Roadmap of Mi-24p ?

[Robo76]

If an enemy unit is unmasked by its own fire, it should immediately appear in Ai Petrovich's target menu. From a greater distance, it could be "unknown". If Ai were not at a sufficient distance to identify it, she should be aware of the threat of hundreds of projectiles flying at her from a given location, and register the threat as a general designation according to the danger, e.g. medium-range anti-aircraft systems, short-range anti-aircraft systems. Because when you attack an airfield full of AAA, there are bullet paths in the air, do you avoid them and Petrovich? He pretends to look for fire units, even though they are already shooting at his sight. In such a situation, I always wonder how much vodka Petrovich drank again and that he must have gone blind....  And then next time I'll take the Ka-50, where I can find the target myself.

 

Edited June 7 by Robo76

[Raffi75]

3 hours ago, Robo76 said:

So for another debater who apparently doesn't use Ai Petrovich :

1. AI doesn't even see the units already firing around the Mi-24 cockpit. If there are trucks next to the AAA firing, it refuses to target the AAA. I would expect the AAA firing to be at the top of the target list and not at the bottom (it often isn't there).

2. Ai cannot see the JTAC smoke signal. If the unit is near a forest or in a city, it doesn't matter if the colored smoke can be seen for miles. Ai "searches" all the trees in the forest and buildings in the city, she doesn't primarily look around the marked location, so she may find the unit after searching everything in sight.

3. Ai It cannot see and therefore will not target a partially damaged unit. I was killed several times by an M1A2, M2A2 and a Gepard, which had red health, but the AI did not see them even after repeated approaches until they shot me down with a machine gun / cannon.

4. Ai can swap an enemy target with a friendly helicopter. All she has to do is momentarily lose sight of the marked enemy and if another helicopter is nearby , Ai will continue tracking it. Ai won't even tell you that she's already tracking another target.

(If you want to argue that this doesn't happen to you, it's just that you don't use the Mi-24 for combat.)

You didn't understand what I wrote. In real life, and therefore in the game, the copilot has only his eyes at his disposal. So detection depends solely on skill. In contrast, Western pilots have various electronic systems to help them see further and detect more.

What you wrote about I also have in the game. Not often, but it does occur.

[Robo76]

27 minutes ago, Raffi75 said:

You didn't understand what I wrote. In real life, and therefore in the game, the copilot has only his eyes at his disposal. So detection depends solely on skill. In contrast, Western pilots have various electronic systems to help them see further and detect more.

What you wrote about I also have in the game. Not often, but it does occur.

But those AI "eyes" suffer from cataracts. And when I look out of the pilot's window myself, I see enemy units before the AI calculates and realizes. That can't happen when the AI operator looks out of the window the same way as the (player) pilot. That has nothing to do with realism, it's just errors in the AI, which you then pass off as an imperfect pilot-operator who makes mistakes, has "bad equipment" and is having a bad day.

Edited June 7 by Robo76

[Raffi75]

18 minutes ago, Robo76 said:

But Ai's "eyes" suffer from cataracts. And when I look out the pilot's window myself, I see enemy units before Ai thinks about it. This has nothing to do with approaching realism, it's just errors in Ai, which you then pass off as an imperfect pilot-operator making mistakes.

I have no problem communicating with Petro. I know his weaknesses and I know how to react. I am not defending him. The AI's bugs and suggestions have been reported, so I am waiting for them to be implemented. Until they are implemented, I have to react to what is happening myself. It is stupid, for example, that when I am flying in pairs and my wingman launches his missile at his target, Petro is able to give a message about the missile. When it comes to ground fire, I have no problem with that. I get information from which side they are firing. What annoys me the most is the lack of reselection of the same target after the first hit.

[Robo76]

2 minutes ago, Raffi75 said:

I have no problem communicating with Petro. I know his weaknesses and I know how to react. I am not defending him. The AI's bugs and suggestions have been reported, so I am waiting for them to be implemented. Until they are implemented, I have to react to what is happening myself. It is stupid, for example, that when I am flying in pairs and my wingman launches his missile at his target, Petro is able to give a message about the missile. When it comes to ground fire, I have no problem with that. I get information from which side they are firing. What annoys me the most is the lack of reselection of the same target after the first hit.

Whatever it is, I don't have the strength and time to prove it to ED for several days, and they won't touch it for several years anyway, with the excuse that it corresponds to realism as they see it. And if I don't like Ai, the most I can get from ED for dissatisfaction with their work instead of a fix is a threat of a ban from the discussion forum.

