator but what it does is the game files can record that seed value for use in testing should the engineering team need to help figure out a problem.
that's what i said to my PhD friend who works with GIS on daily basis: "You have never made a better example on how to use GIS and how it works, use it with your students!"
@@danielkrucekjr.4439 That is by far the best part. See a laser going to town? Better make sure there isn't a fucking little scav behind you when the laser switches targets.
I think this is the perfect video to realize that we are all just beings in the search of things interesting to us. No conflict, no argument, just analysis. Very nicely investigated!!!
I used ArcGIS systems for a few years in college for satellite imagery and seeing GIS in a Helldivers 2 vid nearly gave me an aneurysm. I admire and fear your dedication in equal measure.
@@sc3961 Exterminate missions are arguably the best ones to bring barrages on, especially on the highest difficulties. A whole team of barrages makes quick work of the mission. Throw in some intentional traitor strikes during the cd if you're feeling spicy. Use those reinforcements, soldier!
@@sc3961 From Suicidal vs Bots experience (difficulty 7), as long as you use a defendable point away from center, you can cut a lot of time with glorious orbital barrage spam. It actually keeps your death count down when you do it right (Went from being 0-4 to 6-10 reinforcements left on average). Just remember to bring some tools do deal with heavies once the timer runs out.
That's actually a smart move. They make a formula that is effectively random for the players, but predictably manageable for the game designers. And it doesn't require your client to transfer the location of every point of impact to the server and then to every other player, which may create visual and gameplay lags. Instead, your client only transfers the angle of the throw and point of stratagem landing, which is way less info - and then everyone get the same pattern of explosions. Brilliant! An amazing study, helldiver! Super Earth should be proud of you!
Psuedo Random Number Generators and algorithms solve so many problems I wouldn't be surprised to find them in something like video compression for similar reasons.
Yeah, sorry there is nothing. To me, if you don't hover over the red line coming from the sky and react accordingly, it's a training issue. Perhaps the training that we can retake at any time needs an update.
I always use the 380 barrage. People sometimes just kick me, sometimes i got called shit and so on. I don't know how they let these people through basic trainig these days.
@@thereisa6inthename that’s weird. I don’t play with ransoms, but my friends and I use the 120 and 380 all the time on higher difficulties to soften large enemy bases. They’re excellent for that job. You can throw one and then back up, and let it run its course and then clean the rest up by hand. But I also really love the Spear for bots generally and bile titans, so I’d probably get called crazy and kicked too, but I’m over here laughing as I’m taking out bot turrets from 200 meters so
The most useful information in this video is that the area directly on top of the orbital marker seems to be the safest place to stand if you can convince your friends to play 380 roulette with you.
Never underestimate the desire for humans to search for others to preserve life. One of the great things about a species...and sometimes...across species.
I think a factor you forgot to include in your studies, is the fact that the distance you are from the center of the map (Where the Super Destroyer hovers over) can change the angle of Orbital Strategems and how they're casted across the land. This could explain why the walking barrage seemed to have a different pattern in each direction, simply because you were a certain distance from the Ship causing the shots to be angled weirdly. I've seen it happen with the Orbital Airburst strike, where being near the edges of a large map causes the shots to cascade over an area like a sunsetting shadow rather than straight down in a circle as they normally detonate.
This... And, There are only so many guns and cannons on your ship You can count them from your observation window. And each barrel on your ship is the origin point of a round of that type. and it rotates a cannon barrel out of sequence if it has been used like IRL until it's stepped through them all. So pattern changes. It's a ordered barrage not a free fire.
I don't think this information impacts strike *location*. It impacts AOA, for sure, but I think the actual impact remains consistent. This does influence factors like "Is that cliff in the way?" for sure, though.
@@cashcrop_ So basically if my only orbital is the 380, and I remember the patterns and count my throws, I will always know where my shots will land? As well as know the order those shots will land, so I might know when a certain location becomes dangerous or when it becomes safe? Would be kind of funny finding different pathing options through the different 380 patterns that are safe; and just release a montage of you running around inside a 380 barrages like a madman unharmed. Bonus points if you get teammates to stand at what will be impact locations, and you just call out their name just before they are about to blow up, they could even bring the shield emplacement so they don't instantly die to the first impact.
@@cashcrop_Can you plz explain how throwing the 380mm and 120mm at same time changes them? What changes? I always throw the 380mm first, then immediately throw the 120mm at the exact same spot. I also immediately throw a walking barrage into the same area. It’s kinda hard to tell what’s happening 😂 I have many clips of me doing this (throwing 380mm, 120mm, and walking barrage all at once) but it’s all been when playing against enemies. Still would like to know how it changes them, and is there a “best way” to properly throw all 3 of them together, back to back to back? Which one should be thrown first, should I be throwing them all in same spot or spreading them out more? Any tips would be appreciated 😂
@@avenger4cats72 it has nothing to do with throwing them at the same time. What is happening is every orbital that is thrown by you increases a count of how many you have thrown. When you throw your barrage it uses this count to give you the pattern that matches that number. So if it is the 1st orbital you have thrown it will use pattern 1, if it is the 32nd it will use pattern 32. Using your example the 380 will use pattern 1 and the 120 will use pattern 2 and the walking barrage will use pattern 3.
@@Thatbeardedguy86 I use Arc for work. Had to learn QGIS for this video as I didn't want to mix work with personal. It's...odd. Absolutely stellar application for being completely free. I've listened to some presentations where it was used, but had no reason to use it myself. It took a few hours to figure out how to do what I needed to do with it. I was hoping for some GIS folks to sound off. Glad that you did. We are few, far between, misunderstood, and often...underfunded (probably because of misunderstanding...I'm working on my simplification of terminology to bring more folks into the fold). Who knows, perhaps there is one person out there who finds a new path in life from this video. A path of GIS, Geography, or Earth Science. UC?
@@cashcrop_I know a few of the physical scientists who have played with QGIS and echoed what you said. Odd. Nope not UC. I work for an agency that does ocean floor mapping.
The consistency also works with the traitor mechanic. I believe it fires a volley of 3 walking barrage style shots in whatever direction you are moving in. I’ve survived 5+ minutes being fired at, if you turn 90 degrees after the third shot and keep moving you won’t get hit
Helldiver: CMON THROW THE 380 MM BARRAGE, THE BOTS ARE COMING. Helldiver 2: ALR, JUST WAIT A LITTLE, I NEED TO GRAB MY CONSTELATIONS Helldiver: 2: ALR, NOW I GOT THE CONSTELATIONS, wait where are you? *Gets blown up by pieces*
i think the reason the pattern changes is because of the orientation of your ship relative to where the strategem is placed, since the shells get shot from the super destroyer
One of the drop in tips is that as you get to the edge of the map, orbital strikes come in on shallower angles. Don't think that would change the pattern though.
Not sure if you mentioned this in the video, but the walking barrage pattern is the same pattern as the other barrages, just with each shot offset in the thrown direction, based on how long the barrage has been active. Specifically the pattern you can see at 6:25, before the second shot that lands due north of the call-in point. If you look at the centers of the polygons from the walking barrage testing, you can see that they line up with the distribution at that timestamp. Pairs of shots to the east, south, and northwest, single shots in the north and southeast, and a cluster of 4 a bit to the west. I'm also pretty confident all of the patterns are generated programmatically, i.e. the exact coordinates aren't written down anywhere in memory, they're calculated on the fly from a seeded pseudorandom number generator. The seed is presumably initialized to 0 on starting a mission, and advances 1 for each call-in. It might also persist between missions, meaning calling barrages in a previous mission might affect barrages in the next mission if you haven't restarted the game. Therefore, also, there are a functionally infinite number of patterns. You can just keep asking the computer to make more. The specific distribution of the RNG algorithm could be interesting to investigate, like if it makes any efforts to provide an even spread. Based on how close some shots land to each other, I would say it looks like true randomness distributed in either a circle or possibly a square.
Temporal never crossed my mind. That is certainly a possibility. To ascertain a complete understanding...stand in a map for at least 10 minutes and see what pattern of a N-S Walking Barrage Pattern is presented. EZ Edit: Sorry, missed the second part of your point. The continuation of a pattern in a "session" is also an interesting concept that my method would not account for. Your line of thought is fabuous.
