Enemy Accuracy way to good.

[Hell_Toupee 0]

there are enemy and friendly accuracy sliders....

[Sonar 0]

Hi all

@ Sonar with all due respect

To make a statment that the AI if using communist block weapons should not be able to hit you at 500m can only be true if you can show an experiment to prove they are doing something that a player cannot achieve.

So I say to you, if you have ArmA II go on prove it. Make a repeatable experiment that proves what you say is true.

As to your statement

I would submit it is not I believe a valid one.

We obviusly still have people who do not know the difference between the AK47 and the AK74 and AK107 used by most OPFOR in ArmA II.

You would have thought people could at least use Google or even Wikipedia.

http://en.wikipedia.org/wiki/AK-47

http://en.wikipedia.org/wiki/AK-74

http://en.wikipedia.org/wiki/AK-107

Educate your self.

Comparing the AK47 of myth and legend with the AK74 or AK107 prevalent among OPFOR in ArmA II is like comparing chalk and cheese.

The bullet is different

The mechanics are different

The accuracy is vastly different

As a weapon the AK74 is straight up comparable with the M16/M4.

You are caught up in the mythology of the AK47.

You are thinking that:

1) The AK47 of myth and legend you see in civil wars and conflicts around the world fired by everyone from wizened Afghan sharpshooters to drugged up child soldiers in Africa and manufactured from bits of old pipe and petrol cans or at best the worn out weapons sold from conflict to conflict, by the the likes of Viktor Bout, and exported from every pirate and dodgy factory around the world and used in every conflict since the Korean war, and loaded with bullets from every backyard factory and bazar workshop, and with recovered spent cartridges, filled with everything from fireworks fuel and old match heads to even sand and charcoal.

....is the same as the:

2) The AK74 designed and developed and manufactured in the Izhmash factories in Russia.

...is the same as the:

3) The AK107 designed and developed and manufactured in the Izhmash factories in Russia.

They are not.

So can we please stop having the myth of AK inacuracy perpetuated on this forum? This is not stupid COD land.

Kind Regards walker

:rolleyes:Ok I accept your point on the difference's of AK47 AK74 AK107. But I generalised with the statment the" AK's" I did not state AK47. I will standby the statement that"AK's" are inaccurate.However Wiki is not a quotable source....why. No university allows it to be referenced as it is not controlled.“Wikipedia entries are inherently unreliable because they can be written and edited anonymously by anyone.â€So as my reference I tend to use,Rifles of the World by John Walters isbn 1-85409-9 mines the sixth edition so it is a little dated. I also use Janes Infanty Systems, which I dont have on my desk right now, because I usually borrow it from the library because its so damn expensive.

Another point...I was a senior NCO in the Royal Australian Navy. I served for 16 years as a Weapon System technician and did 2 years in the Australian Army as a Infantryman. I have been shot at by most variants of the AK and are still very much alive at last check. I should point out that I believe M16's and generally all 5.56mm NATO weapons are also inacurate and pretty much usless. This is because my intial training was with the L1A1 7.62x51mm NATO.In fact when playing any version of ArmA I tend to dump 5.56 and grab any 7.62x51mm cal weapon I can find...call me old fashion:D. Also I was rated as a "A"class marksman and today I still shoot competion over ranges of 900metres in all calibres up to 8mm, so I believe my "Education" is adequate.:cool:

The AK47 was a 7.62X39mm round, the newer versions are 5.45 x39mm, however intel sources...cheifly the US Army quotes ...."In it's accuracy AK-74 surpasses AKM in 20 - 60%. So realistically no great leap. Also the mechanics are not different.The differnence is in the size of extractor claws and locking lugs, reciver manufacture, stock and other hardware manufacture. In fact apart from improved manufacture and caliber change there is little differnce. If you were trained to use a AK47 (or in my case a Chicomm Type56 which is a pain cuase I'm 185cm tall and the 56 is made for someone about 152cm tall so the butt is incrediably uncomfortable for me:mad:) you can pick up a AK74 and still use and field strip it!