[Raffi75]

12 minutes ago, Robo76 said:

Whatever it is, I don't have the strength and time to prove it to ED for several days, and they won't touch it for several years anyway, with the excuse that it corresponds to realism as they see it. And if I don't like Ai, the most I can get from ED for dissatisfaction with their work instead of a fix is a threat of a ban from the discussion forum.

Relax, it's just a game.

[Ramius007]

Lack of target distance information is odd, it should be first info that pilot get after target detection, long before identification

[Weta43]

OT, but saying it anyway...

I've been coming here for a long time, & I've never seen anyone banned for pointing out a problem - or perceived problem - with the game.

What I have seen a lot of people banned for being d*cks in the way they go about pointing out those problems.

2 quite different things...

5 hours ago, Robo76 said:

And if I don't like Ai, the most I can get from ED for dissatisfaction with their work instead of a fix is a threat of a ban from the discussion forum.

 

[Weta43]

I agree about the gunfire thing - if Peter's boy can tell there's someone firing, it's not super difficult to work out if the projectiles are aimed at you & if they are, the offending unit should be on the target list

(Actually, skip that - the SIM knows without calculation, because THE SIM is omniscient. The code's just not passing the information along.)

Ordering the threats is probably the hardest part - is an infantry unit that's already firing at you more or less of a threat than the Strella you're still well out of range of - & is that more of a threat than the tank that is currently tracking you & that in 2 seconds you'll be within range of?

[Robo76]

12 hours ago, Weta43 said:

OT, but saying it anyway...

I've been coming here for a long time, & I've never seen anyone banned for pointing out a problem - or perceived problem - with the game.

What I have seen a lot of people banned for being d*cks in the way they go about pointing out those problems.

2 quite different things...

 

It is difficult to maintain a phlegmatic approach in a discussion when people from the ED team or their admins oppose, despite the evidence, arguments and testimonies of more people, that something is different than what they think in ED. When they don't know where to go and it's embarrassing, they admit the error or problem by looking at it - and then usually nothing happens for years (like the unbreakable landing gear on the F-16 reported a year ago, which I reported a year ago and ED initially denied). The cancellation of the stable versions of DCS was made by ED out of beta testers who pay ED and report problems for free in their free time. And as a reward, most " beta testers " encounter an arrogant approach to reporting with the justification that we don't have time for it now, because we are making the next great news here and there (which they then release full of often serious bugs) and they don't care about fixing old problems because they don't make money from it anymore. With every major update, most MP players pray that DCS will work at least as it always has and not paralyze MP. Then comes the standard ED excuse that the fix was very complicated and took longer, but that they have discouraged many players during that time and they still don't get it.  ED is supposedly just people. But so are we DCS players. If someone gets upset by this attitude and expresses themselves more strongly (not in a rude way), they will get demerit points and a warning of a possible discussion ban. The person who constantly excuses ED or defends how they are doing it right doesn't have a problem with such a debater, so you don't have to worry.

 

Edited June 8 by Robo76

[Raven (Elysian Angel)]

12 hours ago, Weta43 said:

Ordering the threats is probably the hardest part

Sure but for example door gunners' ROE can be set to "return fire" only, so that tells me the underlying logic is already present and functioning.

[Robo76]

12 hours ago, Weta43 said:

I agree about the gunfire thing - if Peter's boy can tell there's someone firing, it's not super difficult to work out if the projectiles are aimed at you & if they are, the offending unit should be on the target list

(Actually, skip that - the SIM knows without calculation, because THE SIM is omniscient. The code's just not passing the information along.)

Ordering the threats is probably the hardest part - is an infantry unit that's already firing at you more or less of a threat than the Strella you're still well out of range of - & is that more of a threat than the tank that is currently tracking you & that in 2 seconds you'll be within range of?

It is not necessary to rank the targets exactly according to their danger, It is enough if they are at all at the top of the Ai list and the player will be able to choose the one that is the most dangerous for his next planned attack. When attacking an airport, even in the current state of Ai, the order of the targets in the Ai list is not correct. For example, if you destroy a Hawk-TR or Hawk-CMD, the Hawk's launchers and search radar may no longer be at the top of the Ai menu.

