FlightControl Posted November 5, 2015 Posted November 5, 2015 (edited) I don't know what is wrong with version 1.5, but it seems that the AI of the helicopters and planes is bugged. My Gori Valley is an imersive ground attack / defense mission, and the helicopters with AI are flying around like rookies on the battlefield. They approach the enemy till the last 500 yards and still they are not firing their missiles or cannons. They do react well when fired upon in terms of evasive manoeuvring. As a result, tanks are driving around freely without air threats... Note that air-to-air battles between helicopters is ok though... Sven Edited November 9, 2015 by BIGNEWY Title [TABLE][sIGPIC][/sIGPIC]| Join MOOSE community on: DISCORD :thumbup: Website of the MOOSE LUA Framework. MOOSE framework Downloads. Check out Example Missions to try out and learn. MOOSE YouTube Channel for live demonstrations and tutorials. [/TABLE]
FlightControl Posted November 8, 2015 Author Posted November 8, 2015 Test Mission I don't know what is wrong with version 1.5, but it seems that the AI of the helicopters and planes is bugged. My Gori Valley is an imersive ground attack / defense mission, and the helicopters with AI are flying around like rookies on the battlefield. They approach the enemy till the last 500 yards and still they are not firing their missiles or cannons. They do react well when fired upon in terms of evasive manoeuvring. As a result, tanks are driving around freely without air threats... Note that air-to-air battles between helicopters is ok though... Sven To simulate the problem, find attached a test mission. Follow the MI-28N, and observe. It just does not fire to the tanks. You can play with the mission configuration and test other helicopters.Test Mission Fire Problem 1.5.zip [TABLE][sIGPIC][/sIGPIC]| Join MOOSE community on: DISCORD :thumbup: Website of the MOOSE LUA Framework. MOOSE framework Downloads. Check out Example Missions to try out and learn. MOOSE YouTube Channel for live demonstrations and tutorials. [/TABLE]
ED Team BIGNEWY Posted November 8, 2015 ED Team Posted November 8, 2015 Thanks for the PM's svenvandevelde I will take a look Forum rules - DCS Crashing? Try this first - Cleanup and Repair - Discord BIGNEWY#8703 - Youtube - Patch Status Windows 11, NVIDIA MSI RTX 3090, Intel® i9-10900K 3.70GHz, 5.30GHz Turbo, Corsair Hydro Series H150i Pro, 64GB DDR @3200, ASUS ROG Strix Z490-F Gaming, PIMAX Crystal
ED Team BIGNEWY Posted November 8, 2015 ED Team Posted November 8, 2015 If you use enroute task - search then engage the AI helicopters will engage. Forum rules - DCS Crashing? Try this first - Cleanup and Repair - Discord BIGNEWY#8703 - Youtube - Patch Status Windows 11, NVIDIA MSI RTX 3090, Intel® i9-10900K 3.70GHz, 5.30GHz Turbo, Corsair Hydro Series H150i Pro, 64GB DDR @3200, ASUS ROG Strix Z490-F Gaming, PIMAX Crystal
Grimes Posted November 8, 2015 Posted November 8, 2015 Newy, "CAS" is an enroute task. I reported it. Its a very strange issue as the weapons available and target seems to play a role. If the HAI decides to enter a hover they will be perfectly content sitting there and shooting the targets. However the Mi-28 against tanks doesn't want to enter a hover and prefers to attack moving slightly forward, which is how it get stuck in a loop attempting to extend and maneuver to find a new shooting position but never finds it. Against IFVs the Mi-28 gets into a hover and engages normally. Interestingly the Mi-24 has similar behavior, but it is able to turn back onto the targets to engage. Ka-50 and Apache seem to be just fine due to their primary ATGM having longer range and they hover to attack. The right man in the wrong place makes all the difference in the world. Current Projects: Grayflag Server, Scripting Wiki Useful Links: Mission Scripting Tools MIST-(GitHub) MIST-(Thread) SLMOD, Wiki wishlist, Mission Editing Wiki!, Mission Building Forum