Had to learn QGIS, as I'm used to Esri products. Can't mix work and non-work. I saw the possibility and was excited to meld two passions. I'll say one thing that you may like...isobars.
Good video. Good voice. Excellent experiment (my focus was research psychology). And since I'm early here's some advice to the peeps. The 120 and 380 are area denial weapons. I use them on large enemy bases/hives I don't have the time for immediately. They are also amazing at keeping the reinforcing bug holes supressed and obliterating a horde on their way in. 380 better for bugs due to their spread. Walking is very good at it's 1 job: clearing the way as you advance immediately behind it. All ordinance in the game should be used they way they would be in real life scenarios (there is military doctrine to help if you need it)
12:25 what Arrowhead seems to be trying to simulate here with the Walking Barrage is something called a "sheaf". In Field Artillery, we arrange the cannons in lines so we can use a single correct aiming solution for more then one piece. The rounds land along a generally parallel line at generally uniform distance from all the cannons. This is called a sheaf. If a target is left or right, the line remains parallel to the gun-line, which has not moved. Moving would throw off aim, so each gun will pivot in place. Until the gun-line is ordered to displace and emplace facing in another direction, in which case the new sheaf will be parallel to the new facing of the gun-line. It seems like Arrowhead tried to simulate changing the facing of the platform holding the turrets; potentially at a really tiny level cuz math is easier then moving a ship or multiple pieces/vehicles. Oh, this is different from the direction the target observer is facing. Either way, good use of GIS for impact site analysis (Google "changes in the sheaf caused by a change in the line of fire".)
Game developer here (not Arrowhead): It's possible they're using a pseudorandom number generator, but that the seed is fixed. As for randomness potentially dropping all of the bombs on one location, there are solutions to that, ranging from a simple "reroll if wrong" algorithm, to blue noise and Poisson distribution algorithms, or using a random step algorithm.
Thanks for the information. I’m not a developer, and comments such as yours have given me some areas to research to gain a better understanding. About the only development I do is ETL to push data from various sources into GIS.
Developer here ~ It's possible that they did an experimentation with the seeds to get good patterns that fits well in the weapon purpose and list this selected seeds to use in the algorithm. It wouldn't need any other algorithms other than the one using the seed to generate the shots, so would be more predictable and reliable
@@oArquidruida Some others in comments/replies have theorized that the predictability present in the seeds is to allow for easier troubleshooting in replicating problems, as well as requiring less information to sync between the clients.
@@oArquidruida Yes, this is entirely possible too, it's a solid approach for this kind of thing. Test seeds until you find a good run and fix that as the initial seed.
1) This is the kind of video that makes YT still relevant. 2) Thank you for all this time and work. 3) The only question I had about barrages, I believe was not answered : I was wondering if the barrage shells were influenced by the presence of enemies. Since I have found that sometimes shells would do direct hits on chargers or brooders, I thought there was a little "targeting" process going on, but in your video, I only see barrages in the open, with no (or very few) enemies.
I just responded to your other Super Thanks. So, thanks for both! Incredibly kind of you. 1) I appreciate that. I get a bit tired of how YT content is going. I wanted something else and am attempting to make what I want to see. 2) My pleasure. I'd look anyway, so I may as well share what I've found. 3) I haven't seen this to be the case, as there were times when enemies were near the impacts (and sometimes within them), and the patterns were the same. This happened on Study 1, Study 2, and Study 4. There may be something there, but it wasn't noticeable to me. These were all small (standard) units, so maybe it will target larger units. I couldn't say.
can't wait for my friends to use this information to exponentialy increase their chances of utterly missing every potential objective and turning me into paprika dust jokes aside, this is quite a breath of fresh air regarding helldivers 2 content since most channels are milking "nerf drama" for some views, keep it up!
Yep. Typical nerf drama and meta discussions get old fast. There are about twenty channels that do that stuff. This is the only channel I've seen taking this kind of approach to content, and I was hooked the whole way through. Instant subscribe.
@@_Bungus I would argue that the best players know that. But a lot of scrubs I've seen on 5-7 think that the meta weapons/stratagems will do the work for them, which is why they are stuck failing those difficulties.
The best part of this is that: 1. Arrowhead says that everything we do in mission is canon. 2. The government of Super Earth is as it is with all that implies so there is no way the SEAF bothered to specify the barrage system at this level, pay for anything more controlled or for manually targeted barrages, nor distribute any such information that it might have to super destroyers. And those lead to: 3. It makes sense that some super destroyer went and sacrificed a few dozen helldivers getting this data from the field in order to advise other super destroyers. 4. It makes sense that the barrages are coded by the contractors who designed the guns to follow a system like PRNG, just like a videogame. In short, this testing is canon and it makes perfect sense that it happened and that it found the results that it did. I guess we should salute the helldivers whose minute holding the torch of liberty started with taking an aerial photo while dropping in, and then making sure that it is in fact safe to lay down right in the middle of a 380mm, before conveniently falling to Automatons who interrupted the data collection allowing an excuse for the next diver to collect a new aerial photo on their way in. So naturally the next steps are for someone to set up an AI player that uses this to perform orbital barrage wizardry, they get the account banned for botting, in universe that was the MoT finding an automaton infiltrator. All still canon.
Man, that means somewhere in the world there is A: someone who pushed fheir friend and got executed via shotgun for it B: many people who have died bonking their head on the evac C: way too many people willing to drop nukes on evac cuz they like watching the explosions as they leave.
What I was hoping for was quick tips on how to make the barrages more useful. Instead all I got were visual proofs that the rng algorithm is definitionally not random.
Yea, I think the focus on “this isn’t random” ignores that most people aren’t using the word in a “technical” manner. They mean they can’t reliably predict what their barrage will hit, and this video concludes: right, correct, no way for a player to predict that.
Fun fact, Galileo wasn't imprisoned for his scientific findings, after all, the church funded his research. He was imprisoned later for insulting the pope.
@@cousinzeke4888 not quite - Galileo was confined to his house for his scientific arguments (rather than his discoveries) about heliocentrism, particularly in the Dialogo, which is a treatise/compilation of his and others’ largely scientific theories. IIRC Pope Urban was upset that Galileo had not weighed up Geo- and heliocentrism equally in the text (plus the character Simplicio is a buffoonish Aristotelian), paving the way for the church to charge him with, yes, heresy.
Now normally I hate videos that advertise something and say it at the end and just talk about the boring process to get there, but this was actually so interestinf
As a surveyor with a little bit of a background in GIS, I just want to say this is one of the best things I've seen on TH-cam in ages! You've earned that subscription button and a metal!
Another well done and insanely detailed video pal - the dry humor was on point too. I'm sending these requisitions your way for the primer course on GIS alone - keeping doing super earth proud helldiver
I like how this video fully illustrates the slow decent into madness you described in the beginning. It just keeps getting more complicated with every new finding.
Being someone who's into palaeontology and specifically evolutionary biology(as an interest), over the years I've gotten mad respect for the entire field of geology. A lot of people think it's just mostly mineral composition but just like palaeontology, there are sooooo many specializations in the field. I mean, there are people who specialize in bathymetry(Study of lake and riverbeds). And I honestly love how you can apply geology in a practical manner in video games like Helldivers, as a player. Most my interest in evolutionary biology can take me is hypothesizing which bug the shriekers evolved from(My guess being hunters, not just because they share morphology, but also because Hunters have once speciated into stalkers. Not sure if it's actual speciation or more of a 'royal jelly' type of thing, since we do see hunter nymphs but I'm sure we'll find out some day).
But what most people call "the meta" is pure bullshit. People will sit there with apples, bananas and oranges and say bananas are meta because they're longer. Or apples are better because they're more green. It's nonsensical in a game where you aren't just comparing one value to one value.
The pattern might not be random, but the video certainly is. I was not expecting to see this today. As a guy who's worked with GIS and other such things before, this is so much fun to watch.
Been thinking about this since watching the video. One note that came to mind - given that the strikes are physically simulated as being fired from a point above the map, the actual final firing distribution "in code" is likely a set of angle-offsets from the central aim angle, rather than a set of 2d projected impact points. I'd expect that strikes placed farther towards the edge of the map would exhibit a more oval-shaped spread (a conic section) due to the cone of fire being at a shallower angle with the ground.