The AK107 however is a completly different kettle of fish, but I feel I have made my point.

Xenios, I wouldnt consider the Iraqi army poorly trained or equipped. People forget that proir to the 1990 invasion of Kuwait; Iraq had been in a full on no holds barred war with Iran for 8 years and the USA, UK, USSR, France , Germany, North Korea and others supplied both sides and Belguim, Spain, Portugal and Egypt supplied only Iraq. So a lot of weapons testing went on. I digress. So in 1991 when we hit back the Iraqis had 8 years of experience whilst the majority of the US hadnt been in a major War since 1975. Now having said that the Coalition forces whacked Sadam quite nicely. However the US Army only lost 21 M1A1 to moslty blue on blue with only 2 being confirmed as being destoryed by Iraqi T72's(with only one kia). Other non Republican Guard units had T62's and Type 69's (ChiComm T62) , so I dont know were you get the idea that they only had mostly T-55s and and a few T-72s constructed in Iraq with rejected Soviet parts.I served during the Gulf War(1), did you?

Anyway I still believe the T72 model is a too tough and not realistic compared to a M1A1 and the AK's are too accurate over range. :D

I will do the experiment for you to prove my point and will post on youtube. please feel free to set the parameters of the experiment.

[max power 21]

Allow me to quote myself:

The ak47 isn't that inaccurate. 3 to 5 MOA at 100 meters with a max effective range of 350 meters. 4 moa means 4 inches dispersion every 100 yards, and maximum effective range is the range a trained shooter can expect to get 50% hits on a man sized target.

The m16a2 with milspec ammo can get 1.5-3 MOA at 100 meters with a maximum effective range of 600m.

The ak-74 seems to be able to achieve groupings of 2 MOA. The site I got that figure off of seems to say the shots were fired at 300 meters.

The m16a2 has better sights for long range shooting and has less recoil in rapid shooting than the ak47. I am not sure about the ak74.

If you wish to see the context:

http://forums.bistudio.com/showpost.php?p=1296030&postcount=44

None of this information was found on wikipedia. Most of it was found by keyword searching and looking at shooter's club forums and such.

[Hell_Toupee 0]

20-60% improvement is not too much to you?

Iraq was poorly trained and equiped, their equipment was not only old but crippled export models, their war with Iran did not leave them more prepared but expended and heavily in debt.

Iraqi tanks are just not representive of Russian ones, they did not have quality ammunition,armor or sights.

[whisper 0]

Wasn't BIS bragging about their AI? This seems to be a game killer issue due to lazy programming.

If you blindly believe the statement that "AI cheats and see through all objects", sure. But this statement is wrong in all accounts

[Sonar 0]

"4 moa means 4 inches dispersion every 100 yards"

It does not! Well not really.MOA means Minute of Arc (or Angle) it equates to 1.047" at 100yds and 10.47 at 1000yds so at 400yds it would equal 4.188" So 4MOA would be 4.188 at 100yds and 16.752" at 400yds at 1000yds it's 41.88".

Some people will mistake MOA for Mils which dosnt work. one MOA is 1/60th of a degree one mil is 1/6400th of a degree. ps the M24 scope is a mildot.Heres a good website to learn how to use them http://www.mildot.com/index.htm

I have fired a Type56 and a SKS both which are 7.62x39mm and there is no way I would pull a 3-5MOA group at 350Metres. And to qualify as a "A"class marksman I had to hold a sub MOA group at 300metres, which is less than 3 inch group...from memory I shot 2 3/4" group. I should point out that I still shot sub moa groups at ranges up to 900metres, I have to , to remain competative.

I still do not believe the commbloc weapons are as accurate as people think. Its one thing to read these numbers on a web page but its different on a bi-directional rifle range.

Milspec ammo.....my favorite! One day for shits and giggles we took a 100 round random sampling of Australian made 7.62.51NATO rounds. Put them in a ballistic Hammer, broke em and weighed each component. We found a 10-15% variation in powder weight, projectile weight and case weight ( which does affect accuaracy) Non milspec ie commerical ammo had less than 5% variation and Laupa Target Rifle ammo has less than 1% variation. And these are from countries with good to excellent quality control. My own hand loads for my M82 is less than .1% variation.