FlightControl Posted November 9, 2015 Author Posted November 9, 2015 Well, I tried the ka-50 also, but they just don't shoot... [TABLE][sIGPIC][/sIGPIC]| Join MOOSE community on: DISCORD :thumbup: Website of the MOOSE LUA Framework. MOOSE framework Downloads. Check out Example Missions to try out and learn. MOOSE YouTube Channel for live demonstrations and tutorials. [/TABLE]
ED Team BIGNEWY Posted November 9, 2015 ED Team Posted November 9, 2015 Thanks Grimes Forum rules - DCS Crashing? Try this first - Cleanup and Repair - Discord BIGNEWY#8703 - Youtube - Patch Status Windows 11, NVIDIA MSI RTX 3090, Intel® i9-10900K 3.70GHz, 5.30GHz Turbo, Corsair Hydro Series H150i Pro, 64GB DDR @3200, ASUS ROG Strix Z490-F Gaming, PIMAX Crystal
Grimes Posted November 9, 2015 Posted November 9, 2015 Well, I tried the ka-50 also, but they just don't shoot... It does in your sample mission if I give the flight a short route so they don't RTB immediately. Can you describe in detail what you see on your end or attach a track? "they just don't shoot" is not an adequate description. The right man in the wrong place makes all the difference in the world. Current Projects: Grayflag Server, Scripting Wiki Useful Links: Mission Scripting Tools MIST-(GitHub) MIST-(Thread) SLMOD, Wiki wishlist, Mission Editing Wiki!, Mission Building Forum
FlightControl Posted November 11, 2015 Author Posted November 11, 2015 Test Cases and Test Results Hi Grimes and Co, You asked me to be more specific, well, find the following test results ... :pilotfly: ... I've made test cases with 3 helicopters, in 3 battle field scenarios, testing 3 different engagement types for each helicopter type. Helicopters: - KA-50 - MI-24V - MI-28N Battle Scenarios: - Tanks Activated, SAMs not activated, Not moving. - Tanks Activated, SAMs activated, Not moving. - Tanks Activated, SAMs activated, Moving. Engagement Types: - Weapons Free, Passive Defense - Weapons Free, Allow Abort Mission - Designated Only, Allow Abort Mission For each combination, find below the test results table. The results show: - Effectivity = +++ = very good; --- = very bad - Missiles fired = How many missiles were fired in 10 minutes of flight - Time and Frequency = How long till firing first missile, and how frequent missiles were fired. - Notes = Explanation of a few more details. In a summary, the following can be concluded: 1. All helicopters get really ineffective, if Ground units are moving, and Reaction to Threat is Allow Abort Mission. then the effectivity of AI is completely useless. 2. When ground units are not moving, AI is also ineffective, but depending on the helicopter, AI can still fire missiles. 3. The KA-50 is the best helicopter around. 4. MI-24V helicopters don't hover, the fly directly into the battle zone. 5. MI-28N helicopters are the least effective. I would really study the table, it shows the effectivity and the ineffective behaviour of ROE and RTT paramters. This really needs to be looked at. For each scenario, I've added a ZIP files the for each test case a separate mission (miz-file), so anybody can use that to test quickly the AI effectivity. I've only tested the RED parts right now, the BLUE ones are still to come. One last note: There is an other bug. When the KA-50 fires a missile, the vikr BAY on the wing does not show an empty slot after firing. I'll report that also in a separate mail. Find the table of the findings below: Ref Effective Side Unit ROE RTT SAMs SAMs Missiles Time in sec Additional Notes Active Moving Fired Start/Freq --- --------- ---- ---- --- --- ------ ------ -------- ----------- ---------------- 1.1 +++ Russia KA-50 Weapons Free Passive Defense No No 6 60/10 Perfect! 1.2 ++ Russia KA-50 Weapons Free Allow Abort No No 6 60/10 Heli flies first to safety zone before firing. 1.3 ++ Russia KA-50 Only designated Allow Abort No No 6 60/10 Heli flies first to safety zone before firing. Hovering and firing. Perfect. 