I did test on the edge at South, East, and North (can't recall anything to the West of center). However, we're talking Level 1 or Level 2 maps. They're small. Many others have brought this up. I'm now curious what the edge of a Level 9 looks like. I've been meaning to look into map sizes by grids present across the maps.
This was my initial thought as well when trying to explain the walking barrage. It could make sense that the super destroyers are actually shooting as if they were players with a gun (probably re-using the same code, same for how automatons shoot I guess). That would be consistent with the kind of simulation-y details Arrowhead seems to love putting into this game. The impact of barrages you see on the ground would be the result of the same spread pattern being projected from the ship's point of view. Same as seeing the spread pattern of a shotgun you've shot into a wall. It would be a simpler explanation for the walking barrage's rotating spread than it having some complex behaviour related to cardinal directions. That said, the rest of the video shows it's more complicated than that. Plus, I would expect the barrages' spreads to look noticeably elongated in some cases given the angle the destroyer sometimes fires at. So idk :)
I worked on this game and spent years tweaking and reworking these barrages and I never noticed this. I can't tell you how much joy it brings me that you spent all this time and effort using actual science to figure out how they work. Thank you for spending the time you did making this video!
This is most probably just a case of them using pseudo random number generators with the seed for the number generator being maybe something like position + angle.
Could be the case. The changes to the distribution pattern are predictable. Perhaps they RNG'd it, then locked it in to not have to process it each time on the fly and reduce calcs that need to be done. I have no clue.
This is awesome and it shows that they do put a lot of thought in their decisions, at least when they have the time to do so... I'm fairly certain dev time is scarce nowadays "thanks" to the guaranteed monthly warbonds, so I wonder if we'll ever get stuff with so much thought put into it as we go forward. The idea of having A LOT of patterns that SEEM random to negate the weird oddball scenarios of impacts being too close to each other or actually never hitting anything if they were full random is pretty interesting. Ultimately, I see the barrages as a good diversion/passive type of clearing out zones, I use the walking barrage often on small camps when I'm cornered or just running elsewhere. It's nice to have options!
This was super helpful in my hypothesis using the 380/120 to clear bug nests that sometimes it’s better to offset your beacon, instead of throwing it square in the middle
All I can think of watching this unfold is the “always sunny in Philadelphia” Pepe Silvia scene: “So I decided, Ohh shit, buddy, I gotta dig a little deeper.” And what dose he find helldiver what’s dose he find?!?!? THERE IS NO “RANDOM ORBITAL STRIKE!!!” ITS ALL MATH AND PATTERNS HELLDIVER!?!?!?
@@cashcrop_ I got a feeling it would fit! Be even better if u could get in touch with a HD2 developer or maybe just a picture of one, to be the “oh he’s the guy who tipped me off about Pepe Silva” 😂
Quick (or not so quick, lol) note on randomness from a software developer. Computer randomness doesn't work the way people expect it to. Pseudo-random is the norm. If you're not using pseudo-random numbers, you're either doing something like cryptography or are doing something _wrong._ True random numbers are expensive in that they require entropy. Your computer only has so much entropy per unit time. If everything on your computer that wanted a random number consumed entropy to do so, you'd literally run out of random numbers. No, that isn't a joke, or hyperbolic. You can literally run out of randomness. Your computer generally collects entropy over time from sources of true random noise. This will come from the environment and usage patterns of the computer. When a program asks for a cryptographically secure random number, a bit of that entropy gets used up. When it runs out, the next thing to ask for a random number "blocks" until there's enough entropy. Blocking is the technical term for when a program freezes. So you _really_ don't want to be touching that kind of RNG unless you have some very specific reason to need your numbers to be true random. Video games don't qualify. So instead they use pseudo-random numbers. These are basically functions that produce extremely variable output that is not correlated with itself (or at least has an extremely low correlation). But the thing is, these random number generators generate the exact same random numbers every time. Literally the same numbers, in the same order, _always._ The solution is called "seeding", where you basically fast forward to a random point in the sequence. And now you've probably spotted a problem. We need a random number to generate random numbers. Fortunately this seed value doesn't need to be _that_ random. It's common to just take the current timestamp and use that, because what you really need is a number that won't be the same twice in a row. That's very easy to do without randomness. Arrowhead appears to have not seeded their random number generator. And your tests appear to be finding some of the interesting properties of the random number generator they use.
And, a similar note is now the pinned comment. Others have noted that not seeding the PRNG is probably deliberate. Makes testing and client synchronization so much easier.
Yeah, this is THE most wonderful and nerdy video Ive seen in a while. Thank you! I work within GIS, but not so much the analytical part as the presentation and distribution of services. This? This is fricking brilliant, and made me want to do a LOT of similar studies for certain games. It also highlights how important robust testing is to any proper research, and how it's so easy to become myopic and narrow on just one point of attack. That's what our brain is best at. Varying and seeing new patterns is important, and learning to look for such useful variations (and the patterns within) is a hard-learned skill. Nearly 300k viewers in a week agree: your days were definitely productive, and you've produced a gem of a video here :)
So what I learned is that the barrages skew far too heavily in focussing on destructive force per impact versus being able to aim those impacts, which makes them impractical to use if you want to destroy specific targets. This is why people prefer the Eagle payloads, as they far better balance destructive payload with ease of use.
Could be a seeded random, or even more likely using a noise function with an offset given, to determine the points for the shapes of the function, so you get predictible distributions. Some Noise also has the upside, that you can get more evenly distributed randomness compared to using random. And that you probably know, but rotating stuff with the golden ratio always creates distributions instead of recognizable patterns. Well it creates a pattern, but the pattern feels evenly distributed. If you manipulate that further, like increasing radiuses or differing shapes, you can ensure e.g. that certain stuff will not overlap too much to reveal that your rotation angle is based on an algorithm, so it is often a baseline. I find your findings interesting, that they basicly use shapes and intersections to get pseudo random distributions, that look a certain way. That is pretty dope. But I am pretty sure, they did not hand-draw 360 of those :D
Interesting. I have zero game dev knowledge. I'll have to look into some of this. And agreed, the distribution patterns were very likely not created by hand. Mine were, heh.
From a programming standpoint, I think what you are seeing IS random (or more accurately pseudo-random) number generation using the count of orbital strikes as a seed. You wouldn't want the strikes to look the same, so the easiest way to make sure strike 14 and 15 are different is to just seed your random number generator with the current strike number. If they used something more random like the current time in ms, or current frame, etc, they couldn't check each barrage pattern. By using an integer with known bounds (maximum possible barrages per player over 45 minutes times 4 players), they could loop through every possible barrage to make sure they all looked good while tweaking values.
The fact that each stratagem use alters the pattern for the next hints at a single random number generator being seeded with the same seed value each time, but incremented by each type of strike. One random number generator for each Helldiver, but that same one incremented for different types of strikes which that one Helldiver casts.
That's what I've gathered from other comments here, as well. I'm not a game dev, so it's been nice to get a better understanding of the results of my study.
Gamedev here with some experience with Procedural Generation (specifically procedurally generating terrain/combat environments) . This is 100% caused by using a pseudo-random number generator (or PRNG) to apply offsets to a set of points. The reason why there are consistent patterns is because the instance of the PRNG being used for orbital strikes is always the same on the client for a given session.
PRNG's have some interesting (and sometimes annoying) properties:
They are deterministic: PRNG's ALWAYS return the same sequence of numbers if they are using the same seed. This is why a minecraft world (or helldivers map) will look the same with the same seed.
Access Order Matters: Since the sequence of numbers is the same, if you change the order of requests the output *will* be different. This can be incredibly annoying with multiplayer synchronization and is probably why this is most likely handled server side (which is why the 'randomness' is the same for each player)
This How I think that barrages are handled/how I would do it (this is slightly simplified/not going into technical implementations):
- Each barrage is a file/asset that can be changed by designers that effectively defines a set of points that determine the 'perfect' impact offsets. This asset file would also contain the damage/explosion radius, visual effect settings, etc. But most importantly, it would include minimum and maximum spread values.