Hell Toupee."Iraqi tanks are just not representive of Russian ones" No they are not because they were Polish ones. And they had better sights because the Belgium supplied them with Thermal sights, however this is NOT confirmed and I was never here!!!! And the 2 M1A1 kills suggest that they did have "Quality" Ammo because the hits were caused by APDSFS rounds, and as they had no radiological reading they were not DU, which all collalition forces used. However one tank had a HEAT signature on its engine cowling suggesting that it had been first hit by a Hellfire.....this is not confirmed and I was never here! ( notice how I skip intelligence questions quiet nimbly!)

When you are playing with MOA numbers 20-60% isnt all that big. we are talking 1.047 to 1.704 INCHES so no I dont consider it a big increase.

Anyway you can all say what you like as far as I am concerned combloc weapons are inaccurate for various reason's including I have been shot at by them and am still here, I dont like or play CS and I still think combloc weapons are inaccurate so :p pfffffft!

Edited June 11, 2009 by Sonar

[whisper 0]

Just to add my stupid 2 pennies, and sorry for Sonar even though your input is greatly appreciated, it's one area where I'd prefer to see gameplay considerations being more important than realism ones. Depict the difference between weapons, yes, being exactly what is seen on theater of wars (and tbh I think we lack much much informations about the current state of Russian Army, you could very be surprised seeing the recent events), no.

[bloodbomb 12]

Yeah I've shot an old AKM in both SA and FA, and though its obviously a pretty accurate weapon its not really that amazing, not compared to a highly tuned hunting rifle or even an M16 or M4...you'd do well to listen to Sonar he knows his shit...

A friend that was stationed in Afghanistan saw an Afghan national with a Rangers Marksmanship badge on his arm, he asked the Rangers why they gave him one and they told him that he had nailed a tango at like 900 meters. Apparently the Rangers were all sniping at this running Taliban and joking how no one could hit him, next minute the Afghan national comes out with his old Ak-47 and just plain old ironsights and empties half his clip downrange, walking the fire up till he tags the tango, then calmly walks back inside. They said he basically machinegunned the tango at 900 metres (not surprising as the AK47's workings are basically (sometimes literally) machine gun parts) so they gave him a Ranger badge lol

Not trying to say that makes the AK an accurate weapon, quite the opposite, what I'm saying is that often it comes down to the individual, his skill and technique, and the individual weapon he is using and how it has been tuned/treated etc

It all comes down to the individual most of the time (ie: Karate vs Kung Fu, AK vs M16, Ford vs Holden etc etc)

The average insurgent isn't an amazing shot though, so I hope theres not too much hax shooting on the AI side (as their was in OPF) as I like to crawl around and hide and set up ambushes etc, and I like to have prolonged gun battles (sometimes) which actually is quite realistic at long ranges, just go on LiveLeak etc and watch some of the real life gun battles (or ask a soldier whos been there, I have) and theres a whole lot of spamming (er-I mean "sweeping fire") and missing going on from both parties.

[Sonar 0]

I think we should move this disscussion to the off topic thread as we are going to fill this one up with off topic stuff before much longer.

[KeyCat 131]

It's because the AI cheats, nothing blocks their view, they know where you are everytime you fire a shot.

If true, this sounds alot like the issue we experienced in ArmA (IIRC this was fixed in 1.12) where the AI could pin-point your location with ears alone, see thread here.

The question is... can you back that up with some piece of evidence? A simple test mission that isolates the issue and shows it in a way that can be reproduced along with notes about your AI settings.

I'm still waiting on the 505 release so I don't know how good/bad the AI performs in ArmA 2 but before claiming something like above I would have made sure to have some kind of evidence else it's just an opinion and your'e just blowing "hot air".