1.4 +++ Russia KA-50 Weapons Free Passive Defense Yes No 6 90/10 First scans the SAMs, then fires the SAMS, then the other ground units. Perfect! 1.5 +++ Russia KA-50 Weapons Free Allow Abort Yes No 6 180/20 Heli flies first to safety zone before firing. First scans the SAMs, then fires the SAMS, then the other ground units. Perfect! 1.6 +++ Russia KA-50 Only designated Allow Abort Yes No 6 180/20 Heli flies first to safety zone before firing. First scans the SAMs, then fires the SAMS, then the other ground units. Perfect! 1.7 +++ Russia KA-50 Weapons Free Passive Defense Yes Yes 6 60/10 First scans the SAMs, then fires the SAMS, then the other ground units. Perfect![color=Red] 1.8 --- Russia KA-50 Weapons Free Allow Abort Yes Yes 0 - Helicopter is flying like a rookie. No missiles fired, and flying around... 1.9 --- Russia KA-50 Only designated Allow Abort Yes Yes 0 - Helicopter is flying like a rookie. No missiles fired, and flying around...[/color] 2.1 +++ Russia MI-24V Weapons Free Passive Defense No No 8 90/15 Attacking, firing missiles. Because the heli is a flying tank, it engages on the field. 2.2 - Russia MI-24V Weapons Free Allow Abort No No 5 180/60 Heli flies around for 2 minutes before firing a missile. Flies around a lot before firing. 2.3 - Russia MI-24V Only designated Allow Abort No No 4 60/15 Heli flies around for 2 minutes before firing a missile. Flies around a lot before firing. 2.4 +++ Russia KA-50 Weapons Free Passive Defense Yes No 8 90/15 Attacking, firing missiles. Because the heli is a flying tank, it engages on the field. 2.5 + Russia KA-50 Weapons Free Allow Abort Yes No 5 180/60 Heli flies around for 2 minutes before firing a missile. Flies around a lot before firing. 2.6 + Russia KA-50 Only designated Allow Abort Yes No 6 180/60 Heli flies first to safety zone before firing. First scans the SAMs, then fires the SAMS, then the other ground units. Perfect! 2.7 + Russia KA-50 Weapons Free Passive Defense Yes Yes 4 60/30 Flying around firing missiles. Low effectivity due to missile older technology. [color=Red]2.8 --- Russia KA-50 Weapons Free Allow Abort Yes Yes 0 - Helicopter is flying like a rookie. No missiles fired, and flying around... 2.9 --- Russia KA-50 Only designated Allow Abort Yes Yes 0 - Helicopter is flying like a rookie. No missiles fired, and flying around...[/color] 3.1 +++ Russia MI-28N Weapons Free Passive Defense No No 8 60/30 Perfect! 3.2 ++ Russia MI-28N Weapons Free Allow Abort No No 8 60/30 Fires missiles. Evades first to safety zone. 3.3 -- Russia MI-28N Only designated Allow Abort No No 4 60/60 Sporadically fires missiles. Flying around like a rookie. 3.4 + Russia MI-28N Weapons Free Passive Defense Yes No 8 90/10 First scans the SAMs, then fires the SAMS, then the other ground units. Perfect! 3.5 + Russia MI-28N Weapons Free Allow Abort Yes No 8 180/10 KA-50 flies first to safety zone before firing. First scans the SAMs, then fires the SAMS, then the other ground units. Perfect! 3.6 + Russia MI-28N Only designated Allow Abort Yes No 8 180/10 KA-50 flies first to safety zone before firing. First scans the SAMs, then fires the SAMS, then the other ground units. Perfect! 3.7 + Russia MI-28N Weapons Free Passive Defense Yes Yes 8 60/10 Low effectivity due to moving targets. [color=Red]3.8 --- Russia MI-28N Weapons Free Allow Abort Yes Yes 0 - Helicopter is flying like a rookie. No missiles fired, and flying around... Helicopter was shot down. 3.9 --- Russia MI-28N Only designated Allow Abort Yes Yes 0 - Helicopter is flying like a rookie. No missiles fired, and flying around... Helicopter was shot down.[/color] Helicopter Missile Firing Bug in 1.5 - Cases.zipHelicopter Missile Firing Bug in 1.5 - Results.txt [TABLE][sIGPIC][/sIGPIC]| Join MOOSE community on: DISCORD :thumbup: Website of the MOOSE LUA Framework. MOOSE framework Downloads. Check out Example Missions to try out and learn. MOOSE YouTube Channel for live demonstrations and tutorials. [/TABLE]