- When a stratagem is triggered, the system would grab the location of the stratagem and then add that to each impact offset value. This creates a 'perfect' impact, from there randomness would be added to get the final impact points.
- This is where the PRNG comes in! For each impact the system will query the PRNG and ask it to generate a number between the min and max spread values and add that to the 'perfect' impact point. From there, it's just spawning the effects + explosions in order.
- For the walking barrage, this method needs a slight modification. In addition to adding the stratagem coordinates to the impact offsets, you would need to shift/rotate the coordinates. If the impact points are rotated BEFORE applying the randomness, the min/max values will end up having different effects depending on your angle with the highest differences around multiples of 45 degrees and little differences at multiples of 90 degrees. Note that these rotations are relative to the WORLD not the player, so 0/360 would be north.
Using this or a similar method to generate impacts would explain the consistency for the shape of the first stratagem and explain why the patterns dramatically change based on order.
Another fun fact that can be used to 'prove' this hypothesis: If you fire two barrages WITHOUT the extra salvo upgrade and then WITH the extra salvo upgrade, the first pattern will be identical until the extra shot, but the second pattern will be completely different when done with the salvo upgrade (because the order of access changed :D)!
The reason why the order of applying randomness to the walking barrage matters is that if it's applied AFTER rotating, then the offsets are done relative to the compass direction instead of relative to the direction that the stratgem was facing.
Another (former) game dev chiming in here, this is the correct answer. If AH made the initial seed also random then you would get "true" randomness every time.
If I was coding this and needed a consistent spread, having the perfect spread template + randomness would be a good way to do it. I would also randomize the order of the salvos as well. Upgrades to the spread could simply shrink the template the randomness works off of
So do you mean that each time the orbital is thrown it has like a "default" pattern (for example, first throw has one default pattern, second throw has another default pattern, and so on) that changes between the min and max spread values generated by the PRGN?
He should pin this comment.
@@leisexe yes but we never see the default pattern as it is always modified by the PRNG
Software engineer here, this is what I thought once he got to the degree part. Slightly randomize a set pattern. They might also use an existing number such as your x,y,z for example
Homeboy got a masters in urban planning and development to figure out why he got teamkilled
LOLLL
Hahahaha
As a Geodesy engineer i salute you. GIS MVP
Me before watching this video: *throws barrage randomly in the nest
Me after watching this video: *throws barrage randomly in the nest
same
@@cashcrop_ dude, I bet at this point you've researched this so much, if you know the number, you can see the impact points. Like neo in matrix.
@cashcrop_ what I wanna see next is you calling in artillery on your own location and dodging all of the hits while everything else dies.
@@OniJitsu I've done this on luck and it felt cool. Being able to do it intentionally would definitely be Neo levels of gameplay
@@OniJitsu ultra instinct level dodging
Anon: "the barrage is random"
Bro: "And I took the personally."
Made me laugh.
Yroue'r*
Came for explosions and democracy, left with a PhD in explosive patternology.
Patternologist should be a rank.
This killed me.
Don’t forget your phd in democracy! o7
Brilliant made me smile… :)))
Huge respect to people in geography, geometry and patterns, this is black magic to me but as a mechanical engineer I'm super grateful to those that can do this, definitely vital analysis for any related study.
If legitimate democracy is the body of an advanced civilization, research is the food that will keep it going through eons. And space exploration, both robotic and manned, is the physical training that makes it stronger and more capable in the long term.
Someone get this Helldiver a medal and assign him to the Ministry of Science and Enginnering. This is an amazing feat
A C-01 would suffice.
Man it would be so cool if the devs could give out unique titles for this kind of stuff. That way when you see one in the wild you know you are dealing with a legend. Props to you @cashcrop_ and a well done video
@@cashcrop_aye, only need that for possible procreation…
@@cashcrop_ At least an extra 2 minute break between dives if nothing else.
😂@@cashcrop_
Former Artillerymen. Props to arrowhead for the attention to detail. There are many different kinds of Fire missions. This one most resembles a sweep and zone fire mission. You take the deflection and the quadrant (angles of your cannon) and you adjust them a certain degree every time you fire. Its meant to cover a large area but still be controlled to minimize any collateral and fraticide. The paperwork was a nightmare but always fun and strenuous work
I enjoyed Sweep and Zone missions from the Fire Direction Center side. It was always fun to send the fire commands to the gunline and let them race each other to fire all of their aim points first.
Until you get a sweep 12 mils 7 deflections Zone 9 mils 9 quadrants
Stop saying it was fun to explode people from a distance you legalized murderers
@@StrngeI don't understand a word you said, but I am guessing it translates to "own position"
Because if you aren't willing to shell your own position, you aren't willing to win.
@@Sotanaht0artillery stuff
- So is it random?
- Well yes but actually no
Legit.
computer RNG in a nutshell
There are theory that suggests "true random" meaning absolutely no pattern what so ever is impossible.
- Not random.
- But is selected between 37 patterns.
@@kyousey there are likely an infinite number of patterns, but you're unlikely to see more than a couple dozen per mission because they start over at the beginning of each mission. That's just how pseudo-random number generators work. Helldivers 2 doesn't need 256 bit encryption to randomize their orbital barrage patterns, so they just use a pseudo-random number generator with a fixed seed value for algorithmically-efficient calculations (aka better optimization)
Automaton 1: What are you trying to tell me? That I can dodge barrages?
Automaton 2: No, I’m trying to tell you that when you’re ready, you won’t have to.
This comment is gold👌 Gave me a good laugh🤣
Golden comment! Might see a flying, dodging bullets Automaton soon if this continues.
Your Democracy Officer would like a word with you.
this was a good reference
Best comment in the entire thread!
I always called the 380 "indiscriminate death." I now know it discriminates
@@james2042 “Actually quite discriminate death”
Gerrymandered death
Being able to pack this much info into just 20 mins is almost more impressive than the shear amount of work that went into gathering and analyzing the data in the first place.
I took a heavy hand to my script this time.
@@cashcrop_you earned my sub because of this xd
@@DaxWolf777 Kind of you.
I just started watching the video, but when I saw 20 minutes I was wondering how it could possibly be that long on such a simple topic. Now I am extra curious. Hahaha.
Same I subbed and watched the whole thing such a refreshing and different video to most on Helldivers I hope you do more of them Ty.
You know what this sounds like to me: Seeded Psuedorandom distributions. Unfortunately, the conditions and information available to you where you'd want to rely on an orbital bombardment makes it unfeasible to rely on any more than select basic features like "Pattern x is safe in the middle"
I mean signal noise could be number of orbitals and the pattern generation is a matrix of polar coordinates.
You might be able to work it backwards with some transforms. Already you have distance to polygon based on throw direction, set up a matrix that represents the orbital of each blast as a function of compass angle then compare the difference in polygonal functions to see what the constant offsets would be from the noise input.
If it was a seeded distribution, then there wouldn't be as much consistency between throw patrerns. Chaos theory would start kicking in at like the 15th iteration or something. Me being lazy would make it a single transfer function that I could throw another variable at to offset the pattern in different directions. That offset function could be another transformation matrix internally changing with orbital throws and externally applying to the vectors that represent hit positions afterward.
@@darbodrake89 Having to Google what you're discussing. No clue. Geography guy. Love it nonetheless.
I'll look into seeded pseudorandom distributions. Never heard of it. Sounds like my brain would like an understanding.
One thing is for certain...Orbitals certainly have a perception of randomness to them, and I like that.
@@cashcrop_ It's the technical term for a type of random number generation where a seeded number is used as the starting point but from there it randomizes. If you use the same seed every time then the same pattern will show up every time but each number in the sequence will appear random at first. If you took this exact same system and simply had it randomize the seed number then it would be for all intents and purposes truly random. Likely some sort of simple number generator but what it does is the game files can record that seed value for use in testing should the engineering team need to help figure out a problem.
YES! Former Maxar employee here. LOVE that you're using GIS in Helldivers 2.
The stars aligned.
As an engineer who uses GIS daily for basin planning...this may be the best use case for GIS software I've ever witnessed.
Finally. What i learned in class will be useful
Facts!!!
