Rimworld Testing : Shooting training and Capturing Enemies : Tutorial Nugget
ฝัง
- เผยแพร่เมื่อ 15 มิ.ย. 2020
- When it comes to capturing enemies does the machine pistol have an advantage.... not really as far as I can see. Looks to be random if the enemies end up dead or incapacitated.
When it comes to training shooting skill, things are a little un-intuitive. All weapons generate the same about of exp for time spent shooting. That means their is no difference between using an auto pistol or a sniper rifle for exp gain. So it's just a case of figuring out which gun can miss the most so you can farm the most exp. Winner is the awful Sniper rifle with the bolt action being a close second. Auto pistol gets an Honorable mention.
Capture testing save map
drive.google.com/file/d/1r25K...
Snipe rifle skill gain testing save
drive.google.com/file/d/18nMh...
Spreadsheets
Down chances per weapon
docs.google.com/spreadsheets/...
Exp gain per weapon
docs.google.com/spreadsheets/... - เกม
![บุษบา - เมนทอล [Official MV]](http://i.ytimg.com/vi/IGnWPokSEis/mqdefault.jpg)
The best way I found to level shooting is to take 2 pawns that dislike each other put a shield belt on one and a pistol on the other then have one shoot the other.
Dammm.... I want to say "Spot the Rimworld player" but it feels a little redundant.
Why not shoot alpacas or deer or anything with 0% revenge chance? Really great for training and you get food from it uwu
@@stormtruppen4039 have to wait for em to respawn, wildlife on one map is good for training 1 or 2 people but with trying to level a whole colony it's realllly slow. i do hunting trips. take same wood, bedrolls and meds and shitty autopistols. form caravan with all colonists. settle one tile over. put down a small wood room with bedrolls and medical sleeping spots. equip awful autopistols but keep a sniper up for safety. limit allowed area to 1 quarter of the map at a time (so if someone gets in a bad spot they can be helped) set all weaker wildlife to hunt. let it roll. deal with bigger wildlife manually. might take a few days so stay the night. form caravan, move over another tile and do it again. if you have the spare settlement spots you can leave the settlements there for the future and do it every other week. it's really hard to maintain level 19/20 without two flame passion or great learner/great memory. the skill decay is just too much. maintaining level 16 is usually my goal, or 18 for 1 flame
@@greenhound just use a smoke or emp launcher
@@greenhound you might wanna check out the tents mod, it has teepees and tents you can carry like bedrolls, great for hunting trips
Not directly related but I figured I'd talk about a few relevant mechanics/clear up some myths I commonly see when people talk about Rimworld combat:
Figuring out what body part is hit when a pawn is hit by a gunshot, is more or less just totally random, it doesn't have anything to do with the gun's 'accuracy' (a hit is a hit, a miss is a miss, the body part hit is calculated the same on any hit. Basically a Bolt Action Rifle doesn't hit the torso/head more often then say a Machine Pistol, it just hits more often in general but the ratio of body parts hit is the same). The actual mechanics are such that all body parts are 'connected' to the torso (ie the neck is directly connected, and the head through the neck) and there are chances for each connection and it sort of goes out like a tree. But in practice this more or less means that it's just totally random what body part is hit (although as you might expect, it's a VERY high chance to hit the torso compared to most body parts). And this means that technically there are separate chances to hit the brain, head, skull ect. However, something to note is that if an internal body part is hit, all the parts from the inside out will ALSO get hit. For example if the game rolls that your Charge Rifle shot hit the brain, it will deal 15 damage to the head, 15 damage to the skull, and then 15 damage to the targets brain (which will result in death ofc, but with a lower damage shot just brain damage). This also often happens with vital organs in the torso, where if the heart is 'targeted' the torso also gets hit for the same damage. So in terms of 'total damage' dealt, you take less damage if an 'outer' body part like the head or torso is hit vs an inner organ. However in most cases, this distinction is only relevant in the amount of bleeding/pain caused (due to multiple injuries), as for example damage to the head or torso actually has no negative stat effects unless the body part is destroyed outright (well the pain from the injuries will cause some stat debuffs).
Either way in practice this doesn't matter much, but it is relevant to note that weapons that put a lot of bullets into the air in a given time frame (Machine Pistol is notable culprit, but any 'automatic' weapon often does this, it's just the Machine Pistol doesn't do enough damage to destroy the Brain, so instead of just killing the pawn by destroying the brain like a stronger weapon would it commonly leaves them with brain damage) actually have a much higher chance of causing brain damage on average, as the chance is effectively 'per hit'. Explosions also tend to do damage multiple times to multiple body parts, and since they deal Blunt damage effectively partly ignore armour as well, so they can often cause brain damage (not very surprising from a realism POV). Arpen of course can factor in if using a good helmet, but even cataphract helmets vs the worst weapon will still only have about 60-70% of completely deflecting the shot, and while it will always reduce it to half vs a weak weapon, even 'half' brain damage is still pretty bad.
