Stealth
- Jwk the Hemp Monkey
- Offline
- Addict
22 years 1 month ago #2838
by Jwk the Hemp Monkey
Replied by Jwk the Hemp Monkey on topic Stealth
Elite mod does that. There has been more than one occasion where I have seen some fat juciy ferighters...charged head first with guns blazing only to find 6 patcoms waiting for me.
Please Log in or Create an account to join the conversation.
22 years 1 month ago #2839
by jojobird
Regards, Sergej
Well indeed, it does, but the prolem of elite is that it is one sided. it changes the sensed brightnesses for you, but does not change how others see you... I would like it even for everyone
Regards, Sergej
Please Log in or Create an account to join the conversation.
22 years 1 month ago #2845
by Shane
A couple of things that have changed my viewpoint on stealth:
When I changed my Marauders (brightness at 0.3, min brightness at 0.1)I noticed I could not pick them up on my scanners unless I was within five kilometers. However, they left to engage me (I could see the ships out the window) at just under 15 kilometers, firing missiles. Therefore, while they did not show up on my scans, I obviously showed up on theirs...they could not have targeted me without me showing up.
At least, that's how it would be if the flux engine played fair...
Last night I was playing around with the HackREM mod and generally being a nuisance to the shipping lanes. iFleet was still active, so I decided to see what I could do with the two of them (alot, let me tell you!). I had hijacked a large ship by docking to it and REM'ed it across the system (Hoffer's Wake; I took the ship at the Alexander L-point). After running around for a little bit I decided to see if the iFleet mod still worked. I only had two marines on my ship, so I REM'ed the freighter (with my ship still docked to it) to the Twinkle Underworld base. A good long distance in LDS. I then disengaged from the REM link, undocked from the freighter, and while it sat there, I docked with the base, and bought marines. Then I redocked with the freighter and tried to capture it.
Well, as soon as you do that any ship in the freighter's group should become an enemy, but, those ships were back at the Alexander L point, and there was no way I was still on their screens, so I was safe.
I was really surprised when they showed up (about 3 minutes later) LDSing in after me. This shows there are no scan limitations or distance limitations if the flux engine decides the AI ship is to engage the enemy.
This means that the EoC program has a 'god-like viewpoint', and knows where every ship currently created is, and what it is doing. Stealth cannot blind the computer's eyes. The player, however, does not know what's out there, so stealth can exist for him, but only from his standpoint... AI ships can be hidden from your scans, you cannot be hidden from theirs. Unless the program is instructed to act as if the player is not detected under certain circumstances (like the 5 km limit in the Hide and Seek mission)there is no stealth in the game.
In one night I've completely reversed my first assessment. Stealth can only exist in the game if:
1) The player is decieved by the stealth capabilities of the AI ships.
or
2) The engine is coded to cause the AI ships to act like they don't detect the player ship.
Anyone see anything I missed?
When I changed my Marauders (brightness at 0.3, min brightness at 0.1)I noticed I could not pick them up on my scanners unless I was within five kilometers. However, they left to engage me (I could see the ships out the window) at just under 15 kilometers, firing missiles. Therefore, while they did not show up on my scans, I obviously showed up on theirs...they could not have targeted me without me showing up.
At least, that's how it would be if the flux engine played fair...
Last night I was playing around with the HackREM mod and generally being a nuisance to the shipping lanes. iFleet was still active, so I decided to see what I could do with the two of them (alot, let me tell you!). I had hijacked a large ship by docking to it and REM'ed it across the system (Hoffer's Wake; I took the ship at the Alexander L-point). After running around for a little bit I decided to see if the iFleet mod still worked. I only had two marines on my ship, so I REM'ed the freighter (with my ship still docked to it) to the Twinkle Underworld base. A good long distance in LDS. I then disengaged from the REM link, undocked from the freighter, and while it sat there, I docked with the base, and bought marines. Then I redocked with the freighter and tried to capture it.
Well, as soon as you do that any ship in the freighter's group should become an enemy, but, those ships were back at the Alexander L point, and there was no way I was still on their screens, so I was safe.
I was really surprised when they showed up (about 3 minutes later) LDSing in after me. This shows there are no scan limitations or distance limitations if the flux engine decides the AI ship is to engage the enemy.