It's pretty good for mapping caves too. For food reviews, there's no accounting for taste. Just Google Golden Corral
that's what i said to my PhD friend who works with GIS on daily basis: "You have never made a better example on how to use GIS and how it works, use it with your students!"
@@bofa722You have a beautiful profile picture.
"They're not random bro you just need to conduct a geostatistical study to find a way to make them usable"
Meanwhile the orbital laser.
@@dr.vanilla9017lights me on fire if I’m within 20 feet of anything that moves and isn’t a helldiver
@@danielkrucekjr.4439 That is by far the best part. See a laser going to town? Better make sure there isn't a fucking little scav behind you when the laser switches targets.
-least unhinged Dark Souls/Eldren Ring player
@@dr.vanilla9017 spends 5 minutes on a strider, doesnt kill it, leaves. the laser is overrated dogshit
I think this is the perfect video to realize that we are all just beings in the search of things interesting to us. No conflict, no argument, just analysis.
Very nicely investigated!!!
Thank you.
When someone asks me what I do for a living as a GIS Analyst, I'm going to stop saying "Google Maps", and just link this video - good job dude!
I say I make maps.
@@cashcrop_ Are you a GIS professional? So funny seeing QGIS in a helldivers video.
@@FishyBusiness69 i am
Nerds - love it lmao
When searching parcels, I use the GIS base map and never gave it thought as what the GIS actually is.
I used ArcGIS systems for a few years in college for satellite imagery and seeing GIS in a Helldivers 2 vid nearly gave me an aneurysm. I admire and fear your dedication in equal measure.
When that screen went up i needed to pause the video it is truly an universal experience
"25 days of my life"
as a person who brings 380 in ALL my runs. i thank you for your sacrifice.
Even at exterminate missions?
@@sc3961 i mean, the video proves that its not random, so why not lol
@@sc3961 Exterminate missions are arguably the best ones to bring barrages on, especially on the highest difficulties. A whole team of barrages makes quick work of the mission. Throw in some intentional traitor strikes during the cd if you're feeling spicy. Use those reinforcements, soldier!
@@sc3961especially for exterminate missions!!! A fellow 380 truther I see
@@sc3961 From Suicidal vs Bots experience (difficulty 7), as long as you use a defendable point away from center, you can cut a lot of time with glorious orbital barrage spam. It actually keeps your death count down when you do it right (Went from being 0-4 to 6-10 reinforcements left on average).
Just remember to bring some tools do deal with heavies once the timer runs out.
That's actually a smart move. They make a formula that is effectively random for the players, but predictably manageable for the game designers. And it doesn't require your client to transfer the location of every point of impact to the server and then to every other player, which may create visual and gameplay lags. Instead, your client only transfers the angle of the throw and point of stratagem landing, which is way less info - and then everyone get the same pattern of explosions. Brilliant!
An amazing study, helldiver! Super Earth should be proud of you!
Psuedo Random Number Generators and algorithms solve so many problems I wouldn't be surprised to find them in something like video compression for similar reasons.
@@arthurmoore9488 woah... I never thought of using it in anything but games. What you said makes so much sense!
Helldiver at day, scientist at night
give this man class A citizenship
Super citizenship
time to tell my friends that the barrage isn't random and give them no context to frustrate them further when i take a 380 hehe
Yeah, sorry there is nothing. To me, if you don't hover over the red line coming from the sky and react accordingly, it's a training issue. Perhaps the training that we can retake at any time needs an update.
@@cashcrop_yeah, it's odd that there isn't a point where you learn to identify ally strategems given that it's a one step process.
I always use the 380 barrage. People sometimes just kick me, sometimes i got called shit and so on.
I don't know how they let these people through basic trainig these days.
@@thereisa6inthename that’s weird. I don’t play with ransoms, but my friends and I use the 120 and 380 all the time on higher difficulties to soften large enemy bases. They’re excellent for that job. You can throw one and then back up, and let it run its course and then clean the rest up by hand.
But I also really love the Spear for bots generally and bile titans, so I’d probably get called crazy and kicked too, but I’m over here laughing as I’m taking out bot turrets from 200 meters so
The most useful information in this video is that the area directly on top of the orbital marker seems to be the safest place to stand if you can convince your friends to play 380 roulette with you.
This has the same energy as trying to find planes that crash into the ocean
Never underestimate the desire for humans to search for others to preserve life. One of the great things about a species...and sometimes...across species.
So it really does hit everywhere except where you throw.
And my squad kept calling me crazy for dropping it at my feet.
I think a factor you forgot to include in your studies, is the fact that the distance you are from the center of the map (Where the Super Destroyer hovers over) can change the angle of Orbital Strategems and how they're casted across the land. This could explain why the walking barrage seemed to have a different pattern in each direction, simply because you were a certain distance from the Ship causing the shots to be angled weirdly.
I've seen it happen with the Orbital Airburst strike, where being near the edges of a large map causes the shots to cascade over an area like a sunsetting shadow rather than straight down in a circle as they normally detonate.
They even say this in one of the hints/tips during loading screens.
This... And, There are only so many guns and cannons on your ship You can count them from your observation window. And each barrel on your ship is the origin point of a round of that type. and it rotates a cannon barrel out of sequence if it has been used like IRL until it's stepped through them all. So pattern changes. It's a ordered barrage not a free fire.
Was just about to type something similar but thought i'd check if anyone else had the same thought 🙂
He didn't say it like this, but this is the 360 (Constellation) he was talking about
I don't think this information impacts strike *location*. It impacts AOA, for sure, but I think the actual impact remains consistent. This does influence factors like "Is that cliff in the way?" for sure, though.
I respect the thoroughness even if it didn't help me use barrages better
I'm really good at throwing that first one. Heh.
@@cashcrop_at least you got the strstagem codes down to a T
@@cashcrop_ So basically if my only orbital is the 380, and I remember the patterns and count my throws, I will always know where my shots will land? As well as know the order those shots will land, so I might know when a certain location becomes dangerous or when it becomes safe?
Would be kind of funny finding different pathing options through the different 380 patterns that are safe; and just release a montage of you running around inside a 380 barrages like a madman unharmed. Bonus points if you get teammates to stand at what will be impact locations, and you just call out their name just before they are about to blow up, they could even bring the shield emplacement so they don't instantly die to the first impact.
@@cashcrop_Can you plz explain how throwing the 380mm and 120mm at same time changes them? What changes? I always throw the 380mm first, then immediately throw the 120mm at the exact same spot. I also immediately throw a walking barrage into the same area. It’s kinda hard to tell what’s happening 😂
I have many clips of me doing this (throwing 380mm, 120mm, and walking barrage all at once) but it’s all been when playing against enemies.
Still would like to know how it changes them, and is there a “best way” to properly throw all 3 of them together, back to back to back? Which one should be thrown first, should I be throwing them all in same spot or spreading them out more? Any tips would be appreciated 😂
@@avenger4cats72 it has nothing to do with throwing them at the same time. What is happening is every orbital that is thrown by you increases a count of how many you have thrown. When you throw your barrage it uses this count to give you the pattern that matches that number. So if it is the 1st orbital you have thrown it will use pattern 1, if it is the 32nd it will use pattern 32. Using your example the 380 will use pattern 1 and the 120 will use pattern 2 and the walking barrage will use pattern 3.
"Somewhere between 38 distrubution patterns and hell" killed me
Love the surprise villian in Act 3 being polygons
haha. good one.
Polygons are F tier of geometry
Did find it funny that he used circles to first bring up polygons.
@dineshsadhwani3717 take that back you son of a bot
@@dineshsadhwani3717 take that back you son of a bot
And then it turns out that all the dots in a 360 degree circle form a QR code that, if scanned, gets you rick rolled by the developers.
That would have been wonderful.
The 380 has been my favorite stratagem, even before the buffs, nothing shuts down a bug breach or bot drop like a 380
The production and dedication is second to none. Having a GIS breakdown is mega!
I was excited to meld two passions of mine.
Haha seriously. We use ArcGIS at work and I was super excited to see it used here.
