Jump to content

Exorcet

Members
  • Posts

    3825
  • Joined

  • Last visited

  • Days Won

    5

Everything posted by Exorcet

  1. I second TWS Auto in the Hornet. It will center on the target you're tracking making things a lot easier. In general the F-18's controls are sometimes lacking and a bit unintuitive in my opinion. If I'm not using TWS I'll just STT on the biggest threat.
  2. More or less the same thing. The real world testing I mentioned was level release, 1 g, at 300 ft. I had to fly up to 1000 ft to keep the Sparrow from crashing. Granted the test plane was a F-4 in real life, not a F-15 which I used in DCS, but I would expect a similar launch envelope between the two.
  3. I noticed when trying to replicate real world Sparrow tests that low altitude AIM-7's would crash into the ground, I suppose this is why.
  4. The F-20 was basically a fully designed plane as far as I know. It just didn't get any sales. It's not much different from the Su-25T or Ka-50 in DCS, which barely existed.
  5. The AIM-7 isn't the most reliable missile, especially the older versions. You should be prepared for it to lose track. Sometimes you can see this happen visual by watching the smoke trail. The other tool you have is the missile impact timer on the HUD. It's not perfectly accurate, but if you take a shot within correct missile range it will give you a good idea of how close the missile is to the target. If it reaches 0 and the target is still flying a few seconds later, fire again. Also don't underestimate the range of the AIM-9. Under 7 miles, depending on altitude, be prepared to fire one head on if Sparrows fail. Fly high and fast, this helps your missiles. Right now the limit on height is the limited look down capability of the F-18's radar, so try to keep the enemy less than 5 degrees or so below you. You can launch early to make the enemy go defensive, but if you do you'll probably want to fire a second AIM-7 that actually has a decent chance to hit. After you fire your first missile, crank (put the target at your radar gimbal limits) and dive to defend against enemy missiles, even if you don't get a launch warning. Pull your nose back on target to fire additional missiles and then go back into the crank. If you do end up switching to IR missiles, remember that you don't have to keep lock after firing. Launch distance information is on the HUD/radar in the form of the Launch Acceptable Region (LAR) It's a range scale representing missile range (look in the manual for more detail). Use this to determine when to shoot.
  6. Sadly it's the nature of AI. We can't give the AI PFM flight models, nor can we code artificial brains that 100% mimic human behavior. The AI gets simple models that take shortcuts. It's the nature of simulation. This is what the preplanning option is for. You (the person flying the mission in DCS) align Jester's priorities with the mission breifing before you takeoff. Again we need a priority system here. If the targets are the transports, then they get priority unless the other contacts are a threat. Jester would ideally not focus the radar on one group over another without a reason (ie attacking the mission objective or self defense). Humans don't cheat, but AI must. If they didn't we would just be flying by ourselves. If you don't like the word cheat, think of it as simplification, shortcut, or modeling. We can't model a human brain, only lines of code that externally look sort of like a human actor. You're right that guessing and experience is not coded* because it's hard, but it's also unnecessary to copy 1 for 1. *If you really wanted to make the best RIO possible maybe you could try using learning algorithms to come up with an AI program. This might become really good at emulating human behavior, but it still needs to run on our computers and fit within developer budgets. It's fine, not everyone agrees on everything, though I'm completely on board with what you said. However the cheating and dice rolling isn't a problem when done right. In fact, it's the only option we have.
  7. I was replying to multiple posts, more than one of which used the term script. Replying to your post specifically I was trying to address what you brought up: The point of a preplanning mode for Jester would be to give the AI context on mission tasking, RoE, success conditions, etc. Regarding your questions, this goes back to what I said originally. AI sometimes has to cheat. As long as this doesn't come with blatant unfair advantages, it's OK for the AI to cheat to mimic human decision making/intutition. Q1 - There are a few ways to handle this. The easiest is to make Jester all knowing when it comes to target ID, like most AI in DCS. A step up from this would to be base target ID on parameters like speed, altitude, RCS, and/or location. If we know the enemy is sending C-130's at 250 knots/15000 ft, then Jester knows to prioritize searching for radar contacts matching that information over fighters at 35000 ft and Mach 1. Q2 - This is a bit vague, and there is no real target priority given here. One thing I can say is that I dislike Jester's tendency to laser focus on TWS tracks, even if they're miles away, when there are closer targets to engage. In general, shorter range should increase priority. Q3 - This is similar to 1, the AI will probably have to cheat a bit and just know where/what the target is if it reappears soon enough. If it vanishes from radar for a long time, then there should probably be a reset function and Jester will then have to reattempt ID when it next appears based on speed/alt/etc. One thing to keep in mind is that I don't know how Jester/DCS is coded, so these are just examples. They may or may not be feasible as described, but I hope they get the idea across.
