Hraf

Could the AI actually have heard you turning the turret? From your last example, it seemed that turning the turret triggered the enemy action. Patch 1.03 is said to remove "super hearing" from the AI.

Well, my simple tests indicate that neither distance nor dimensions of target matter very much. See my post above.

The only sort of documented value is spotDistance in the Wiki where it is said that the low skill probability applies to the low skill value and lower skill and that high skill probability applies to high skill value and higher skills. For example, if the values are: {0.2, 0.3, 0.5, 0.6} then an AI with skill 0.2 or lower will have 0.3 probability and an AI with skill 0.5 and higher will have 0.6 probability and skill values in between will have probabilities in between. Of course I don't know the basic skill values of eg an OpFor riflemen. That skill value is then probably multiplied with the general AI difficulty modifier, eg 0.6 for Normal mode. But I don't think all the values work this way. For example spotTime seems more closely related to the actual time it takes for the AI to react. With a low spotTime value the AI takes foerver to even get prone if you fire a few shots in the vicinity, with a high value it reacts must faster.

No one knows what the values represent.

Seriously, a prone target at 300 meters with only iron sights is a HARD target just to hit. It is a pretty hard target to even spot. Shouldn't the AI on average have to work a bit more to kill you at that distance than if you are walking towards the enemy at 100 meters?

Did some testing of my own, trying out various values in the CfgAISkill table. My somewhat disturbing conclusion is that regardless of the values, is that it doesn't matter if I'm prone, kneeling or standing up at a 100 meters, 200 meters or 300 meters, the AI will still detect me and kill me in the same time. I can make it take longer or shorter time, depending on the values in CfgAISkill, but it doesn't matter if I'm standing up at 100 meters (pretty easy target) or lying down at 300 meters (pretty hard target). I might be exaggerating a little, but there wasn't any real difference. If I'm right, then it probably is a bug. I did my testing on the flat grass between the runways at the Vybor (Chernarus) airfiled using an opfor and a marine rifleman. Someone care to try to verify my result?

Note that my current CfgAISkill (just extracted from my unmodified config.bin) looks like this: class CfgAISkill { aimingAccuracy[] = {0,0,1,1 }; aimingShake[] = {0,0,1,1 }; aimingSpeed[] = {0,0.500000,1,1 }; endurance[] = {0,0,1,1 }; spotDistance[] = {0,0.200000,1,0.400000 }; spotTime[] = {0,0,1,0.700000 }; courage[] = {0,0,1,1 }; reloadSpeed[] = {0,0,1,1 }; commanding[] = {0,0,1,1 }; general[] = {0,0,1,1 }; }; There has evidently been changes to eg spotDistance and spotTime from those previously reported. I'm of course running 1.02.58134. Can someone else confirm this? Edit: Maybe those values always has been like this by default. But Thunderbirds custom CfgAISkill actually increases spotDistance and spotTime to: class CfgAISkill { aimingAccuracy[] = {0, 0.16, 1, 0.22}; aimingShake[] = {0, 0.18, 1, 0.22}; aimingSpeed[] = {0, 1, 1, 1}; courage[] = {0, 1, 1, 1}; commanding[] = {0, 1, 1, 1}; endurance[] = {0, 1, 1, 1}; general[] = {0, 1, 1, 1}; reloadSpeed[] = {0, 1, 1, 1}; spotDistance[] = {0, 1, 1, 1}; spotTime[] = {0, 1, 1, 1}; };

Did some more playing with the mod yesterday. Worked extremely well on longer distances, but led to some silliness on shorter distances. There was one instance where I was prone and a standing enemy was shooting at me from behind at perhaps 30-50 meters and was missing me all the time. Maybe that was a fluke?

I think this mod is a great step forward. I think AI detection AND accuracy is too good. The lethality of ARMA 2 firefights is way way too high. There are statistics on the average number of small caliber bullets fired per casuality in firefights and usually many hundreds or even thousands of bullets are needed to kill one enemy. Just check any of the numerous firefight videos from Afghanistan or Iraq on youtube. This ofcourse doesn't apply to a short range ambush situation, but a long range firefight where all participants are prone or behind cover. I also think the enemy can pinpoint your position way too easy. I belive that the enemy quite easily should detect your general position once you eg fire a weapon, but that they hit and kill you with just a few very well aimed bullets when you are kneeling or prone under a bush at 400 meters is just silly. There is also a playability element - if I have problems spotting the enemy at 400 meters, it gets kind of boring if he (or my AI teammates) spot each other every time without fail.

I had the exact same experience. Drove to the range, started doing AT practice. Then the MOUT guys show up and kill me.

I don't think the modded textures are larger, they are just prettier. So, it shouldn't have any performance impact. It is an interesting theory that Chernarus maybe uses the wrong mid detail textures, or maybe is missing those altogether. Any experienced modder could probably have a look and compare the textures between Utes and Chernarus.

I had the exact same problem. First the base is wrongly named (it is called Starforce, not Reforger), ie it is the base where you start the mission. Then I couldn't activate the UAV until I pointed my crosshairs at THE SOLDIER standing next to the UAV control trailer. Then I just recently learned that you can change waypoints by opening the map view when controlling the UAV. You can also change the flying height of the UAV with PgUp/PgDown