@@Thatbeardedguy86 I use Arc for work. Had to learn QGIS for this video as I didn't want to mix work with personal. It's...odd. Absolutely stellar application for being completely free. I've listened to some presentations where it was used, but had no reason to use it myself. It took a few hours to figure out how to do what I needed to do with it. I was hoping for some GIS folks to sound off. Glad that you did. We are few, far between, misunderstood, and often...underfunded (probably because of misunderstanding...I'm working on my simplification of terminology to bring more folks into the fold). Who knows, perhaps there is one person out there who finds a new path in life from this video. A path of GIS, Geography, or Earth Science.
UC?
@@cashcrop_I know a few of the physical scientists who have played with QGIS and echoed what you said. Odd. Nope not UC. I work for an agency that does ocean floor mapping.
@Thatbeardedguy86 discord.gg/QCjzSwA
This is where I typically am when gaming. Good group of folks. We like testing things. C'mon. (expires in 7 days)
The most captivating Helldivers 2 video I've watched in weeks! Bravo!
This is a very humbling comment. Thank you.
You’re one of the scientists we have to save in the evacuation missions.
I cloned myself. So, I'm actually all of them.
This is the kind of research the government needs to be funding.
Holy what are you doing here?
They are, the Forest Service has both howitzers and GIS. All in the responsibility of the GS5 Mafia.
The consistency also works with the traitor mechanic. I believe it fires a volley of 3 walking barrage style shots in whatever direction you are moving in. I’ve survived 5+ minutes being fired at, if you turn 90 degrees after the third shot and keep moving you won’t get hit
Fabulous. Some have asked about this. I'm going to try to figure out how to link to a comment if it's possible.
Watching things like this is humbling because it reminds me that there’s people that are at least 300% smarter than me
20:30 I love how you just run up, vibe check that automaton, and board the ship. "I like ya cut, G"
He was on my landing pad. Can't have that.
Helldiver: CMON THROW THE 380 MM BARRAGE, THE BOTS ARE COMING.
Helldiver 2: ALR, JUST WAIT A LITTLE, I NEED TO GRAB MY CONSTELATIONS
Helldiver: 2: ALR, NOW I GOT THE CONSTELATIONS, wait where are you?
*Gets blown up by pieces*
i think the reason the pattern changes is because of the orientation of your ship relative to where the strategem is placed, since the shells get shot from the super destroyer
One of the drop in tips is that as you get to the edge of the map, orbital strikes come in on shallower angles. Don't think that would change the pattern though.
Not sure if you mentioned this in the video, but the walking barrage pattern is the same pattern as the other barrages, just with each shot offset in the thrown direction, based on how long the barrage has been active. Specifically the pattern you can see at 6:25, before the second shot that lands due north of the call-in point. If you look at the centers of the polygons from the walking barrage testing, you can see that they line up with the distribution at that timestamp. Pairs of shots to the east, south, and northwest, single shots in the north and southeast, and a cluster of 4 a bit to the west.
I'm also pretty confident all of the patterns are generated programmatically, i.e. the exact coordinates aren't written down anywhere in memory, they're calculated on the fly from a seeded pseudorandom number generator. The seed is presumably initialized to 0 on starting a mission, and advances 1 for each call-in. It might also persist between missions, meaning calling barrages in a previous mission might affect barrages in the next mission if you haven't restarted the game. Therefore, also, there are a functionally infinite number of patterns. You can just keep asking the computer to make more.
The specific distribution of the RNG algorithm could be interesting to investigate, like if it makes any efforts to provide an even spread. Based on how close some shots land to each other, I would say it looks like true randomness distributed in either a circle or possibly a square.
Temporal never crossed my mind. That is certainly a possibility. To ascertain a complete understanding...stand in a map for at least 10 minutes and see what pattern of a N-S Walking Barrage Pattern is presented. EZ
Edit: Sorry, missed the second part of your point. The continuation of a pattern in a "session" is also an interesting concept that my method would not account for. Your line of thought is fabuous.
As a meteorologist and fellow geographer it’s so awesome to see someone using GIS like this
Had to learn QGIS, as I'm used to Esri products. Can't mix work and non-work. I saw the possibility and was excited to meld two passions. I'll say one thing that you may like...isobars.
discord.gg/QCjzSwA
This is where I typically am when gaming. Good group of folks. We like testing things. C'mon. (expires in 7 days)
It expired fast :( i test things too sometimes :,( @@cashcrop_
@@cashcrop_ Already expired :p
Bro you should start to write your PhD thesis about Helldivers barrage distribution lol… This is insane job well done!
Good video. Good voice. Excellent experiment (my focus was research psychology).
And since I'm early here's some advice to the peeps. The 120 and 380 are area denial weapons. I use them on large enemy bases/hives I don't have the time for immediately. They are also amazing at keeping the reinforcing bug holes supressed and obliterating a horde on their way in. 380 better for bugs due to their spread. Walking is very good at it's 1 job: clearing the way as you advance immediately behind it.
All ordinance in the game should be used they way they would be in real life scenarios (there is military doctrine to help if you need it)
i did not expect helldivers to overlap with my experience in autocad engineering/gis but here we are, thanks for putting in the work!
A point is a point, a polygon is a polygon...but...WTF is a Donut? Do you use Object Data?
discord.gg/QCjzSwA
This is where I typically am when gaming. Good group of folks. We like testing things. C'mon. (expires in 7 days)
@@cashcrop_ Happy to be there, I'm gonna learn a lot from all you!
This lowkey inspired me to keep being passionate about both gaming and science
The obtuseness of Fatshark has truly prepared all of their former youtube squadies for Helldivers
Yeah I'd noticed a whole lot of dark/vermtubers have been making the best at helldivers 2 content
I don't know what's happening over at FS. I wish it was more than what it is. DT is a really fun game.
obtuse rubber goose
@@cashcrop_ I couldn't agree more! I'm fucking loving Helldivers 2 as well! xD
Fatshark has a long history of similar behavior. Do you remember war of the vikings?@@cashcrop_
12:25 what Arrowhead seems to be trying to simulate here with the Walking Barrage is something called a "sheaf". In Field Artillery, we arrange the cannons in lines so we can use a single correct aiming solution for more then one piece. The rounds land along a generally parallel line at generally uniform distance from all the cannons. This is called a sheaf. If a target is left or right, the line remains parallel to the gun-line, which has not moved. Moving would throw off aim, so each gun will pivot in place. Until the gun-line is ordered to displace and emplace facing in another direction, in which case the new sheaf will be parallel to the new facing of the gun-line. It seems like Arrowhead tried to simulate changing the facing of the platform holding the turrets; potentially at a really tiny level cuz math is easier then moving a ship or multiple pieces/vehicles. Oh, this is different from the direction the target observer is facing. Either way, good use of GIS for impact site analysis (Google "changes in the sheaf caused by a change in the line of fire".)
*than
Then is for sequence; than is for comparison. ‘If the guns are moved more than a bit then helldivers will die.’
Game developer here (not Arrowhead):
It's possible they're using a pseudorandom number generator, but that the seed is fixed.
As for randomness potentially dropping all of the bombs on one location, there are solutions to that, ranging from a simple "reroll if wrong" algorithm, to blue noise and Poisson distribution algorithms, or using a random step algorithm.
Thanks for the information. I’m not a developer, and comments such as yours have given me some areas to research to gain a better understanding. About the only development I do is ETL to push data from various sources into GIS.
Developer here ~
It's possible that they did an experimentation with the seeds to get good patterns that fits well in the weapon purpose and list this selected seeds to use in the algorithm.
It wouldn't need any other algorithms other than the one using the seed to generate the shots, so would be more predictable and reliable
@@oArquidruida Some others in comments/replies have theorized that the predictability present in the seeds is to allow for easier troubleshooting in replicating problems, as well as requiring less information to sync between the clients.
@@oArquidruida Yes, this is entirely possible too, it's a solid approach for this kind of thing. Test seeds until you find a good run and fix that as the initial seed.
Why not just type artilleryisrandom=true? Game dev are stupid
I feel sorry for your sanity after that one! Good work HELLDIVER!
Thanks. It is finally over.
@@cashcrop_ Sounds like treason to me /j
Good work my friend! This content was truly incredibly.
what the actual fuck.... this is insane
1) This is the kind of video that makes YT still relevant.