Grimes Posted November 11, 2015 Posted November 11, 2015 Very methodical tests. :) The behavior is pretty much what I've replicated for the bug report, though I did it without sams ^_^ Will add the reaction to threat information to the bug report. Also I noticed the Ka-50 vikhers, thanks for reminding me. The right man in the wrong place makes all the difference in the world. Current Projects: Grayflag Server, Scripting Wiki Useful Links: Mission Scripting Tools MIST-(GitHub) MIST-(Thread) SLMOD, Wiki wishlist, Mission Editing Wiki!, Mission Building Forum
FlightControl Posted November 11, 2015 Author Posted November 11, 2015 Hi grimes. Thanks for the feedback. The presence of SAMs makes ai even perform worse... So it is an important attention point for the developers. Otherwise the helicopters won't properly start evasive maneuvers and flares. Once the fix is there, we can reuse the test cases for regression tests. [TABLE][sIGPIC][/sIGPIC]| Join MOOSE community on: DISCORD :thumbup: Website of the MOOSE LUA Framework. MOOSE framework Downloads. Check out Example Missions to try out and learn. MOOSE YouTube Channel for live demonstrations and tutorials. [/TABLE]
vicx Posted November 11, 2015 Posted November 11, 2015 I don't see how you be completely unhappy with Allow Abort outcome ... it seems like that setting is in there to encourage SEAD flights to be effective. If a player is flying SEAD and doesn't do the job then down stream effectiveness of mission elements fails. This is kind of what you want. Of the choppers only the KA-50 should be partially effective in this instance and it looks like they are. Again not perfect simulation but in general not a bad result. As for individual AI unit performance in DCS. :noexpression: No comment.
FlightControl Posted November 11, 2015 Author Posted November 11, 2015 Hi vicx, I don't understand your feedback. I think you miss the point of the report. The ka-50 performs very bad also in certain cases. The issues are that helicopters should be able to engage at safe distance, regardless of the threat reaction settings when they are out of the engagement zone of the enemy. And they don't seem to do that. Both the KA-50 and the MI-24N should engage from a safe distance. That is what they were designed for. [TABLE][sIGPIC][/sIGPIC]| Join MOOSE community on: DISCORD :thumbup: Website of the MOOSE LUA Framework. MOOSE framework Downloads. Check out Example Missions to try out and learn. MOOSE YouTube Channel for live demonstrations and tutorials. [/TABLE]
vicx Posted November 11, 2015 Posted November 11, 2015 Sorry, I mean to say that where there is buggy behaviour it is not always clear where the problem lies. You might need to split up the report and that is what Grimes has done.
FlightControl Posted November 11, 2015 Author Posted November 11, 2015 (edited) Vicx, What split would you suggest? You are right, buggy behaviour needs root cause analysis. That is the reason why the scenarios were made with the report. Just note that I know the simulator very well, been in it for years, designing and scripting many missions. As such, I have a good feeling on what I can expect the AI to do and what not. During the history of dcs (from version 1.1), I've been helping dcs with these kind of reports detecting and reporting bugs, as many others did. That is why there is a beta. DCS will decide together with the moderators what to fix and how. But it is great that the guys leave us a possibility to test and improve stability and behaviour before release. I hope they appreciate the time we spend on it, as it may become useful for others and for themselves. Edited November 11, 2015 by FlightControl [TABLE][sIGPIC][/sIGPIC]| Join MOOSE community on: DISCORD :thumbup: Website of the MOOSE LUA Framework. MOOSE framework Downloads. Check out Example Missions to try out and learn. MOOSE YouTube Channel for live demonstrations and tutorials. [/TABLE]
Recommended Posts