On the subject of downing, while I think almost everyone already knows this (and is mentioned in the video), basically whenever a NON player controlled pawn is downed (I THINK prisoners are included as player owned and are 'safe' from this but not 100% sure), the game rolls a chance for them to instantly die. This chance is based on your population, so the higher your pop the more enemies just magically die for no reason. For non-player animals I think it is just a flat 50% chance of death on down. This is why you often see enemies with no clear causes of death that are dead. This also is how a 'Wimp' trader can die in a single mad rat bite, as any combat related 'downing' counts even if it isn't by the player or their animals. As some have already mentioned, heatstroke and other ways to 'pass out' without getting directly hurt will bypass this chance, so using 'ovens' to down people can be effective as a reliable capture method.
Speaking of downing, there are basically three methods to down a target; lower their 'movement' stat below 15% (or it might be 10% I forget), lower consciousness below I think 10 or 15% (there are some caveats to this, as low but not 0% consciousness can cause death, more on this later) or cause enough pain that they go into pain shock. Note that lowering consciousness directly lowers movement so in practice lowering either one is the same anyways (but consciousness affects other stats too, although notably NOT sight). But in terms of combat, this basically means either damaging the brain, destroying/heavily damaging the legs, or just doing a lot of painful damage. This is also why pawns on yayo/gojuice, which basically eliminates pain, can seem almost invincible, as they can effectively ONLY be downed by either outright killing them or destroying their legs (and due to the random hit rules I mentioned before, this can be a massive crap shoot). On the subject of pain, burns cause triple the pain, so in theory using fire is the best way to down people, although in practice most fire based weapons are so inaccurate that this isn't very practical.
And since I'm talking about downing, might as well talk about death. Pawns die when their consciousness is reduced below 10% (this is why people can sometimes die from a 9.1 damage brain hit), or any 'metabolic' stat is reduced to 0%, or some vital body part is destroyed. For humans this basically means if the Torso, Neck, Brain, Head/Skull, Heart, Liver, or BOTH Lungs or Kidneys are destroyed, they die (ofc as I said NPC/enemy pawns can magically die from being downed as well). Stomach used to be included in this list until v1.1. This is why Flak Vests/Helmets are so important, as they protect all of the vital body parts (vests protect the neck as well as the whole torso), while Flat Jackets/Pants are kind of optional (and honestly Dusters made out of good materials are often as good or even better anyways). These rules are again why a yayo/go-juice infused pawn can be so damn hard to kill as you need damage to very specific body parts. For Mechanoids, they basically die whenever they are 'downed' from having their movement lowered too much (or a vital part destroyed, but this only usually happens with the weaker mech types). This is why Centipedes often die without having any destroyed body parts (basically almost every body part you would commonly hit on a Centipede affects it's movement, so you just need to hit it enough times to cause it to be 'downed' and then it just dies.)
There is a caveat to 'consciousness below 10%' thing however; things that set consciousness 'max' to below 10% are fine (which is why sedation doesn't kill pawns), it's actually only direct penalties that count. This makes stuff like smokeleaf combined with early stages of heat stroke/hypothermia/blood loss/drug overdose or stuff like joywires/circadian halfcyclers VERY dangerous. In fact, because 'lower' levels of heatstroke/hypothermia/pain/blood loss/overdose subtract consciousness while higher levels set the max to such that you automatically pass out (and setting to max obviously doesn't stack with other penalties, it just caps the max), this can actually mean that in some cases having LESS pain or partly recovering from heat stroke/hypothermia/blood loss can actually cause a pawn to go from passed out to dead, if stacked with enough other subtractive penalties. This is why pawns can sometimes die from smoking smokeleaf, or from a partial drug overdose stacked with partial heatstroke + a bunch of pain from injuries + partial blood loss and either something like brain damage, or a joywire/circadian half-cycler that lowers consciousness. Ironically in those cases you are safer having 'worse' tiers of afflictions as the worse tiers of those debuffs just cap consciousness instead of subtracting it, and trigger passing out instead of death. I have straight up had a pawn die from 'recovering' from major heat stroke when they had partial blood loss and a ton of burns and brain damage (was a very messy fight vs fire using enemies indoors, so that is where the combination of effects came from).
Also a final 'myth' to address. A lot of people mention stuff like 'use maces to down people instead of longswords because bruises are less likely to kill then cuts'. This is completely false. Bruises and cuts basically only differ in the sense that cuts cause bleeding (and bleeding injuries can become infected). A 'bruise' can still easily destroy a vital organ/body part, as the damage value matters, not the damage type. Now it is true that maces do less damage then longswords, so it is less likely to deal enough damage to outright destroy a body part, but at that point you are better off using any crappy melee weapon (knife/club) then a mace. Also, because I see this commonly brought up, Plasteel Longswords actually do better then Uranium Maces vs Centipedes, due to the Longsword having very high armour pen (meaning that the effective armour of the Centipede vs it, is actually very low), and WAY higher raw DPS, which more then compensates for the lower blunt armour of the Centipede (well more specifically it depends on the quality since melee arpen scales with quality and maces are already 'overcapped' on arpen even at low qualities vs Centipedes, but if you are talking about Excellent or better it's a win for the Longsword and approx a tie at Good). Not that it really matters these days because with Royalty Monoswords beat the pants off everything else anyways as they basically ignore armour and do insane raw DPS.
summary?
