On your second paragraph/idea. There is a lot of other ways to improve the a.i. and it could use a lot of work in that regard as all multiplayer coop games currently do=)
Your first paragraph regarding accuracy where you mention random headshots, and what is considered fair or jarring for the player I would mention that this is subjective and should first and foremost be fixed with suitable difficulty settings that can be voted on so that players can experience what they themselves regards as fair, fun or jarring.
Way too much speedrunning in my opinion as well, and it makes it unplayable because the game feels so outrageously easy for me subjectively and I have zero patience for easy games where death ratio is low and every headless chicken can work their way across the finish line=)
Also what do you mean by accuracy when you want to avoid random headshots from bots, you mean tunnel sighting on certain areas with cooldowns?
You seen this?:
EDIT: Not that it is anything wrong with lower difficulty, less headshotting bots or people being able to rush maps. It is just subjectively not what I prefer and a solid game should envelope that as well.
Andrew Spearin was very important in the development of Ins2014. He is currently "pursuing other opportunities". Look at vid if you want, he explains it better himself what his role was and it makes sense Sandstorm took a different path when he is not onboard:
You probably heard about LEAN production and the principles about asking "why?" until you expose root causes of problems.
Why is the game not succesful?
Because it still feels like a beta.
Why does it still feel like a beta?
Because the game is not finished.
why is the game not finished?
Because there has not been put in the work hours needed, miscalculated time schedule, difficult to adapt to a new engine etc
Now we get a lot of branches on our tree, and must decide what branch to follow. For example: Why has it not been enough working hours to finalize the product?
Because game devs are underpaid and those who complains gets fired, leading to less motivation, exhaustion and unrealistic time schedules.
Quote from Spearin:
"Public stories are the tip of the iceberg floating in NDA Ocean. The waters are frigid below the surface where hard working game devs find themselves drowning from crunch, harassment, burnout and depression."
One of the root causes I would guess;)
Depends if its player controlled or enforced by server. I agree that it is sometimes serves a purpose to keep on deadchat, in other case some random guy starts backseat driving telling the guy what to do and shouting commands so it drowns out all enemy sounds, making it impossible to gain intel on enemy presence, which is really important for me anyway. I prefer teammates being able to write text or use a discord as you mention.
So is the inconsisty made on purpose to create variety in "bot difficulty" or something else. Personally and subjectively I would prefer them all being elites. Bots are most often easy to make counter tactics against also at higher difficulties, but if players meet the slow bots and the elite bots in a variety, it may elude players in to a false sense of control when experiencing dominating bots in some cases and being owned in other cases. Thanks for your perspective, Im glad someone can say speedrunning is actually a thing here.
I know a lot of people see it like you because difficulty is subjective. I did not need to even watch the whole video because I know the game is quite solid already, I can trust the game being itself without needing to cherrypick a video for demonstration. Cherrypicking is a bad thing remember. Imagine a difficulty scenario ( a situation u can meet in the game) that is impossible for 100% of gamers. All the gamers will agree that it is impossible. Now imagine a difficulty scenario (still a situation that u can meet in the game) that is impossible for 80% of gamers. Now most people will agree with u that it is bullshit and impossible, just a few gamers here and there (20% for example) will tell you the game is fine. Now most people will agree with you and maybe say this guy have no idea what he talks about. Sounds familiar? ( No answer needed).
I made a link to the same video where we discussed the death at 17:03 and will go as far as saying there is not even close to being a unfair death in this situation from my subjective point of view, because the death could have been avoided. The player had room to improve in mutiple ways. I rewatched the death again more closely because I thought maybe I had overseen something because of your reply, but no, not in this case. You can slow the video down in youtube settings if the kill feels instant;)
Lets walk through the situation one more time:
at 16:58 the bot reveals itself by opening the door
at 16:59 he attempts to shot the bot first time.
at 17:01 he sees the bot for the second time but hesitates and does not shot
at 17:02 he shots at the same bot a second time but misses
at 17:02 he looks away from the bot to his left again.
at 17:03 he gets killed by the bot. The same bot he first saw at 16:58
All the time he stays at the same position. He has made a lot of noise at 16:59 and the bots are aware of his position. His choice to stay still is relying on luck.
He have had 3 chances to kill the bot but failed each time. This is not an instakill by any definition, this one actually extends for approx 5 seconds which is a lot for this game in many cases.