Having an opionion is fine but it won't help fix things (if they are indeed broken?). Best (proven) way to get BIS attention is to provide a simple testmission along with your system spec. and enough details so they can reproduce the issue/bug on their end. They may not fix it overnight but they will look into it when time permitts, at least thats my personal experience since bying OFP June 2001.

@Junker: I'm also very interested to hear if the setSkill array command works this time around? As Carl Gustaffa says it's kinda confusing if you look at the Biki?

/KC

Edited June 11, 2009 by KeyCat

[GoOB 0]

Hopefully I can say this without repeating anyone else's words.... But wasn't this (perceived) problem in already present in OFP? I vividly remember the early days when you and your co-op mates would get single shots to the head by PK gunners in watchtowers whilst crawling through some forest in Malden. The problem I feel isn't detection, but the fact that the AI is so bloody accurate.

[LeadCommando65 10]

About cheating AI,

found this vid http://www.youtube.com/watch?v=Y6mI410K_ZU

I would guess this kind of test that walker have been preaching about

This shows very good how the AI reacts, but I think it's only because the creator of this video was far away from the enemies.

I got multiple times in the campaign situations where my teammates spotted enemies/objects *behind* hills (in forests or not) and at a really far distance (~ 1000m) as the team was on foot. So I predict if my teammates can even spot such enemy/man locations, the enemy AI is able to that, too. And if you are about 200m near to the enemy it becomes really deadly to you...

Additionally I ran various tests with soldiers & static weapons in the editor to see how precise the enemy AI can shoot. There's little to complain about the standard soldiers. BUT as I tried the Russian KORD machine gun (mini), the first shot was always a headshot. There must be something wrong with the weapon...

[KeyCat 131]

Hopefully I can say this without repeating anyone else's words.... But wasn't this (perceived) problem in already present in OFP? I vividly remember the early days when you and your co-op mates would get single shots to the head by PK gunners in watchtowers whilst crawling through some forest in Malden. The problem I feel isn't detection, but the fact that the AI is so bloody accurate.

Hehe, yes clearly remember the PK gunners in OFP :)

But since ArmA 1.00 BIS allowed us to easily change the AI's precision by editing XXXXX.ArmAProfile (as well as the skill slider in editor).

	skillFriendly=0.820000;
	skillEnemy=0.820000;
	precisionFriendly=0.480000;
	precisionEnemy=0.480000;

It's not a issue anymore (or at least wasn't in ArmA) and you are able to set AI's precision to your own liking. I'm assuming it's the same in ArmA 2, please correct me if I'm wrong...

Edit: Just to be clear.... I'm not saying that ArmA II 1.01DE have no issues/bugs/flaws regarding this topic. I'm unable to test myself until after 19:th June.

/KC

Edited June 11, 2009 by KeyCat

[Hell_Toupee 0]

"4 moa means 4 inches dispersion every 100 yards"

It does not! Well not really.MOA means Minute of Arc (or Angle) it equates to 1.047" at 100yds and 10.47 at 1000yds so at 400yds it would equal 4.188" So 4MOA would be 4.188 at 100yds and 16.752" at 400yds at 1000yds it's 41.88".

So otherwise 4 inches every 100yards then.

Hell Toupee."Iraqi tanks are just not representive of Russian ones" No they are not because they were Polish ones. And they had better sights because the Belgium supplied them with Thermal sights, however this is NOT confirmed and I was never here!!!! And the 2 M1A1 kills suggest that they did have "Quality" Ammo because the hits were caused by APDSFS rounds, and as they had no radiological reading they were not DU, which all collalition forces used. However one tank had a HEAT signature on its engine cowling suggesting that it had been first hit by a Hellfire.....this is not confirmed and I was never here! ( notice how I skip intelligence questions quiet nimbly!)

I'm not aware theres any evidence of any Iraq tanks using thermal sights, most did not even have range finders. They did not have quality ammo, they had early generation APFSDS rounds with steel penetrators. Russian tanks had thicker armor as well as composite and reactive armor.

Majority of their tanks were chinese type-69s.