Wow, the longest meaningful YT comment I've seen. Unexpected. Interesting, though.
Super insightful.
Felt like I read a thesis
Hey I wanted to thank you for the in-depth reply. I'm at about 100 hours and know there's still a ton I haven't learned, so thanks for all the information. If you have anywhere you can point me for more, or if you decide to work more closely with Francis, that would be amazing. Your knowledge seems to be at the top tier and I'm sure plenty of people would love for you to inform. Rimworld has a serious lack of in depth resources, Francis is great, but I know he's still learning himself. If you feel like making a video, or working with Francis on one, the community would be better for it. Cheers.
Going to subscribe to you, in case you decide to post some Rimworld content
Locust: "This sniper rifle is so awful. The long range scope misplaced by 30° degrees. I can't hit anything."
White: "Well, that's the best way to train shooting skill."
when an enemy pawn is downed by fighting (none fatal dmg). The game will rolle for instantly killing them, if the enemy pawn is downed by bloodloss/hypothermia/heatstroke etc they will skip the rolle
I think the first enemy that attacks your colony skips the roll as well.
Did they fix the infinite heat generating door-inside-wall trick yet? Used to be you could down entire raids with some creative placements of perma-2k temperature permeating into enclosed rooms. Sadly mechanoids have no upper temperature limit, so this was only good as a pawn recruitment aid rather than something optimal for base defense more generally.
A good budget option is just to smack a desired recruit with a knife a bunch then disengage until they get downed from bleeding, or to do similar with pistols/machine pistols and stop early so they don't take the down check.
That was probably throwing off his numbers, I use Custom Death Randomness to get rid of that.
Roll*
@@TheMelnTeam check out the rimworld strategy subreddit, ive posted that thing there
You can click the stat in that window (2:40) and it'll freeze the right pane, btw.
Thanks, going to have to remember to take advantage of that.
Thanks Francis, for putting in the time and effort to bring us this Rim-Nugget. (needs a better word, I know)
cheers mate.
I suspect that the "magic number" for damage is anything less than 10.
Basically most organs have their hp as multiples of 10. For example torso has 40. So a weapon with 10+ damage hits torso 4 times the target is gone. But if a weapon has 9 damage and hits torso 4 times, the target is still alive, but in intense pain, likely downed.
the most important thing about it is that you don't instantly kill with a hit to vital organs.
honestly it's useless to worry too much about it because you don't have control over enemy armor, a heavy smg with 12 base damage is very unlikely to actually hit for 12 damage reliably so it's more feasible than using a machine pistol that will mostly deflect on any enemy with flak armor
Even when this comment must be old. Here is some INFO
- Once a pawn gets enough dmg to be considered downed, the next impact of damage will roll the % you have to get pawns killed.
- This is resolved when you create the game difficulty or you choose a value with the special difficulty options.
- The machine pistol puts a lot of bleeds on the target lowering the chance the pawn will go down due to damage but rather goes down due to pain shock or bleeding affecting the consciousness.
- Once the consciousness drops below 35% i believe the pawn goes into a downstate without rolling the % to die factor.
So bleeding or painshock are the 2 methods to help preventing death from a pawn you really wanna add to your colony. I personally use bleeding.
I would love for you to update some of your Rimworld tutorials for the newer updates. These have been insanely helpful. Im just getting started in the game and stuff like this makes the learning curve so much easier to handle. Great work.
Alright crafter with a skill of 5, now is your time to shine. Get in there and make me a sniper rifle.
pawn has produced excellent sniper rifle... classic randy
A useful tip for showing stats: click on a stat to focus on it, so you don't need to keep your cursor over it!
I like how Dead Alpacas became a scale.
i'm dying
I love your Rimworld Vids! Thank You
low damage but lots of wounds add up to a lot of pain, so they go down faster without much blood loss, so the roll on them dying is lower due to the lower blood loss. Or at least that is how I understand the machine pistol's efficency at capture
if you're trying to recruit blood loss is actually the safer way, enemies get an arbitrary dice roll to die instantly ignoring actual health status the second they are downed by an attack
if they collapse due to blood loss, heat stroke and pain etc. that builds up over time (like from disease) they never randomly die.
But this involves stopping to attack that pawn
the difference is a 6 damage hit that hits an unarmored enemy can deal a maximum of 6 damage to vital body parts like heart, brain, liver etc. which have 10HP each so it takes 2 incredibly unlucky hits to the same vital while a heavy smg would instantly kill
you're correct though that the pain collapse itself is favorable when you're simply attacking until the enemy is downed because it can be tricky to capture and heal them in time
If the only thing that matters is time spent shooting, there should be an absolute amount of time spent shooting to get shooting skill from 1 to 20. Will have to investigate next time I play. I really like how you do the tests and share the data! Cheers!