This means that the EoC program has a 'god-like viewpoint', and knows where every ship currently created is, and what it is doing. Stealth cannot blind the computer's eyes. The player, however, does not know what's out there, so stealth can exist for him, but only from his standpoint... AI ships can be hidden from your scans, you cannot be hidden from theirs. Unless the program is instructed to act as if the player is not detected under certain circumstances (like the 5 km limit in the Hide and Seek mission)there is no stealth in the game.
In one night I've completely reversed my first assessment. Stealth can only exist in the game if:
1) The player is decieved by the stealth capabilities of the AI ships.
or
2) The engine is coded to cause the AI ships to act like they don't detect the player ship.
Anyone see anything I missed?
Please Log in or Create an account to join the conversation.
22 years 1 month ago #2846
by jojobird
Regards, Sergej
Shane in my oppinion you are correct but there is one little thing.
This is pure specualation and unless we get someone who wrote the game we are certainly not going to get anywhere.
In my experience with AI it is usually a split personality:)
What I'm saying is that when you remhacked it you gave it a situation wich could never be predicted. The general instruction says that when one of the convoy is attacked the "group AI" instructs the "ship AIs" to attack you and then these ships ask the "game AI" where to and so they go there.
That is obviously computer knowing everything.
But my point is that if the AI is hierarhical (only IF) the ship itsel could still not pick you up in the first place until you engaged it and during combat it would use its own sensors. Otherways I don't see the point of individual ships having sensors.
The "game AI" must know all the positions for the game to function. Otherways it would be impossible to formate for example through LDS. I've had my wingmen catching me after a full minute in LDS and then formate with me. Somehow special tules apply in cases like that one.
I still believe that it would be possible to limit the targeting in combat to ships that are in ship AI's contact list. Even in Shane's case it seems logical. Freighter is attacked and it sends out a distress signal but when actual ships arrive you could wait somewhere cloaked and they would arrive and not see you.
If you have time you could change the experiment a bit... Do everything as before but after attacking the ship and becoming enemy fly away... to lets's say 500 km and droip a waypoint there (UNIGUI OR SOMETHING) or something then wait... Will the ships attack you or will they go to the transport and try to protect it and only attack targets nearby (incapacitate transports drive if you can)
Theese are just a few ideas of mine and i'm not claiming anything is correct... I wish there was someone who actually wrote the AI on engagement and tell us all about it:) Nevertheless if there is an able modder out there I believe this could be added to the game...
This topic has become less practical and more theoretical but it is interesting nevertheless. So if anyone can help us understand stealth and brightness issues please do. It would be a great tactical element in the game if correctly done.
Sergej
This is pure specualation and unless we get someone who wrote the game we are certainly not going to get anywhere.
In my experience with AI it is usually a split personality:)
What I'm saying is that when you remhacked it you gave it a situation wich could never be predicted. The general instruction says that when one of the convoy is attacked the "group AI" instructs the "ship AIs" to attack you and then these ships ask the "game AI" where to and so they go there.
That is obviously computer knowing everything.
But my point is that if the AI is hierarhical (only IF) the ship itsel could still not pick you up in the first place until you engaged it and during combat it would use its own sensors. Otherways I don't see the point of individual ships having sensors.
The "game AI" must know all the positions for the game to function. Otherways it would be impossible to formate for example through LDS. I've had my wingmen catching me after a full minute in LDS and then formate with me. Somehow special tules apply in cases like that one.
I still believe that it would be possible to limit the targeting in combat to ships that are in ship AI's contact list. Even in Shane's case it seems logical. Freighter is attacked and it sends out a distress signal but when actual ships arrive you could wait somewhere cloaked and they would arrive and not see you.
If you have time you could change the experiment a bit... Do everything as before but after attacking the ship and becoming enemy fly away... to lets's say 500 km and droip a waypoint there (UNIGUI OR SOMETHING) or something then wait... Will the ships attack you or will they go to the transport and try to protect it and only attack targets nearby (incapacitate transports drive if you can)
Theese are just a few ideas of mine and i'm not claiming anything is correct... I wish there was someone who actually wrote the AI on engagement and tell us all about it:) Nevertheless if there is an able modder out there I believe this could be added to the game...