When you are playing with MOA numbers 20-60% isnt all that big. we are talking 1.047 to 1.704 INCHES so no I dont consider it a big increase.

Its range to hit a man sized target would also be increased by 20-60%, i would consider that pretty big.

Anyway you can all say what you like as far as I am concerned combloc weapons are inaccurate for various reason's including I have been shot at by them and am still here, I dont like or play CS and I still think combloc weapons are inaccurate so :p pfffffft!

Accuracy is more about the shooter than the weapon,

[NoRailgunner 0]

AI "eyes" or "ears" needs bit tweaking - run a test in dense fog: players arent able to see anything but the AI could easily spot and engage their opponents.

[Sonar 0]

So otherwise 4 inches every 100yards then.

I'm not aware theres any evidence of any Iraq tanks using thermal sights, most did not even have range finders. They did not have quality ammo, they had early generation APFSDS rounds with steel penetrators. Russian tanks had thicker armor as well as composite and reactive armor.

Majority of their tanks were chinese type-69s.

Its range to hit a man sized target would also be increased by 20-60%, i would consider that pretty big.

Accuracy is more about the shooter than the weapon,

Totally and utterly agree with that. Here's a little known factoid. US Armay Tropps do .5 hour dry firing for every hour of live, USMC do 2 hours dry to 1 live Australian Army do 3 to 1 The Russians do 5to1. Even with inaccurate rifles Soviet infantry were better shots than US Army.

The NON REPUBLICAIN GUARD Divisions were armed with Type 69. The Republicain Graud Regiments WERE armed with Babylon Lions. Which were T-72M1A. So whilst not as good as the Russian ones they were better than the Type69's. You will not see refernces to thew Sights because it is NOT confirmed ( however it has never been openly denied either) Some of them also had laser range finders, which, also, is not confirmed. The Iraqi T72's did have laminated armour on the front and slope. But my argument is that the T72 model used in game isnt accurate I used the Iraqi tanks as a reference point because it is one of the few wars where M1's came head to head with T72's. I will admit its probably not the best example, but it does annoy me when people make statments likepoorly trained and led militia using mostly T-55s and and a few T-72s constructed in Iraq with rejected Soviet parts. They wernt militia and they wernt poorly trained. The Republican Guard were enough of a threat that the Airwar proir to Desert Storm specifcally targeted RG Divisons from the outset. You dont waste Billions of Dollars of Ordanance on militia poorly trained or otherwise.(well the USAAF wastes it on who they want really:D) Some people forget I served during GW1 a point I made earlier and asked the person if they had served....still no answer.

You obviously dont understand exterior ballistics, accurary may incresae range but it wont increase kinetic energy, and an increase of 60% is only 180 odd yds. Boy and I thought my maths skills were bad!:D

To the MOA question roughly yes accuratly no. 1moa at 100yds is 1.047".so for every increase of range and/or angle it incresases by that much again. So at 1000yds its not a 10 inch increase but an 10and 1/2 inch increase and when you are shooting at that range 1/2 inch is the differnece between life and death. I know Im being anal about this but it is fact and facts I have had to work with. If I make a mistake with a calculation on elevation and windage at +500yds it is the differnce between a vbull and 5 points and I will lose a competition. That is the one saving grace now, because before if I made the same mistake I would lose my head!

[walker 0]

Hi all

Can people please keep up

AI Cheat when spotting you has been disproved

http://www.youtube.com/watch?v=Y6mI410K_ZU

MYTHBUSTED

AI have eyes in the back of their head

This myth was thoroughly debunked and scientificly examined in the old forums in this thread

http://www.flashpoint1985.com/cgi-bin/ikonboard311/ikonboard.cgi?;act=ST;f=64;t=73748;st=0

MYTHBUSTED

AI can see through bushes

This was tested by fabrizioT in this very thread

http://forums.bistudio.com/showthread.php?t=73900&page=6

And the video:

MYTHBUSTED

fabrizioT has done several other experiments into the AI capabilities.