I never tested trigger happy, I'm going to have to revisit that at some point.
12:29
"And then i fired, and then i missed, and then i fired and missed again. And then i got sad, i had a simple meal, and then i passed out in the snow. And then i woke up, and then i fired, and then i missed..."
Interesting test results, Mr Mengele.
That's DOCTOR Mengele. He didn't go through 8 years of college and 6500 victims to be called "Mister," thank you very much.
@@christopherdobney2927 thats from phenias and Ferb right? Doctor ninja balgeet
@@vitaestludum I thought it was from Austin Powers
OH NO XD
These tests show two things about statistics:
1. Small sample sizes are easily misleading, as the average of one test can be very different from another.
2. Large sample sizes can deliver a steady average... which can also be wildly misleading, if you don't mention the _variability_ (when was the last time you saw _that_ mentioned in advertising?).
It also shows that Rimworld's "simulation" of shooting is incredibly bad (it literally applies a randomised modifier to your pawns' "chance of missing", rather than their _accuracy,_ so a pawn with bad shooting firing a minigun one metre from a barn manages to magically "miss" 90% of the shots, which would take _more_ skill than hitting half of them).
Therefore you have things like Variance, Standard error etc :)
It tells you exactly what you want.
Of course, large sample size requiered, but that is standard in anything close to statistics
@@patricktho6546 - "You have" if the source supplies it. As mentioned in #2 above.
My thought on misses sort of “going out of their way” to happen was that low skill shooters did not take recoil into account and started flailing the barrel around while firing
The reason why the legendary machine pistol does that more effectively, its because it also increases the % to land the hits on the target.
Increasing the number of wounds = more bleed = more pain shock.
Thanks for putting this out. I have to say, though, I don't understand how the hunting animals simulation could be valid without hundreds of simulations. A lucky headshot, unlucky miss, accidental hit, leg injury that keeps the animal from moving out of range, or a quick down skews the results dramatically. There are just too many factors at play.
For example, the pattern of animal movement determines whether/how far the pawns need to move to get in range and this affects things like the accuracy, efficiency of bursts, and chance for an accidental hit. You can see that sometimes the animals bunch together right in front of several pawns, and sometimes they disperse. This movement in turn is affected by how close any individual pawn got to an animal, which is affected by weapon range and RNG like how close the individual pawn decided to get before shooting.
Also, it appears that the pawns hold their fire when there is a pawn between them and the target, meaning that some pawns are either firing less OR firing more (not sure if holding fire counts towards XP or not) and the animal is surviving longer, providing more opportunity to shoot by moving into a low-accuracy range or less opportunity to shoot by into a high-accuracy range, or a combination of those factors.
Also, running the simulation in a box dramatically improves the efficiency of the hunting and therefore the amount of time shooting vs. walking: on an actual map a hunter with a short range weapon has to move much more between shots. You can see the effect here in the simulation when you went from small to medium sized boxes, but it's much more pronounced on a large map. A pawn with a shotgun is going to spend most of the time walking; a pawn with a bolt action will spend more time shooting.
Thank you, I appreciate these videos a lot
My experience has been that the machine pistol causes so many wounds due to small damage, that bloodless KOs between bursts happens more often, causing the better down rate.
I don't know if its been patched out of the game since I saw it initially, but there used to be a way to target a sleeping spot with a drafted colonist to force shooting practice. Since it has no HP it was never destroyed and allowed shooting to the daily cap.
Since the experience was calculated only on shot, and not on hit micromanaging the pawn to always fire at max range and at downed/ covered targets(lowest total hit%) you can milk out some more shooting exp as well.
I don't know the code, but a lot of variability in your testing probably has to do with how the damage is calculated. a good sniper hit on a vital will likely drop an alpaca, where if it hits a non-vital it may incapacitate or even leave them unaffected (ear, eye, finger etc). additionally there are likely breakpoints with the area struck: going from awful to poor may take a torso shot from 2 to 1 hit kill, or leave the target wounded enough to bleed out before the next shot connects.
if shooting at non-living targets still awards exp, you may be able to get more consistent results by having the colonists drafted and shooting at sections of walls or furniture (anything with High HP) at max range.
Again, love the coverage. keep it up.
If you are going to send out a pawn to train with a terrible sniper rifle, I would advise giving them a combat-trained pet to keep them company. That should help protect them against animal revenge and can nuzzle them to give them a mood boost.
Always hunt in packs. Turtles are awesome guards!
Wow, great! Just what I wanted to know.
Totally ethical weapons performance testing
While testing, were the enemy pawns' traits checked for ones such as Tough or Wimp? I've noticed with my playthroughs, pawns with a lower pain threshold will survive more often when downed. Clothing such as the War Veil, as mentioned in another comment, also alter the pain threshold. Awesome content as always, keep it up mate.
I think the reason you had trouble getting repeatable results is that your sample size was still on the small side: you need a bigger sample size to capture the frequency of something with a smaller chance, so for something like downing pawns you'll need the larger sample size for the results to have a somewhat accurate mean.