Ideally it should be more complex bot behaviour and not only superaim that is used to increase difficulty - If NWI don't have the time and/or human resources to do that, they could consider the methods used by modder of Bef server in ins 2014. I never went to the download page myself, because i stumbled upon this server in the server list ingame which download the files when you just join. If this link is not for the whole mod, they could contact Artos, which I believe knows more since he is listed as author of said page:
Because super aim in itself will make all rushing players die very fast, so unless having a super organized team that watch all directions from inside a relatively small safe perimeter the team will fail - This I have experienced being the actual result for hundreds of hours on said server. Combining this with other adjustments like the bots not spawning too close to players etc, could at least make a option for higher difficulty that a certain type of players will enjoy.
The new enemy that bef server introduced who had shield was also pretty cool. You always played as insurgents on this server, so some enemies having a shield was somewhat more believable. Its also believable that it should be more difficult to play as insurgents as in most real world scenarios they are the underdog.
We think very differently, I will try to leave out all text that is just noise in that regard and stay on the topics. For me this is not a battle.
What I have said about ue4 and programming here was trying to tell u about principles about adjusting parameters that is common in games without in any way trying to pretend I am sure exactly how NWI have solved it, just probabilities.
I said the a.i. is bad in the gaming industry in general and that devs use godlike aim on the bots to compensate - I don't have a problem with that, both in Ins2014 and in Sandstorm. What I have said is that I didn't want the devs to lower the difficulty because it [the difficulty] is not bugged from my point of view, it is merely the devs doing a good job compensating in the current situation we have with stupid a.i. in the whole gaming industry. I think they do that in a fun way, that is why I replied to the post by kanyo to begin with. I think very good aim from the bots is a necessity on higher difficulties.
Did not watch the whole video myself, just skimmed through some parts of it. Like I said watching the whole thing would be boring for me. I looked at the highlight u mentioned at 17mins in the sandstorm video and I find no bugs in the way the player dies - And that is the core of where I was going with my reply to begin with. The player looks at the door where his teammate died being open (blind) from both right side and behind, he then see a bot at his right side and goes for a shot but fails to kill the bot. He then keeps staying in the same position and looks away from the bot to his left door again and gets killed - That is the kind of difficulty I enjoy playing at, and he made several mistakes in that situation he could work on improving - Or he could play at a more suitable difficulty for him if he does not enjoy that kind of challenge. There can never be proof of balanced difficulty when difficulty is subjective, what I am trying to show is that it exists customer segments that enjoy that difficulty and that it should remain as an option. @Flair had a recent post that said it quite well imo.
I am not cherry picking examples from the games you mention, I simply just don't have the time to spend hours looking for them, so I tried finding some relevant videos. I would like to see more relevant examples that you may know of from the same games if you find the time to post them here, I am genuinely interested to see some cool bot behaviour somewhere.
I am aware the payday 2 video focused on team a.i. , a.i. is a problem in shooting games no matter the side they are on. I also think payday 2 is a fun game and played it quite much, I still think it is not very good a.i. wise both on the team and on the enemy side. The enemies shows up in extremely large numbers and are bullet sponges on the higher difficulties exactly for the reason that they cant make better a.i.- Stupid a.i. is not necessarily a problem when a game is fun to play. I never said scripts are not used also in multiplayer games, by all means they are - bots being spawned in to strategic positions when you have capped a point in Sandstorm is an example - What I tried to address was that people who complains about a.i. in multiplayer games and use singleplayer games as a reference have some standards and benchmarks of what good a.i. is, that the multiplayer coop games still haven't figured out, so comparing those types of games is misleading.
I looked at two vids made by Worth a buy (wab) today of brand new games 2019ish , new metro game (at least 1 stalker dev worked on that title) and the division 2. Both games are new and have according to dear old Mack poor a.i. - As I said it is a situation across the whole gaming industry with bad a.i. and I think NWI solves bad a.i. in a brilliant way with what they have to work with (compensating with bot aim being very good) - Ins2014 and Day of Infamy is both really good games, and I just hope NWI will keep having at least 1 high difficulty option (hopefully a higher max difficulty than they had in ins2014).
From u, about pistols:
"So the problem isn't the fact that they can headshot you from 100 meters with a handgun. Nope. It's the fact that they choose to use a handgun at 100 meters.
First of all, some bots only have pistols. In fact, many of them only have sidearms.
Second, pistols are somewhat viable at range, especially the PF940 and L106A1. The M9 can hold its own as well. That being said, they probably aren't accurate enough to actually hit a headshot at 60 meters.