2) Thank you for all this time and work.
3) The only question I had about barrages, I believe was not answered : I was wondering if the barrage shells were influenced by the presence of enemies. Since I have found that sometimes shells would do direct hits on chargers or brooders, I thought there was a little "targeting" process going on, but in your video, I only see barrages in the open, with no (or very few) enemies.
I just responded to your other Super Thanks. So, thanks for both! Incredibly kind of you.
1) I appreciate that. I get a bit tired of how YT content is going. I wanted something else and am attempting to make what I want to see.
2) My pleasure. I'd look anyway, so I may as well share what I've found.
3) I haven't seen this to be the case, as there were times when enemies were near the impacts (and sometimes within them), and the patterns were the same. This happened on Study 1, Study 2, and Study 4. There may be something there, but it wasn't noticeable to me. These were all small (standard) units, so maybe it will target larger units. I couldn't say.
can't wait for my friends to use this information to exponentialy increase their chances of utterly missing every potential objective and turning me into paprika dust
jokes aside, this is quite a breath of fresh air regarding helldivers 2 content since most channels are milking "nerf drama" for some views, keep it up!
Not here. I'd hate myself.
Yep. Typical nerf drama and meta discussions get old fast. There are about twenty channels that do that stuff. This is the only channel I've seen taking this kind of approach to content, and I was hooked the whole way through. Instant subscribe.
@@_Bungus I would argue that the best players know that. But a lot of scrubs I've seen on 5-7 think that the meta weapons/stratagems will do the work for them, which is why they are stuck failing those difficulties.
This guy is actually fucking insane great vid dude holy
Not clinically, though.
The best part of this is that:
1. Arrowhead says that everything we do in mission is canon.
2. The government of Super Earth is as it is with all that implies so there is no way the SEAF bothered to specify the barrage system at this level, pay for anything more controlled or for manually targeted barrages, nor distribute any such information that it might have to super destroyers.
And those lead to:
3. It makes sense that some super destroyer went and sacrificed a few dozen helldivers getting this data from the field in order to advise other super destroyers.
4. It makes sense that the barrages are coded by the contractors who designed the guns to follow a system like PRNG, just like a videogame.
In short, this testing is canon and it makes perfect sense that it happened and that it found the results that it did.
I guess we should salute the helldivers whose minute holding the torch of liberty started with taking an aerial photo while dropping in, and then making sure that it is in fact safe to lay down right in the middle of a 380mm, before conveniently falling to Automatons who interrupted the data collection allowing an excuse for the next diver to collect a new aerial photo on their way in.
So naturally the next steps are for someone to set up an AI player that uses this to perform orbital barrage wizardry, they get the account banned for botting, in universe that was the MoT finding an automaton infiltrator. All still canon.
Wait. Real talk, Arrowhead said that literally EVERYTHING we do ingame is canon?
Man, that means somewhere in the world there is
A: someone who pushed fheir friend and got executed via shotgun for it
B: many people who have died bonking their head on the evac
C: way too many people willing to drop nukes on evac cuz they like watching the explosions as they leave.
What I was hoping for was quick tips on how to make the barrages more useful. Instead all I got were visual proofs that the rng algorithm is definitionally not random.
Your sentences in order:
1: Sorry, not a quick tips kind of guy (I try)
2: Yup
Yea, I think the focus on “this isn’t random” ignores that most people aren’t using the word in a “technical” manner. They mean they can’t reliably predict what their barrage will hit, and this video concludes: right, correct, no way for a player to predict that.
I dub thee "The Galileo of Videogames". Now report to your cell for heresy....
Heresy? This ain't 40k lad.
@@insertnamehere6659 Galileo reference, not a 40K reference.
Fun fact, Galileo wasn't imprisoned for his scientific findings, after all, the church funded his research. He was imprisoned later for insulting the pope.
@@insertnamehere6659 Spotted the Warhammer 40K fan.
@@cousinzeke4888 not quite - Galileo was confined to his house for his scientific arguments (rather than his discoveries) about heliocentrism, particularly in the Dialogo, which is a treatise/compilation of his and others’ largely scientific theories. IIRC Pope Urban was upset that Galileo had not weighed up Geo- and heliocentrism equally in the text (plus the character Simplicio is a buffoonish Aristotelian), paving the way for the church to charge him with, yes, heresy.
Now normally I hate videos that advertise something and say it at the end and just talk about the boring process to get there, but this was actually so interestinf
As a surveyor with a little bit of a background in GIS, I just want to say this is one of the best things I've seen on TH-cam in ages! You've earned that subscription button and a metal!
This guy is actually insane ... the ministry of science does approve of this.
The notion of barrages not being random is undemocratic automaton propaganda and I will report him to my democracy officer!
Bro i’m high and listened to every word. Didn’t understand anything but you entertained me for 3 hours with this video on a loop. Thank you.
Comment of the day.
I'll admit, I clicked on this expecting to click away 5 minutes in. I stayed and watched the whole thing. Thank you for ALL of this work.
This is a humbling comment. I don't write that often. Thank you.
Minimizing 380 mm barrage friendly casualties
Minimizing?... no, MAXIMIZING 😈
Another well done and insanely detailed video pal - the dry humor was on point too. I'm sending these requisitions your way for the primer course on GIS alone - keeping doing super earth proud helldiver
Very much appreciated.
SES Agent of Science is the perfect ship name for this video. So much research. This is why I love helldivers.. it's more than just a game.
I like how this video fully illustrates the slow decent into madness you described in the beginning. It just keeps getting more complicated with every new finding.
When you like Helldivers 2, but you really LOVE math, and maps.
Being someone who's into palaeontology and specifically evolutionary biology(as an interest), over the years I've gotten mad respect for the entire field of geology. A lot of people think it's just mostly mineral composition but just like palaeontology, there are sooooo many specializations in the field. I mean, there are people who specialize in bathymetry(Study of lake and riverbeds).
And I honestly love how you can apply geology in a practical manner in video games like Helldivers, as a player. Most my interest in evolutionary biology can take me is hypothesizing which bug the shriekers evolved from(My guess being hunters, not just because they share morphology, but also because Hunters have once speciated into stalkers. Not sure if it's actual speciation or more of a 'royal jelly' type of thing, since we do see hunter nymphs but I'm sure we'll find out some day).
Geology rocks. =D
Devs that think they can fight the meta seem to not understand that people like this exist. lol. The data will be found.
Exactly! XD
But what most people call "the meta" is pure bullshit.
People will sit there with apples, bananas and oranges and say bananas are meta because they're longer.
Or apples are better because they're more green.
It's nonsensical in a game where you aren't just comparing one value to one value.
@@kamikeserpentail3778 Especially since the community (especially Reddit in that case) confuses ease of use and reliability for power.
Her: why is he not responding to my texts? ? I bet he is cheating on me.
Him :
Bro casually dropped the most scientific, in depth, informative, well presented, and entertaining Helldivers 2 video of all time.
I think the center of the circles is potentially where the orbital artillery is positioned physically above the map.
So many helldivers lost to give us this information. Your service to super earth iscommendable sir
The pattern might not be random, but the video certainly is. I was not expecting to see this today. As a guy who's worked with GIS and other such things before, this is so much fun to watch.
Been thinking about this since watching the video. One note that came to mind - given that the strikes are physically simulated as being fired from a point above the map, the actual final firing distribution "in code" is likely a set of angle-offsets from the central aim angle, rather than a set of 2d projected impact points. I'd expect that strikes placed farther towards the edge of the map would exhibit a more oval-shaped spread (a conic section) due to the cone of fire being at a shallower angle with the ground.
I did test on the edge at South, East, and North (can't recall anything to the West of center). However, we're talking Level 1 or Level 2 maps. They're small.
Many others have brought this up. I'm now curious what the edge of a Level 9 looks like.
I've been meaning to look into map sizes by grids present across the maps.
This was my initial thought as well when trying to explain the walking barrage. It could make sense that the super destroyers are actually shooting as if they were players with a gun (probably re-using the same code, same for how automatons shoot I guess). That would be consistent with the kind of simulation-y details Arrowhead seems to love putting into this game.