Your Awesome dude
More Rimworld! More Rimworld!
You can click on the stats in the bio window in order to keep them up in the window on the right.
Cool stuff
Looks good!
Out of everything in this video I’m just surprised the incendiary launcher has such low damage
I try to always keep a machine pistol in awful quality for leveling up shooting. Once a centepede is downed, draft and have the colonist stand at max range and fire away. Sure, it will hit the centipede every once in a while, but the damage is so pathetic compared to the HP that it lasts for a good week or so of training.
Can also do this with a downed thrumbo, and have a second pawn doctor it up (again, make sure you're at max range to avoid friendly fire, or use a shield). This requires a little bit of micromanagement and finesse. You do run the risk of killing the thrumbo regardless from blood loss, but you also get the chance of the doctor bonding with the thrumbo to sell/slaughter later.
They have since introduced the EMP cannon, you can use it to target animals and gain shooting skill without every killing the animal. You can even target your own animals with it.
my first working theory would be that the machine pistol causes a larger # of wounds. The increased pain from these wounds downs them -- whereas a rifle would have killed them in so many wounds. many times I see downed enemies they have the mind-shattering pain debuff. Also did you standardize the opponents that the pawns were shooting? Anyway, excellent video and thanks so much for sharing! super cool, the experiments you came up with and helpful info!
I have a theory/guess: the amount of exp doesn't depend on the number of shootings, but the amount of aiming time. So the slow aiming weapon with a single shot like snipe rifle out performances the quick aiming weapon with multiple shots like minigun
*@Francis John*
6:31 Not exactly, it has to do with how many times you get to pull the trigger per X time.
So you want the weapon with minimal: cool-down, charge-up AND also minimal burst size too. Since it doesn't count the bullets at all, but only the trigger pull.
I think you don't even get extra XP for the kills? So you can literally shoot at a wall or something to train the shooting skill. I think regular *single shot pistol is best.*
Note, I think the pistol would be better per time, not per animal hunted.
You can test it with that "walking timer pawn" in the first "downed" tests ( eg: 0:05 ).
I'm afraid it's not down to trigger pulls or the auto pistol would have dominated being the most rapid fire weapon. While the Exp is only applied when the trigger is pulled the amount of exp you gain is directly related to rate of fire. The faster the rate of fire the less exp per pull, the slower the rate of fire the more exp.
from what i remember also, shooting at walls does nothing to the stat, and melee is also the same, so hunting would be better or maybe the firing squad for prisoners with weak ass weapons while prisoners are placed behind cover
One thing to note is that with automatic hunting, pawns using the sniper rifle in practice just dont get to take as many shots because animals accidently move out of range while they are charging up that atrocious 3.5 second aiming time.
Its fine if u recruit the pawns and use optimal movement to avoid that, but it gets so tedious. I still think sniper rifles are useful for hunting dangerous animals with good enough shooters, since u can run back to base if u get manhunter notification
I have since learned in Royalty you can shoot at your own animals with an EMP or Smoke launcher and train your shooting skill that way. Makes things so much easier.
@@FrancisJohnYT thats kinda cheap lol. I didnt know that launchers leveled up skill
When it comes to capturing pawns, the best way is to set them on fire, they will pass out from pain caused by a ton of burns and have (nearly) no chance of dying. The burns will also completely heal without any permanent damage.
Burns can leave scars
@@dylan__dog Only if you are to slow to rescue them, burns can destroy eyes, fingers, toes, and noses. Which can not be brought back very easily, however bullets can do the same thing. As long as your fast to rescue the person they will recover.
@@crunch.dot.73 how fast you rescue doesn't matter, it's RNG whether a wound leaves a scar or not, burns can leave nasty scars, torso scars especially hurt since you have to use healer serum on them
@@dylan__dog I assumed that burn scars worked the same way as normal scars, my bad
@@crunch.dot.73 they do, it's just like scars from shots and lacerations except burns have a tendency to leave scars more often from my experience, often quite painful ones due to burns generally giving more pain for the same amount of damage
im new to rimworld but the downing mechanics from my personal experience depends on pain threshold just like some animals have higher and lower stats for max pain b4 downed.
which leads to the conclusion that scatter shots that are more likely to damage alot of body parts but low enough damage to not shoot limbs off, meaning MLG and AP that have a lot of scatter shots that the damage is devided between more organs and limbs and each one of those has a pain % once most hit 80% they go down or at 2-3 H bleed 2 death
Just a heads up. My sister in law works in statistics. And according to her. 300 tests are needed for an average
For a sigma 5 I would need 3.5 Billion, but I'm not sure my computer could handle that ;)
It'd be quite cool to have a mod which lets you tell your pawns whether to shoot to kill or to down. I would imagine only pawns at or above a certain skill level would be able to do it, and the higher the skill, the higher the magnitude of the effect that changing modes would have on determining which body part is hit. Could also introduce an increased aiming time for when pawns are told to shoot to kill (harder to hit a head or an organ purposefully than it is to shoot larger body parts or at random) that decreases with skill level. It would be realistic if medical/animals/intellectual skill also had an effect on this since choosing a target like that requires some knowledge of anatomy (medical for human targets, animals for, well, animal targets--though could see medical just applying to both as it applies to veterinary medicine--and intellectual for mechanoids), but this might be getting a bit too complex at that point. I love mods like Combat Extended and the one that lets you set defensive positions, they make the game's combat more tactically interesting and easier to play without a killbox.