  8. The option is "Visible" but it's not an acquisition requirement, it's a tickbox that makes the target visible to the AI even if it's beyond the AI's sensor's ability to detect it. In other words it's a cheat to force the AI to be able to find its target. Visual ID options for RoE is something we really need given how often it was used in historical combat. It shouldn't be too hard to add. All AI aircraft have a visual sensor modeled. I have a simple script for testing it: env.setErrorMessageBoxEnabled(false) -- Prevent error popup local _group = Group.getByName('AWACSMain'); if not (_group and _group:isExist()) then return false; end local _controller = _group:getController(); local _targets = _controller:getDetectedTargets(Controller.Detection.RADAR); for _index, _target in ipairs(_targets) do if _target.visible then return true; end -- I see the enemy!!! end return false; -- I don't see the enemy... You can change the Controller.Detecton.RADAR to be different sensors (RWR doesn't seem to work?). Visual detection when I lasted tested on an AI MiG-29 was about 3 miles distance on me and limited to the sides, front, and top of the canopy.
  9. Where are your based air defenses? Also this shouldn't really happen if you using the advanced waypoint options like Search then engage. If you use the range setting included in those tasks the AI should stop chasing targets at 150% of the range set. I'm not sure what this, though the AI will usually stop what it's doing if it's being shot at by a missile.
  10. Maybe I wasn't clear enough. I don't want a scripted RIO. I want to be able to give Jester context so that it's possible for him to make some of the complex decisions that a RIO would make. I don't like scripted events and missions. The ones I make I try to make as unpredictable as multiplayer. However each mission still has a goal and that goal is going to influence tactics. Right now Jester has no way of knowing what the goal is and won't adapt to your mission. That's one of his flaws in my opinion. The way to change that is to provide a way to tell him what the mission and the goal is. This won't be a step by step scripted list of events. It would instead be something like a set of priorities. And of course it would be optional so if you don't want to set it up, you get what we have now. For example if your mission is preventing transports from reaching a certain airbase, Jester would avoid committing on fighters until they are a threat to you. Instead of focusing the radar on the first thing he sees, he's try to keep searching until he finds a transport, or some non transport flights become threatening. On the other hand if you're flying escort, Jester would prioritize fighters over transports. The system would not be "Jester AI task push - Attack Group "Ariel 1"" with the visual checkbox enabled. I hope this clears things up a little.
  11. In a way it makes sense, but I think something is being left out. Not everyone can get a human RIO. Jester is pretty much my only option with how and when I play and how he compares to a human AI really doesn't matter to me as a single player pilot because I can't choose another RIO. And to be honest, even if Jester was a 99 percentile RIO in every respect, he still couldn't replace human RIO unless he was linked to cleverbot or something and could hold conversation while navigating to the target or speak with me during a mission briefing to discuss how we're going to get through enemy defenses. Multicrew sells itself. It doesn't need to be artificially inflated by a limited AI RIO, and I think attempts to do that just end up frustrating those that can't easily access multicrew. Again I kind of get it, but I don't think the experience I have with the Tomcat is what you were aiming for. I've never wanted Jester to replace a second human crew, I just want him to not have immersion breaking faults, even if he's only average at what he does. I've actually asked for similar functionality from AWACS in DCS. As long as it works well*, it will only add to the sim. It's all configurable anyway, not everyone has to fly the same way or at the same level of realism. Even one specific player might go full simulation one day but the next turn on infinite missiles and just go for a joyride. So make it an option and we're set. *Jester kind of does this as he'll select targets as