 

Edited June 8 by Robo76

[LuseKofte]

22 hours ago, Robo76 said:

If George has a problem, he just needs to turn on the FCR radar, which can find and shoot at targets in the bushes.

yeah, in rotorhead server it is common FCR users kill off friendly units due to lack of situational awareness. I am not at all happy about Petro myself, I find him a bit simplified and easy solution when it come to problemsolving

[Weta43]

11 hours ago, Robo76 said:

It is not necessary to rank the targets exactly according to their danger, It is enough if they are at all at the top of the Ai list and the player will be able to choose the one that is the most dangerous for his next planned attack. When attacking an airport, even in the current state of Ai, the order of the targets in the Ai list is not correct. For example, if you destroy a Hawk-TR or Hawk-CMD, the Hawk's launchers and search radar may no longer be at the top of the Ai menu.

 

You say it's not necessary to order them exactly, but what you're asking for is for them to be ranked correctly +/- a little bit of wiggle room.

You're still asking for them to be ranked well enough for you to not be annoyed by their stupidity (which is what irritating you now...)

Getting the ranking correct (so the most dangerous are "all at the top of the Ai list") is the only trick bit, as the SIM already knows who's tracking who (including engagement parameters) and who's shooting at who & who's going to do the most damage if they hit seems both non-trivial, and not always what you want to happen anyway

For you - how would you rank threats in a way that's useful across a range of scenarios?

Should the AI show the 10 infantry that are already shooting at you at the top followed by the 30 that are tracking you but not yet firing, then the Shilka that you're in range of but that hasn't yet detected you followed by the Igla that you're out of range of but has detected you..

Or the Igla, then the Shilka, then the active infantry then the rest (armed, then unarmed)

If there are 100 units that could theoretically engage you, do you list them all, or trim the list to most threatening first?

If you trim the list, what if your mission is to destroy some howitzers that pose no threat to you, in the middle of a range of targets that do?

How long do you want your head down scrolling through targets to find the howitzers at position 132 (after everything with a gun & a few trucks)?

I agree Petrovich's listing isn't great now, but I don't think it's as easy to get right as simply telling someone to fix it.

That said, I hope they're actively trying to figure it out, not just consigning it to the 'too much effort for the increase in sales" bucket. It's one thing to sell things "as is, where is", it's another to say things will be finished to a particular standard in a reasonable amount of time, then not finish them because 'finishing' is more work than was initially thought.

[Robo76]

On 6/9/2025 at 3:09 AM, Weta43 said:

You say it's not necessary to order them exactly, but what you're asking for is for them to be ranked correctly +/- a little bit of wiggle room.

You're still asking for them to be ranked well enough for you to not be annoyed by their stupidity (which is what irritating you now...)

Getting the ranking correct (so the most dangerous are "all at the top of the Ai list") is the only trick bit, as the SIM already knows who's tracking who (including engagement parameters) and who's shooting at who & who's going to do the most damage if they hit seems both non-trivial, and not always what you want to happen anyway

For you - how would you rank threats in a way that's useful across a range of scenarios?

Should the AI show the 10 infantry that are already shooting at you at the top followed by the 30 that are tracking you but not yet firing, then the Shilka that you're in range of but that hasn't yet detected you followed by the Igla that you're out of range of but has detected you..

Or the Igla, then the Shilka, then the active infantry then the rest (armed, then unarmed)

If there are 100 units that could theoretically engage you, do you list them all, or trim the list to most threatening first?

If you trim the list, what if your mission is to destroy some howitzers that pose no threat to you, in the middle of a range of targets that do?

How long do you want your head down scrolling through targets to find the howitzers at position 132 (after everything with a gun & a few trucks)?

I agree Petrovich's listing isn't great now, but I don't think it's as easy to get right as simply telling someone to fix it.

That said, I hope they're actively trying to figure it out, not just consigning it to the 'too much effort for the increase in sales" bucket. It's one thing to sell things "as is, where is", it's another to say things will be finished to a particular standard in a reasonable amount of time, then not finish them because 'finishing' is more work than was initially thought.

enemy units in the AI menu is simple:

1.       If AI sees units, it will sort them by their firing range in the menu

2.       If an undetected unit within sight fires (anywhere), it is revealed and added to the target list and ranked according to point 1 (I wouldn't worry about who it's shooting at - although not me, it's shooting at a friendly aircraft in the area and mine is included in that too)

3.       If units important for the functionality of the entire group have been destroyed in a SAMs group , the remaining units of the group will move to the end of the list (before unarmed targets).