You actually said that you'd refuse to comment on this one at all, even though I have a point I'm making here. GG."
From me about pistol:
Bots could be tweaked to not use pistols at unrealistic distance, that would solve the situation for players that like the challenge of bots able to oneshot from long distance at the high difficulty I hope stays as an option. I did not mean to refuse any point you had, I just did not have anything more to add. You seem to have a lot more knowledge than me on weapons. Maybe the devs could take the different realistic ranges that applies for pistols into account and solve the bug that you have noticed in this regard?
QUOTE from u:
EDIT: Let me sum up your posts so maybe you'll understand.
You keep telling me that I'm wrong without actually telling me why. Almost every "fact" that you're using to back up your arguments is objectively false, and everything else is your opinion and no one else's.
Still of this opinion or did my last few posts clarify?
QUOTE from u:
You've been accusing other people for things you do yourself, and you don't seem to have the ability to understand things from other people's points of view and will never seem to accept or even consider someone's else's position over your own.
Other people is currently approx 7.7 billion, most other people I encounter both professionally and socially does not give me that kind of feedback. Communication is a vital part of my real job, I would quickly lose my job if your statement was true=)
I said: "Is anyone enjoying the difficulty as it stands? Is there videos showing ppl plowing through the bots with ease? It is, so the perceived difficulty is subjective."
You answered: Why are you asking me instead of posting those videos yourself?
I did not ask you. If you read all three sentences you can notice I answer the question myself. This is known as rhetorical questions.
On a.i. in other shooters, you made a list. I will look at it here:
-> Sven Coop (really improves Half-Life AI) * Do you have any video examples? "Sven Co-op is an eclectic collection of user-created Half-Life maps that did one crucial thing differently: it allowed players to explore those maps together, battling AI enemies and solving simple environmental problems as a team."
All games based of half life can have similar ai as ins2014, as ins 2014 is also based of half-life. I havent played sven coop. Its after all 17 years old and there is many versions from it as I understand. Which version have this good A.I that you mention? Does it not have the same problems as half life ai and ins2014 ai?
Cry of Fear (also drastically improves Half-Life AI) *video below is a complete coop playthrough, I failed to find encounters with any smart ai, though I could never look at it all as it was not my type of game. Please highlight encounters with good ai or provide other examples if you like.
-> Chivalry: Medieval Warfare
Comment from community manager of the game.
"Bots are only in the game for testing purposes. This is a multiplayer only game the fact that there are bots is a miracle in itself."
-> Mount & Blade: Warband
Do I need to comment here?
-> PAYDAY 2 has decent AI, actually. It's based around Coop, after all. *You sure its decent compared to NWI?
Comment on video is by Generalmcbadass, 1 of the most respected players in the payday community. You can here see he is also talking about bullet sponges.
Not [multiplayer coop] = irrelevant. (takes some reading to understand why it is irrelevant though, I could dig up an old post if you don't like to read about those differences in the source material by learning about a.i. in different games.)
Any STALKER game (singleplayer, for the most part) * Only singleplayer or pvp, no coop.
End note: Do you want me to continue to answer all your questions as you asked me to, or did my answers suffice? I could make a post about differences between coop multiplayer ai and singlelayer ai if you don't see the significant difference - >The differences are numerous. For example in coop you have multiple human players and the bots must use tactics viable against groups, so all "a.i" resources can not be focused on one single blob of coordinates as the enemy - This in itself should tell you a lot. In singleplayer games you can also use trigger points to a large degree, which is an action happening when you step on a certain area etc ->scripted events like that helps cover flaws of ai.
btw: I love the stalker series its one of my favourite games of all time, though I mostly played those games with mods. Misery (https://www.moddb.com/mods/stalker-misery) is my favourite, though based on your feedback you will hate it tremendously, it is the hardest (singleplayer) game I have played, excluding ironman runs in for example doom (havent finished one)
Welcome to Unreal Engine 4. This isn't exactly Source here. A "few lines of code" isn't exactly accurate.
Okay, so if you know some basic principles about programming and such, can you actually explain why that would be easy?