he pleases. The problem is you can't really get him to follow a mission briefing, or sometimes any kind of self preservation. I've had situations where he will ignore escort fighters that are very much a threat and become focused at lobbing missiles at things a hundred miles away. Then I've also had situations where we could bypass enemy fighters and launch at a transport or something, but Jester just won't do it because he has no way of knowing that the Transport is the mission objective. In these cases, I override whatever Jester tries to do. I have to be honest, I'd like to avoid using the Jester menu as much as possible. I think instead of a menu option, some kind of mission pre planning needs to be available. We need to be able to tell Jester what the mission is, what the objectives are, and how to prioritize targets. All ahead of time. Then when we're in the air Jester wouldn't need to be told what to do (within reason, I'm not trying to eliminate communication entirely, but I think it's reasonable for Jester to know to connect with a specific datalink based on mission brief or to focus more on fighters than bombers to target for example) and what he does would align with the player's understanding of the mission.
  12. AI seems to have better chaff/flare than players. The only release mode they have is to lazily deploy a couple CM's every second or so, but this is enough to distract even the more resistant missiles like 9X and R-73 while flying in AB. It would be nice if the playing field was more even and the AI would have to use different CM programs. I suppose the slow release used now is fine for preflaring, but if the AI sees a missile incoming they should change their release program to something much more aggressive.
  13. AI is hard. Jester doesn't really have common sense. When things go as expected according to how he's coded, he usually does an OK job. Anything outside of the coding can produce frustrating results. I think Jester needs to "cheat" a little more like enemy AI do. Of course too much cheating is bad, but AI needs to cheat a little to compete with humans. If Jester loses a target he should know where to look, roughly, to reacquire. Jester should be able to see and lock targets outside of the current radar FoV if it makes sense (like the situation described above with enemy aircraft taking off). There are too many situations where Jester just throws up his hands at trivial problems. Again, AI is hard so it's not like this is because HB is slacking, but I'd like to see some changes to make Jester act more like a human RIO. The F-14 being a 2 seater is what holds it back the most in DCS.
  14. This is a bit citation needed. Between multicrew and added weapons/capability this could sell just as much as a USAF F-16. I don't think it's easy to predict how they would compare. The same goes for Hornet and Super Hornet and Eagle vs Strike Eagle. The F-5 isn't an equivalent to the F-18. It's a low cost export fighter that didn't really see use by the US. The MiG-21 is maybe a bit more well known, but it's Russian. The real test would be seeing F-4 sales numbers as the F-4 does have a place similar to that of the F-18.
  15. https://forum.dcs.world/topic/299207-enemy-pilot-vision/?do=findComment&comment=4952431
  16. That would be the way to go. Given DCS's nature as a sim, I don't know why this is so often opposed. To me it goes hand in hand with non historical scenarios. Of course there is the developer work load to consider. There are an infinite possible number of "field modifications" that could have been so we can't have them all, but the inclusion of a prototype weapon or something similar should be easy enough. As stated though it needs to be clearly separated from the historical weapons via ME checkbox or something like that.
  17. You could build many convoys and only choose one to activate. This is also probably better than a purely random generator because one of those could pick an immobile SAM, a building, troops, etc and ruin the convoy.
  18. It should be the other way around shouldn't it. Units should be hot because they are active. The checkbox should be to make them cold since that would be less common.
  19. They're not. Subject to change. The features list was never set in stone, this should be well known by now.
  20. It's an easy problem to fix. Random failures already exist, they just need to be turned on. The same should be true for randomly variable airframes.
  21. Hopefully this is a simple request, the ME currently lacks the ability to assign groups into larger collections of units like a package. Such grouping options could be helpful in terms of mission organization (providing a tickbox for setting a frequency for an entire package at once instead of doing it group by group for example) and in terms of triggers. Consider a situation where you want an attacking force to retreat when it takes 50% casualties. Currently the only triggers supporting this are Group Alive % triggers, but this can be tricky to apply across multiple groups. If we could consolidate many groups into a package or supergroup, then we could have a trigger such as Package Alive Less Than X%, etc, making the tracking of casualties much easier.