4.       On the command "check the target marked" - "orange", AI would check for the presence of units marked by JTAC with the nearest orange smoke. (This should not be automatic, because sometimes there are many marked places near the target and this would unnecessarily delay the search).

5.       Damaged units MUST be listed in order to be destroyed - most MP missions will only destroy a base if there are no enemy units there. (This is more of a bug where AI doesn't see heavily damaged units when they have some health in the red color.)

 

Using these rules, it will sort into groups in order:

 

EWR radar (does not shoot but is very important for situational awareness on the battlefield)

Long-range SAM (Patriot, S-300)

Medium-range SAM ( Hawk , Sa-6 Kub, Sa-11 Buk, …)

Short-range SAM (Roland, Sa-19 Thunguska , Sa-15 Thor, …)

SAM IR ( Linebacker , Avenger , Sa-13, …)

MANPADS

AAA radar (Cheetah, Vulcan , Shilka , …)

AAA ( Bofors , S-60, …)

Airplanes and helicopters on the ground

Armored ATGM

Armored

No armored

Infantry

SAMs remnants without control and guidance units

Unarmed targets (trucks, tankers, ....)

 

If these rules were introduced into AI , the main problems would be eliminated. I definitely don't want to create AI that serves the player a list of targets and he just stupidly destroys them one by one. When attacking multiple units, the pilot must always evaluate the threat and destroy it primarily based on its range. Disable the first Long range SAMs , medium range, short range Fox 2 and manpads , radar-guided AAA guns, radar-free AAA, armoured vehicles, infantry, other non-firing units. You need to use your wits, know what enemy units are capable of, don't throw yourself into a strong defence with kamikaze attacks, but gradually dismantle it according to its range.

 

If Ai doesn't see the infantry Manpad for example because it is a very small target, that's fine. Ai should register it every time it fires. The pilot must have a reconnaissance of the area from JTAC, or must preemptively drop flares in the hot area (on the Mi-24, the option is after 4 seconds). If Petrovitch can see the incoming missile, he can drop flares automatically.

There are usually up to 50 units defending the airport, with everything included, of which about half are more or less dangerous. And even if there were more, if they were arranged according to the rules above, it doesn't matter, they must be destroyed in order of range. It is up to the player-pilot's intelligence to judge whether it would be necessary to coordinate the attack with other players in different machines, or whether he is not enough to do it alone. I'm talking about MP missions, where the number of ground units is adjusted to the difficulty. If there were few units, several players would wipe them out in a few minutes. If there were more, one player would usually not be enough for them. That is why cooperation is necessary, like in real life.

Regarding the destruction of specific units in SP missions that are defended. You can try a mission created in this way as a challenge, but it has little to do with real combat. In reality, such missions are solved using the procedures I described above. The first attack wave overwhelms and destroys the Fox 1 SAMs , the next drops the GBUs from a height where they are safe from the Fox 2 SAMs . If an attack has to be made from a low altitude, all SAMs and AAA must be destroyed first . If one phase of the attacks cannot be completed, the other attack waves are called off due to high risk and then, after evaluation, the entire attack is repeated until the targets are destroyed.

Edited June 10 by Robo76

[Robo76]

Was this discussion thread started by someone from the ED team that worked on the Mi-24 ? You can't tell from the posts that ED is interested in the unfinished state they left Ai Petrovitch in. They are now dealing with other things that will make them money. Repairing Ai Petrovitch is not making them money, so it is way down the list of their plans, trz. that repairs can be expected at best in 1 year, but probably in 2-3 years. For newcomers to DCS, this realization may be shocking, as they won't encounter this approach from developers of other games elsewhere today, as similarly working developers have gone out of business.I don't want to be mean to ED, I'm as much of an aviation fan as they are, but without feedback, the DCS situation can't change (and then there won't be enough paying players either).

[BIGNEWY]

6 minutes ago, Robo76 said:

Was this discussion thread started by someone from the ED team that worked on the Mi-24 ? You can't tell from the posts that ED is interested in the unfinished state they left Ai Petrovitch in. They are now dealing with other things that will make them money. Repairing Ai Petrovitch is not making them money, so it is way down the list of their plans, trz. that repairs can be expected at best in 1 year, but probably in 2-3 years. For newcomers to DCS, this realization may be shocking, as they won't encounter this approach from developers of other games elsewhere today, as similarly working developers have gone out of business.I don't want to be mean to ED, I'm as much of an aviation fan as they are, but without feedback, the DCS situation can't change (and then there won't be enough paying players either).