Are you still of this opinion or did my previous answer clarify? Here is an example for you on the user friendly unreal engine. You can also search 1000s of other topics free on the web. I can make a more comprehensive guide for you if u want:
Ah, a common problem for some people. You have to look aaall the way to the right of the screen and you can see the numbers actually match your chosen input. So the numbers located on the right side of the picture of the loadout reflect the numbers you have customized. I hope this helps=)
youtube video found by doing the heavy lifting of writing ins sand coop in search field. @MarksmanMax i can also make a more comprehensive guide on how to do it yourself if you want. Hope this answered your question where you wanted me to find videos of the not difficult coop. video was the top result, its a bit boring to watch if u know the game (I know you have a lot of hours, so this sentence was not directed at you specifically). enjoy
This is exactly why we need difficulty settings that appeal to the gamers among us who does not find the rushing through the maps either tactical or difficult enough. Higher difficulties forces players to play tactical or die and then adjusting. Custom servers will hopefully at some point fix this if the brilliant modders behind servers like Bef from Ins2014 finds the time to make mods in Sandstorm as well, but a difficulty like @Flair describes should also be in vanilla, as there is many gamers who enjoy a tougher challenge.
My background is in engineering, math and physics, using matlab and having no time or patience to pick up coding.
Both Math and coding is similar in the way that it is very logical, so by learning math at university level you can understand several principles of coding as coding is also very logical. I also know coders and in engineering you learn to use a lot of formulas that is making coding looking a lot less confusing in general.
I did fool around in fallout 4 creation kit just to see how that worked (nuclear shelter 077b) is a free mod I made there if u wanna download it. But the "coding" I did there was extremely trivial as I just used some basics in the creation kit.
When you have defined a behaviour in bots that makes it aim and shot, defined cooldowns until firing and whatnot, you have used the premade tools in ue4 or made software that supplements the engine. (Good programmers will not only use the premade stuff as it will not be suited for all needs).
So we have gotten to a place where the bots work as they do now. Then it would be logical that they simply have a hit ratio defined as a parameter for how often the bot will hit the player when the player is in sight or other similar conditions are met. Like if the sound of the player which for the bot will probably be read like closer than xyz coordinate in a certain radius from the player etc.
The coders can add all kinds of layers to this, like the bot shoting in the general xyz direction of player and having hit probabilities in the area close to the player etc.
Somewhere in this lines of codes there will be numbers in percentage where 0 is no chance and 1 is a clear hit or a distribution of areas across the body of the player or similar mechanics (can be many layers). In the ue4 you can probably adjust many of these numbers just by changing the value of the numbers to increase or decrease the hit chance in easy to look at displays by default, as it is quicker to navigate and more user friendly than only looking at lines of code. Most devs would and should define a bunch of their own layers to suit their game, sometimes even having another gameengine working together with the ue4 engine. You can see many features of ue4 just by accessing a bunch of tutorials free online and looking in their forums.
So the devs will most probably have the possibility of just changing many of the different hit probabilities by changing a few numbers like adjusting a number for bot reaction time in seconds, the time a bot is staggered, how often it hits a certain area etc just by adjusting a few parameters-> Thats what I already said earlier and is why I mean it is not a lot of work to change many those parameters with the effect of changing the difficulty of the bots.
If the bots have zero ms delay and the players have x ms delay (no matter how fast your internet connection is you will have a significant delay compared to what happens in the server) it can create a challenge for the devs to make a system that accounts for different players having different delays and whatnot, so bugs can occur also here, and I was not referring to anyone in particular when you thought I talked about your internet connection.
The reason I dont adress all your questions or accusations as I would call some of it is because it doesnt make sense.
For example you say I am contradicting myself when saying that it is fair enough you dont understand what I write, and also I am saying its pretty straight forward so if u cant comprehend it who cares. There is no contradiction there, I am simply saying it is okey to not understand what I write, after all a huge percentage of the population dont even believe in climate change, but even if someone dont understand, I cant be responsible of needing to clearify every sentence I write as it is everyones business of reading up on subjects that interest them themselves ey=)
Yes it is easy to adjust the flinch and aiming of the bots. why in the holy jesus would I want to be a game developer just because I understand some basic principles about programming, I probably earn more than most developers you derp, and you have a problem with metaphors and started insulting me about it calling me drunk, and now you high horsing. Why would I need to make youtube videos of it its already plenty of them. Its fair enough you dont understand what I am writing and mean it doesnt make sense, but why do I care about that, its all pretty straight forward so if you cant comprehend it who cares
To be fair it might be bugs related to internet speed, latency etc that will give buggy experiences for some people, rising storm 2 vietnam has solved this well as it is the only game I can kill ppl in pvp with a ping in the top hundreds, therefore all feedback is valuable, just hope the devs see the difference of what is beard and what is snot ( another metaphor) what is the actual bugs and what is diversions.