Lower armor pen is probably very important for capturing
So when and what will be the next playthrough? Don't get me wrong, I do enjoy testing but a let's play is a class of its own.
Lmao Been playing so much rimworld I just hit 3 trying to speed this video up
So I'd say maybe you want a shooting range with an alpaca at maximum sniper range, with a barricade, in the dark, preferably asleep on a bed so they are lying down that way your pawn has like zero chance of hitting them and can spend all night shooting at the target and missing but getting skill in shooting. If you are saying it takes 11 of them to level up to maximum I reckon barricade, distance and darkness would take that down to 2-4.
Been a while since the non CE rim called. But when there were no ores i replazed the miners with shoters. Gave them low damage single shoot guns with low cool down and carve the mountain with bullets.
Force targeting does not give exp any more, from what i can tell you don't get exp from shooting walls or even downed people or animals. Has to be a living non disabled target. (Assuming mechanoids can be classified as alive)
I think it's basically like gambling. If you use a high damage weapon and you shoot them in the head, you will kill them. However, if the same happens to their leg, he will be downed and only have one wound.
If you have a low damage weapon, you have more chance of letting the opponent survive, however you need to shoot more. It basically ends up with the same chance.
It's like gambling: playing a game with low success chance once is just as bad as playing a high success chance game a lot.
I think the best weapons to down people is a low accuracy, high damage weapon.
Francis, first off, big fan of your Oxygen and Rimworlding!
Now I was curious, have you done testing for the LATER levels? Say level 15? I imagine since the "Missing" factor is so key which is why you choose the sniper for early levels. I'm wondering if when you reach a certain threshold if perhaps a weak bow might actually surpass the sniper in the later levels due to damage and penetration ability?
Thoughts?
Does it mean you have to hunt on small animals to miss even more often?
Low penetration weapons turn half.of the damage into blunt/non.bleeding more often tha not. in terms of downing a non-drugged opponent its not only the damage of the weapon before armor,but the actual final damage inflicted at the body parts themselves. (Armor pen is god :) )
If you check the consciousness of the pawns downed by the machine pistol you will see its around 35
A little curious for why no bows were tested. As far as lvling shooting goes getting an awful bow or greatbow should work quite well, especially when you aren't concerned about raw output.
Does the Trigger-Happy / Careful Shooter trait affect the experience gain? Is a shot from a Sniper Rifle by a Careful Shooter worth more experience than a shot from a Sniper Rifle by a Trigger-Happy shooter?
Also, while you're clearing up Rimworld myths, I've seen contradictory data on the internet about which weapon are better for either trait. Might be a subject for another video, if you're interested?
Question are you going to do a similar one with the bows?
A question for code readers watching this video for some reason: Does the game calculate the shooting XP gain, or is it just a fixed value?
If it's actually calculating it: does it use both the aim and cooldown time?
I forget who's video it was, but they said they get xp per shot fired, it didn't matter if they hit or miss or got deflected. And there's a (daily?) cap on xp
@@mikebytheway The daily cap on XP doesn't stop the XP gain but it does decrease any further XP to 10%
Could the machine pistols lack of deadliness be related to the burst function or maybe armor penetration?
By the end I'm not sure it is less deadly, a sample size of several hundred may not have been large enough.
@@FrancisJohnYT Maybe they should shoot at naked targets, so there are no armor from clothing that can mess up the results?
( Why does it always looks so "war crime'y" when talking about Rimworld mechanics? ) :D
Did each pawn have the same gear every time? That’s an important control I’m not sure you mentioned
There is lot of body parts which have 10 HP so 9 DMG means damaged but not killed. Seems fair.
If you click on the damage, then the information to the right will stay there.
Here's my explanation for why the machine pistol downs enemies. Now I've only watched up to 3:15 so if he comes to the same conclusion I'll delete this. Machine pistols only have 9 armor penetration (for comparison the auto pistol has 15) and if the armor of your enemy blocks the bullet the damage is converted to blunt damage. (This is to stop stacking layers of sharp resistant armor) So with it's super low armor penetration it's very easily blocked (Essentially 109% sharp armor is all thats needed to convert 100% of the damage too blunt) more often than not it's dealing blunt damage like a mace and that's what's downing them.
No wonder my bolt action pawns level up like crazy
It’s pretty clear there are so many random factors that you need a VERY large sample size to determine odds of downing a target without killing it.
That said, the only real mechanic that’s going to affect the odds is instant death due to destruction of a vital body part. If there’s no death due to destroyed head, torso, brain, heart, it’s just the pain vs. total damage progression, which doesn’t care about weapon type at all.