Can AI see you in ground clutter when a human could not?

The current statment that is under review and has not yet been scientificly tested with a repeatable mission is: that AI can see an entity in grass cover when a human would not be able to see them.

This would be a fairly easy experiment.

Simple mission

Place one soldier lying in grass place another standing beyond, waypoint AI untill they see the soldier lying in grass, use the knows about command to find when they spot the soldier in the grass.

User camera mode to test if soldier lying in grass is visible to a human player at that moment.

Improvements to the experiment

Reduce variables by making the laid down entity static, by lobotomizing the AI in it.

Addendum experiment: Consider also the effects of sound. Does any one know of a way of finding out how the AI spotted you whether it is sound or vision?

Kind Regards walker

Edited June 11, 2009 by walker

[maddogx 13]

Does any one know of a way of finding out how the AI spotted you whether it is sound or vision?

It's been a long long time since I tested this in ArmA, so take this with a grain of salt:

I seem to remember there being a certain threshold in the knowsAbout value, something like 1.2 - the detection by ear could never go above it. Was probably the informational equivalent of "some human being is definitely there, but identity is completely unknown". As soon as the AI had actually seen the unit, the knowsabout value would go above the threshold.

Like I said, I'm no longer quite sure of this. I may be mixing it up with something else.

[Sonar 0]

http://forums.bistudio.com/showthread.php?p=1306856#post1306856 for those who wish to battle further....have at you scoundrals!!!!

[Hell_Toupee 0]

You obviously dont understand exterior ballistics, accurary may incresae range but it wont increase kinetic energy, and an increase of 60% is only 180 odd yds. Boy and I thought my maths skills were bad!:D

I was not meaning maximum range the bullet travels, but max effective range.

To the MOA thing, its not target shooting its an assault rifle you arnt going to be taking 1000yard shots let alone worry about 1/2 inch :P

Edited June 11, 2009 by Hell_Toupee

[An-225 0]

What about the AI knowing when one of their group members dies from several hundred meters away?

I tested this in ArmA, using a silenced weapon with subsonic ammunition. Upon killing a group member, the squad leader of that group turned and faced my direction. I think the auto-detection of a downed squad member is quite unrealistic, and can ruin some stealth missions.

[walker 0]

Hi all

An-225 please post your experiment so it can be verified and examined. A video on youtube would also help.

How many shots did you fire?

Entry and exit wound and smoke could all give a human in reality your direction of fire. The experience and training an officer would have would also point them to likely attack directions.

There is no such thing as a silencer for rifles. It is a suppressor, it merely suppresses flash, sound and smoke it does not remove them. If you were in the frontal arc when firing or moving or not in proper cover each would lead to a spot.

At what distance where you?

As I said make the mission you are performing the experiment in available so others can test your hypothesis.

Kind Regards walker

Edited June 11, 2009 by walker

[sandzibar 0]

I think the whole "hive mind" aspect of current arma AI (and presumably Arma2) needs some tweaking.. their intra squad communications is just too damn accurate and efficient.

I believe that this is contributing much to the the perceived accuracy issues.

[maddogx 13]

What about the AI knowing when one of their group members dies from several hundred meters away?

I tested this in ArmA, using a silenced weapon with subsonic ammunition. Upon killing a group member, the squad leader of that group turned and faced my direction. I think the auto-detection of a downed squad member is quite unrealistic, and can ruin some stealth missions.

True. I've noticed as squad leader that you can send your group members miles away and everything will be fine. But when that group member dies (there's a delay, not sure how long) you automatically ask that member for his status. When he doesn't answer he is declared dead. While he is alive there can be no communication for ages, no problem.

[whisper 0]

True. I've noticed as squad leader that you can send your group members miles away and everything will be fine. But when that group member dies (there's a delay, not sure how long) you automatically ask that member for his status. When he doesn't answer he is declared dead. While he is alive there can be no communication for ages, no problem.

Seeing the delay put before leader asks for his troop, I really don't see that as a problem. what is the issue with this behavior?