Dear Robo 

The Op is not ED team. 

I get you are not happy and you may think development times are to long, or you may not like a the way something works, and you are welcome to give your opinion and share your thoughts. We don't ban people for sharing their thoughts or expressing discontent ( you will see plenty of it if you search ), we have rules which you can find at the top of the forum and they apply the same to everyone, please take a look at our 1.10 rule, you will get the idea, its very simple. We ask that you keep it respectful that is all. If you do have a problem with moderation than you should PM me or Nineline to discuss it, discussing it in a forum thread will get you warning points.

Petrovich will continue to get tweaks, the work being done helps with George and Bob AI also, and when we are ready for more updates they will be in change logs.

Myself and Nineline ( we are both associate producers and community managers ) interact here on the forum, social media, and discord, we chat with many people from all over the world, we are fans of DCS like you are, so all we ask is you keep it civil and try to understand that what we do at Eagle Dynamics is complex, it takes a lot of time for development which is why we have long early access periods, without them we wouldn't be able to bring these modules to you all and develop the core of DCS. You can see the work we do in our change logs and already this year the work being done is huge. We do try to get to as many posts as we can but it is not always possible, fortunately we have a lot of help from community members and also our volunteer testers who also report issues to the team directly. 

We have been making DCS for well over 16 years now and we continue to develop and continue to grow as a business, this can only be done with the support of the community.

thank you 

 

[kotor633]

Hi,

I'm still really curious as to when the best, proofread manual will finally appear in DCS...? Then another item on the Mi-24P roadmap would finally be ticked off.

[Robo76]

15 hours ago, kotor633 said:

Hi,

I'm still really curious as to when the best, proofread manual will finally appear in DCS...? Then another item on the Mi-24P roadmap would finally be ticked off.

Chuck's Guides

[kotor633]

Yes, I know Chuck's Guide.

[Raffi75]

14 hours ago, Robo76 said:

Chuck's Guides

You see, you didn't understand the question.

[Esac_mirmidon]

So we need to relief on Chucks shoulders EDs work on complete module manuals?

Dont get me wrong, Chucks manuals are excelent. But if ED is not going to provide one after so many years just offer him a job as an official partner for manuals. ED can dedicate resources and time in other things and we can have manuals from a dedicated and profesional fellow, quicker.

Like CurrentHill, or Massun92, i think an official partner with Chuck is a win win

 

Edited June 14 by Esac_mirmidon

[Robo76]

10 hours ago, Raffi75 said:

You see, you didn't understand the question.

But I understood and counted on the fact that these unofficial manuals are known by most people in DCS and I put the link for sure. Since they are available from the community, then there is no need to ask the ED for them. Why do something like this again. I'm much more bothered by the functional bugs of modules like Ai Petrovitch. It would have been much more sensible to leave it in its original state until the new version was completed to a reasonably usable state, rather than immediately go work on something else and abort the completion of Petrovitch. Then it's hard to find the time, during that grace period the programmer forgets enough things and has to fully understand the code again.

[Raven (Elysian Angel)]

35 minutes ago, Robo76 said:

Since they are available from the community, then there is no need to ask the ED for them.

Yes, there is. An official manual is an important resource and part of the module, therefore we paid for it, therefore it's far from unreasonable to insist ED provides one.
Their logic is probably that it's a waste of time and resources to write a manual while features are in flux and thus the manual would need to be re-written every time they change something. But that logic falls flat on its face when you consider how 3rd parties do things. If a particular feature is WIP, simply leave it blank in the manual, stating it's WIP and add it later. This leaves plenty of opportunity to write the parts of the manual for stuff that won't see significant changes and/or is considered complete 
Besides, it's probably a lot easier for developers to explain a feature when they just finished developing it and is thus fresh in their minds. Compare this to writing about something years later after you have been working on other modules as well...

[kotor633]

Please don't forget: Chuck creates his guides in his free time and on his own 'account'. And we're all glad he puts so much heart and soul into it. I think we should also mention TOViper, who is responsible for the Viggen manual. However, I get the impression that manuals are generally the 'unloved child'. That there are no uniform specifications. But that's going too far and in a different direction that I don't want to go in this thread.

If only they had treated the Mi-24 manual the same way as the F-16, FA-18, AH-64D, or Mirage F1... I've said a lot here, but my main message is: Please, ED Team, publish the manual for the Mi-24P asap. It's such a great module and deserves a good manual.