Mostly that’s about the raw damage per hit. Minimum threshold’s about 15 damage, since the heart has 15 HP. Heart shots are pretty rare, though, so it mostly comes into play when you’ve got a weapon that totally vaporizes the head with a single hit.
Keep in mind the game randomly rolls for instant death when a character is downed due to sustaining damage in a fight, so while using lower-damage weapons like the machine pistol will make it more likely that enemies will be downed instead of dying, you're still gonna see a lot of instant deaths.
Shooting training for my pawns looks like this:
wall in a critter that wont attack back, remove the eyes of a character you want to train (store then in a fridge) send in your eyeless pawn, put in eye again when hes a god
what am i doing with my life
Stuff you only hear in Rimworld.
I tip my human leather hat to you.
I think you need to remove the outliers in the tests to get a better average. Remove the whole row.
Small box, dump an animal in and give the dupe(!) a smokelauncher... in the end the animal will love the dupe and have smokerslung..
Is there a de-buff if you're shooting smaller animals like squirells or rats? It might be a better way to train.
Yes smaller animals are harder to hit. I currently use a bit of an exploit by shooting my own animals with an EPM or Smoke launcher. Trian the shooting skill with no need for enemies.
Student's T would be interesting here.
When some dude start roofing (@7:52 the dude on the second to last box just gives up completely).
What down setting do you have on for this video?
Ah, yes, game science
Does the target affiliation matter? I've heard hostile targets grant more XP per shooting cycle, so do wild neutral Animals give more XP than colony Animals when targeted with a Smoke/EMP Launcher?
Don't know about wild vs domesticated animals, but worst case you can use an EMP launcher against wild animals.
@@FrancisJohnYT Also, I know the Video is a little bit old, but I'm surprised no one in the comment seems to have brought blindfolds for an effective -9.6 Levels in shooting instead of spending resources on awful guns.
I was wondering ehy my charged pawn is lagging so bad behind lol
Charge rifles tend to make a mess of things rather quickly.
This is also without a passion so they learn 35 percent of the way which means a fast learner with a huge passion is just so much more worth it to train just cause you dont have to spend hours training it up one level >.>
If you can get them with a passion it is great. For testing here I got pawns without passion. I used passion pawns at first but they would gain skill so fast that they would max out for the day which then nerfed their skill gain and mess with the testing.
10:34 "Snifer Riple"
A Snifer Riple, like a regular sniper rifle but it operates on smell instead of sight.
need much, much larger sample sizes. you're rolling a dice a few times to make a conclusion
The general trend is the less you hit the target, the more shots you take the more experience you gain. So a poor quality sniper rifle or bolt action look to be the best. The sample sizes were to small and I would have liked to have more data to work with but it took many hours to get what I got. If you want to do more extensive testing the save game files are attached.
You're experiencing too low sample size for the variance in your experimental design. With proper statistical testing, you would never be able to reject your null hypotheses. In other words, you can't draw conclusions from your current experiments. It is however a good start, and I'd love to see you take it the next step forward!
When the accuracy is so low I would need to run literally hundreds of tests to account for variance correctly. The hours required are beyond me. With the Bolt action and Sniper scoring so high so consistently I am going to stick with them for training.
Do the "shooting the sleeping spot" trick still works for levelling up shooting skills?
Not as far as I'm aware, but you can shoot animals with a smoke launcher or EMP launcher and that gains you points. Can even be your own animals.
@@FrancisJohnYT
Can it be done with an EMP grenade?
@@remliqa Grenades don't train shooting last I checked.
Machine pistols are best for downing enemies cause its low dmg, it tends to cause way more wounds on enemies making them drop from pain, cause normal uncooked raiders drop only when thier legs are too damaged or incapacited by pain (15% movement), machine pistol has low accuracy and dmg so it spreads damage over raider body downing them from pain.
Low accuracy is most important cause highly accurate guns and skiled pawns tend to go for torso and head, i had moded gun with high accuracy and only 2 dmg (Bulletstorm from glittertech mod) and it was causing a lot of brain injuries and death.
Also I belive that your population has inpact on chance for raiders to be downed( cause game rolls death chance on raiders when they go down), I incresed maxDesired population to 30 and bumped tresholds up, and I'm getting more downed raiders despite using asalut rifles.
This may be the cause for people saying that machine pistol is so much better for capturing pawns, cause it is low tech early game weapon.
Barr! Barr! Barr! Barr! Barr! Barr! Barr! Barr! Barr! Barr! Barr! Barr! Barr! Barr! Barr! Barr! Barr! Barr! .... barr.
I want all my guys in trench coats and helmets like that!! How do I do it? Do I need mods?
I want a Wehrmacht style colony 😍
The trench coats are dusters, the white colour is because they are made of thrombofur. I'm pretty sure the helmets are just marine helmets.
Just disable critters from getting stored and hunt them, almost no effort for big gain.