  22. Improvements to DCS ground AI have been wanted for a long time, but one challenge to implementing this is getting the AI to understand what is supposed to be happening in the wider battlefield. I don't think it's realistic for them to figure things out on their own without guidance. In order to provide the AI some guidance I've come up with the idea of Objectives that can be passed to ground groups. The objectives would have a position and radius. When a group is assigned an objective, probably by triggers or waypoint actions, the group will advance to the objective automatically and seek out enemies within the specified radius. The groups will probably want to use roads by default as much as possible but there should also be an option for exclusive off road driving (perhaps the roads are mined or heavily defended) if the AI can path find around/through forests and terrain. When all enemies in the objective radius are destroyed, ideally the group should form some kind of defensive formation around the objective area. Example image: Note: In the image the objective belongs to blue side, but it may make more sense to have objectives be accessible by both coalitions. Expanding on the image above, the blue group on the left would be assigned the objective over the city. This would cause them to automatically advance to the objective and destroy all enemies in the objective radius (1 red group on the map). When finished, the blue group would then take up a defensive position around the objective to prevent red forces from entering. Objectives wouldn't just be offensive in nature either. Let's say in the example pictured above there was a second objective to the west away from the city. It could serve as a safe area that is either far away from the front lines or defended by additional friendly forces. This objective could be activated if the attacking blue force takes too many casualties, which would result in them retreating and abandoning "Blue Objective 1". Another use for defensive objectives might be as hiding spots for convoys. Currently the AI will stop and spread out when attacked. We could have the option to set objectives along their route that are located in forests or cities. When attacked, instead of stopping immediately, the convoys would head for the nearest objective (probably needs new trigger AI Task Push Nearest Objective for example) and hide or fortify there.
  23. This seemingly got lost at some point, but it's still an issue. I understand the request for specific info, but I really think the issue is every port. 99% of them have some area where the depth is as low as 7-10ft. You could nearly stand in them safely, yet they have static tankers as scenery in the sim. @draconus listed some ports in the post I quoted earlier: https://en.wikipedia.org/wiki/Port_of_Jebel_Ali https://en.wikipedia.org/wiki/Mina_Rashid US SuperCarrier at Dubai Jebel Ali: https://nara.getarchive.net/media/the-aircraft-carrier-uss-george-washington-cvn-73-pulls-into-port-jebel-ali-5d9757 In addition the port at Fujairah has a depth listed of over 50 ft: https://en.wikipedia.org/wiki/Port_of_Fujairah Like most ports on the map it can reach around 10 ft min, barring any passage by a large or even mid size ship I also found a quick list of port depths, not sure if this is accurate but it makes more sense to me than some of the DCS depths: And I have a mission with various boats trying to leave ports around the DCS map. They fail in 99% of cases that aren't the patrol boat which can get through 1-3 ft of depth it looks like. The only non patrol boat to be able to leave a port is a Combate from the Bandar Abbas port Hormuz Shallow Shipping.miz
  24. If HB has modules A, B, C planned, the modern F-4 can go to the end, safely preventing the delay of anything. If it's a new developer that's going to be making the module and they wanted to make a modern F-4 in the first place or have info on it already, well it would make a lot of sense for them to pursue that module over something else and we lose nothing. And even if we encounter what I guess you might call the worst case where a modern F-4 replaces something more "unique" well, I don't see the big deal. It's not like having somewhat similar modules will degrade the quality of the simulation. They're also no so similar as to be redundant either. Asking for plane has zero cost, opportunity or otherwise. It just shows that there is demand for it. Now if we want to get into trying to schedule releases and ranking planes by priority, then what you're saying might apply. That doesn't leave you with much to worry about though since it's typically the devs that have the final say.
  25. No I am not because I'm not demanding that anything be put aside for the sake of anything else. HB is free to prioritize however they like. They don't even have to be the developers to make a modernized F-4.
×
×
  • Create New...