This topic has become less practical and more theoretical but it is interesting nevertheless. So if anyone can help us understand stealth and brightness issues please do. It would be a great tactical element in the game if correctly done.
Sergej
Regards, Sergej
Please Log in or Create an account to join the conversation.
22 years 1 month ago #2850
by MrZUK
well Hi there,
somebody emailed me about my mod (Firepower Mod V0.4 is the latest)
I am not sure whether it does affect the AI, I would like some feed back though! <img src=icon_smile_8ball.gif border=0 align=middle>
somebody emailed me about my mod (Firepower Mod V0.4 is the latest)
I am not sure whether it does affect the AI, I would like some feed back though! <img src=icon_smile_8ball.gif border=0 align=middle>
Please Log in or Create an account to join the conversation.
22 years 1 month ago #2851
by Flamineo
Yet more speculation...
AI ships don't have a 'contact list' in the same sense as the player. Hamsterlord and others have done some testing with player sensors on NP ships; consensus was that the AI won't initially see a player until it 'should', but after that the attack/pursue/defend/run code is omniscient.
Last point I gave up on this area was trying to get any meaningful return value from the BrightnessOf pog function. You feed it two sims, a distance, and it returns brightness of one from the other at said distance using linear or squared attenuation. Unfortunately, as far as I could tell, it returns the queried sim's exact brightness at zero distance, and 0.0 at any other . All sims do have a brightness FloatProperty, so it might be possible to write a function to get attenuated brightness according to custom rules, then attach an AI handler to every NP ship to simulate responses to it. To establish a custom formula that makes NPS sensors equivalent to the player's, you'd have to do some pretty painstaking testing of player sensors' cut-off distances for targets of various brightness.
Incidentally, some of the peculiar behaviour of stations regarding visibility occurs because their brightness is -1.0, which appears to be a special case meaning 'always targetable'. Even to the player, sims within ~5-10km are always visible as contacts according to StandardSensorVisibility, regardless of whether their brightness is below the supposed minimum sensed brightness for the sensors the player is using (minimum sensed brightness is why low-grade Elite mod sensors are useless at the ranges in their ini files; the ~5-10km effect is why they work at all).
One other difficulty: player brightness changes with heat and power output, but NP ships don't have any power or heat modelling, so their base brightness is fixed. I don't believe this can be changed in the current engine, although you might be able to write a pog layer to fudge it somehow -- the brightness FloatProperty is read/write.
AI ships don't have a 'contact list' in the same sense as the player. Hamsterlord and others have done some testing with player sensors on NP ships; consensus was that the AI won't initially see a player until it 'should', but after that the attack/pursue/defend/run code is omniscient.
Last point I gave up on this area was trying to get any meaningful return value from the BrightnessOf pog function. You feed it two sims, a distance, and it returns brightness of one from the other at said distance using linear or squared attenuation. Unfortunately, as far as I could tell, it returns the queried sim's exact brightness at zero distance, and 0.0 at any other . All sims do have a brightness FloatProperty, so it might be possible to write a function to get attenuated brightness according to custom rules, then attach an AI handler to every NP ship to simulate responses to it. To establish a custom formula that makes NPS sensors equivalent to the player's, you'd have to do some pretty painstaking testing of player sensors' cut-off distances for targets of various brightness.
Incidentally, some of the peculiar behaviour of stations regarding visibility occurs because their brightness is -1.0, which appears to be a special case meaning 'always targetable'. Even to the player, sims within ~5-10km are always visible as contacts according to StandardSensorVisibility, regardless of whether their brightness is below the supposed minimum sensed brightness for the sensors the player is using (minimum sensed brightness is why low-grade Elite mod sensors are useless at the ranges in their ini files; the ~5-10km effect is why they work at all).
One other difficulty: player brightness changes with heat and power output, but NP ships don't have any power or heat modelling, so their base brightness is fixed. I don't believe this can be changed in the current engine, although you might be able to write a pog layer to fudge it somehow -- the brightness FloatProperty is read/write.
Please Log in or Create an account to join the conversation.