Here is a great way to train pawns shooting AND constructing (repair) that I (Sion) invented WAAAY back in 2014: ludeon.com/forums/index.php?topic=3169.0
Unfortunately most of the pictures have been lost due to time, but there is *still one picture left on page 2:* ludeon.com/forums/index.php?topic=3169.15
I havn't tested if it still works, I remember a glitch back then where bullets could go through walls, but I think that's fixed now (or it will train medic too, use shield belt) ;P
shouldn't the experience gain be given out when they actually hit a target? Makes more sense to me, this game is weird.
so we should give an awful sniper/bolt action rifle to our hunters and tell them to hunt hares? because you know he'll be shooting all day long to hit it
That's a lot of dead alpacas :(
We needed some sort of metric to measure by.
you realize this is less a tutorial than it is you going through your experiments almost live. I love it, but it's kinda misnamed
You could be right, I gave the title a little tweak and called it testing. Though it's not close to a live version, that would be many many hours long. Next up is medical leveling and testing the earliest surgery setup you can get away with. The issue is the failure rate can be as low as 2% so I have a map with 100 patients, 100 doctors, 100 rooms, 100 beds, 100 stacks of medicine and 100 queued surgery's. Even with Dev mod, blueprints and mods there is no way that construction project would end up on a video. Now I just need to test each scenario 10 times for a 1000 operation average.
@@FrancisJohnYT I mean if you want to level medical, the obvious answer is to just 'experiment' on prisoners ie attach/remove peg legs constantly. Or get mods for more creative 'experiments' on prisoners if you want to be even more evil. The non evil but inefficient method is 'rescuing' downed animals from hunting and euthanizing them for medical exp. As long as you aren't doing it in a dark/filthy room you won't fail often enough to matter. Frankly any sort of leveling on your own pawns is going to come with pointless risk. As for surgeries on your own pawns, with the recent change that failed surgery damage ignores worn armour, doing any sort of brain implants (like many of the Royalty ones) is insanely risky even with the best surgeon possible in the best conditions possible as ANY fail = death or massive brain damage which frankly really annoys me as it happens way more often then it really should and basically makes me think you should never install any brain implant unless it's 100% necessary.
you REALLY need to turn down the highs on your record equalizer.
Its really bad, its been bad for years.
Do you perhaps have a link to a tutorial on the subject? When it comes to recording/editing I'm very much an amateur as you can tell.
@@FrancisJohnYT try this theproaudiofiles.com/sibilance/ maybe here. Treble is the issue. Love your stuff though.
I like the current audio, but I'm also half deaf.
The high treble helps me differentiate the words better.
Cool test, but the sample size is just way too small to draw a meaningful conclusion. I don't expect you to want to repeat this a thousand times, but that's what you'd have to do in order to be able to say one weapon is better than another at downing pawns and not killing them or lobotomizing them.
Turns out their is a chance to down mechanic built into the game. It means most enemies just die instead of being downed. You can customize it now since the release of the Royalty DLC.
Maybe hunt on rats can increase output of training? Rats just replicant themselves by eating themselves and corps of enemies. It's like training dupes on manual generators i think, need some research :D
is it still true in 1.4 ?
Yes, you can also train shooting by shooting your animals with a smoke launcher that was introduced in version 1.1 it is slower than shooting enemies for exp gain but is totally safe.
Thank you, this will simplify slave trade and organ harvesting.
LOL! Rimworld is so morbid!
You hate alpacas, don't you?
I'm no statistician, but I think your sample size was too small to get any meaningful information.
At skills that low it is way to messy and a couple of lucky shots can really skew the numbers. But the theory and the data we did get seems to line up.
To train your shooting you want to do more of it and the best way to get that is to miss a lot. So get a really long range weapon with low DPS and awful quality.
Write a script and run larger sample sizes.
Don't know how to do that, is their a tutorial available on how to do it?
@@FrancisJohnYT The simplest solution I can think of is to use a scriptable autoclicker. What doesn't seem to be super simple is to analyze the output in an automated way (surely correlating pixels can't be efficient). Maybe save the game at set intervals to unique savefiles and then using grep on your *.xml savefiles, finally resulting in some raw data sheets which in turn could be analyzed using awk or something comparable. This could actually be fun :)
@@FrancisJohnYT There is probably no tutorial for this exact problem, but for all the tools in my first answer there is a near endless amount of material in the internet.
@@janniserhard2810 it's more time investment vs return. It would take so long to figure out how to automate it that it would be faster to do it manually.
@@FrancisJohnYT Depends on the sample size. If you would like to run a million batches, doing it by hand and retaining sanity might not be possible. In my experience it always pays off to invest the time to automate but that's porbably only true if acquiring data is your goal. If your goal is a youtube video, that might just be overkill. But consider, you could apply theese tools to all your benchmarks.
Animals that got mad give much more XP, I like to do drafted hunt to make it less risky.
For some reason I have soft spot for alpacas. So many of them fave their lives in defense of my colonies...
The Great Alpaca Charge of 5501 will be remembered!