The impact of barrages you see on the ground would be the result of the same spread pattern being projected from the ship's point of view. Same as seeing the spread pattern of a shotgun you've shot into a wall. It would be a simpler explanation for the walking barrage's rotating spread than it having some complex behaviour related to cardinal directions.
That said, the rest of the video shows it's more complicated than that. Plus, I would expect the barrages' spreads to look noticeably elongated in some cases given the angle the destroyer sometimes fires at.
So idk :)
Holy shit, dude. As a Veteran getting his BS in GIS, my head turned so fast with the mention of GCP's and Georeferencing. Solid video, hell yeah.
I worked on this game and spent years tweaking and reworking these barrages and I never noticed this. I can't tell you how much joy it brings me that you spent all this time and effort using actual science to figure out how they work. Thank you for spending the time you did making this video!
Aw, man. You're making me blush. Thank you for taking the time to watch me tip-toe around insanity.
Curious that your account was made today.
Did you know that AFK Journey is a game you can not play yet play?
Having worked in GIS before this is making my IRL nerdiness and Helldiver nerd happy. This is awesome work!
This is most probably just a case of them using pseudo random number generators with the seed for the number generator being maybe something like position + angle.
Could be the case. The changes to the distribution pattern are predictable. Perhaps they RNG'd it, then locked it in to not have to process it each time on the fly and reduce calcs that need to be done. I have no clue.
This is awesome and it shows that they do put a lot of thought in their decisions, at least when they have the time to do so... I'm fairly certain dev time is scarce nowadays "thanks" to the guaranteed monthly warbonds, so I wonder if we'll ever get stuff with so much thought put into it as we go forward.
The idea of having A LOT of patterns that SEEM random to negate the weird oddball scenarios of impacts being too close to each other or actually never hitting anything if they were full random is pretty interesting.
Ultimately, I see the barrages as a good diversion/passive type of clearing out zones, I use the walking barrage often on small camps when I'm cornered or just running elsewhere. It's nice to have options!
This was super helpful in my hypothesis using the 380/120 to clear bug nests that sometimes it’s better to offset your beacon, instead of throwing it square in the middle
if you remember these patterns while playing your stratagem barrages must be insane,
dude probably walks through his own barrages untouched
All I can think of watching this unfold is the “always sunny in Philadelphia” Pepe Silvia scene:
“So I decided, Ohh shit, buddy, I gotta dig a little deeper.” And what dose he find helldiver what’s dose he find?!?!? THERE IS NO “RANDOM ORBITAL STRIKE!!!” ITS ALL MATH AND PATTERNS HELLDIVER!?!?!?
Heh, someone on our Discord asked if I was going to do a thumbnail using that.
@@cashcrop_ I got a feeling it would fit! Be even better if u could get in touch with a HD2 developer or maybe just a picture of one, to be the “oh he’s the guy who tipped me off about Pepe Silva” 😂
Ive grown to love the walking barage i use it on bot bases like alot of dudes use the laser but i get way more uses out of it
Quick (or not so quick, lol) note on randomness from a software developer. Computer randomness doesn't work the way people expect it to. Pseudo-random is the norm. If you're not using pseudo-random numbers, you're either doing something like cryptography or are doing something _wrong._
True random numbers are expensive in that they require entropy. Your computer only has so much entropy per unit time. If everything on your computer that wanted a random number consumed entropy to do so, you'd literally run out of random numbers. No, that isn't a joke, or hyperbolic. You can literally run out of randomness.
Your computer generally collects entropy over time from sources of true random noise. This will come from the environment and usage patterns of the computer. When a program asks for a cryptographically secure random number, a bit of that entropy gets used up. When it runs out, the next thing to ask for a random number "blocks" until there's enough entropy. Blocking is the technical term for when a program freezes.
So you _really_ don't want to be touching that kind of RNG unless you have some very specific reason to need your numbers to be true random. Video games don't qualify. So instead they use pseudo-random numbers. These are basically functions that produce extremely variable output that is not correlated with itself (or at least has an extremely low correlation). But the thing is, these random number generators generate the exact same random numbers every time. Literally the same numbers, in the same order, _always._ The solution is called "seeding", where you basically fast forward to a random point in the sequence.
And now you've probably spotted a problem. We need a random number to generate random numbers. Fortunately this seed value doesn't need to be _that_ random. It's common to just take the current timestamp and use that, because what you really need is a number that won't be the same twice in a row. That's very easy to do without randomness.
Arrowhead appears to have not seeded their random number generator. And your tests appear to be finding some of the interesting properties of the random number generator they use.
And, a similar note is now the pinned comment. Others have noted that not seeding the PRNG is probably deliberate. Makes testing and client synchronization so much easier.
A lot of fancy words and studies to ultimately learn nothing practically helpful. Still subbed as this is the level of insanity I like to see.
Yea man i cant learn anything helpful but respect his hard work
Science takes time! You do the science and then you realize that you made a mistake (or several), then you do more science!
Yeah, this is THE most wonderful and nerdy video Ive seen in a while. Thank you! I work within GIS, but not so much the analytical part as the presentation and distribution of services. This? This is fricking brilliant, and made me want to do a LOT of similar studies for certain games. It also highlights how important robust testing is to any proper research, and how it's so easy to become myopic and narrow on just one point of attack. That's what our brain is best at. Varying and seeing new patterns is important, and learning to look for such useful variations (and the patterns within) is a hard-learned skill. Nearly 300k viewers in a week agree: your days were definitely productive, and you've produced a gem of a video here :)
"if you study something long enough, it begins to study you" great video Helldiver, very informative. May Democracy be with you. 🦅
So what I learned is that the barrages skew far too heavily in focussing on destructive force per impact versus being able to aim those impacts, which makes them impractical to use if you want to destroy specific targets.
This is why people prefer the Eagle payloads, as they far better balance destructive payload with ease of use.
ok, you get a like for simply geo-mapping the grid-squares onto tour chosen map-location.
Video continues.
The absolute dedication to the craft.. jeez, dude
Love your stuff and looking forward to the next 🍻
You conducted your experiments very well. Great video, you're the kind of obsessed that I aspire to be
Could be a seeded random, or even more likely using a noise function with an offset given, to determine the points for the shapes of the function, so you get predictible distributions.
Some Noise also has the upside, that you can get more evenly distributed randomness compared to using random.
And that you probably know, but rotating stuff with the golden ratio always creates distributions instead of recognizable patterns. Well it creates a pattern, but the pattern feels evenly distributed. If you manipulate that further, like increasing radiuses or differing shapes, you can ensure e.g. that certain stuff will not overlap too much to reveal that your rotation angle is based on an algorithm, so it is often a baseline.
I find your findings interesting, that they basicly use shapes and intersections to get pseudo random distributions, that look a certain way. That is pretty dope. But I am pretty sure, they did not hand-draw 360 of those :D
Interesting. I have zero game dev knowledge. I'll have to look into some of this.
And agreed, the distribution patterns were very likely not created by hand. Mine were, heh.
Nice touch with your ships name being the 'agent of science' ;)
It was between that or Fist of Peace.
From a programming standpoint, I think what you are seeing IS random (or more accurately pseudo-random) number generation using the count of orbital strikes as a seed. You wouldn't want the strikes to look the same, so the easiest way to make sure strike 14 and 15 are different is to just seed your random number generator with the current strike number.
If they used something more random like the current time in ms, or current frame, etc, they couldn't check each barrage pattern. By using an integer with known bounds (maximum possible barrages per player over 45 minutes times 4 players), they could loop through every possible barrage to make sure they all looked good while tweaking values.
What 1 day of no updates does to a community
THIS WAS AN AMAZING VIDEO I COULDNT STOP SMILING! BIG PROPS!
Gracias.
The fact that each stratagem use alters the pattern for the next hints at a single random number generator being seeded with the same seed value each time, but incremented by each type of strike. One random number generator for each Helldiver, but that same one incremented for different types of strikes which that one Helldiver casts.
That's what I've gathered from other comments here, as well. I'm not a game dev, so it's been nice to get a better understanding of the results of my study.
Sheldon Cooper stated there is no such thing as random in computer games.
Funnily enough, there is no such thing as random anywhere, only unknown variables
Let's gooo!!