Hind tempting, obviously, but what exactly are we going to do with it?

[Ramsay]

2 hours ago, Johnr666 said:

What are the thoughts of Mi8 owners about buying the Hind?

If you like the idea of a Mi-8 "upgraded" with better weapons systems, ATGM's and R-60's it's a no brainer.

 

Start up and navigation is built round the Hind's pilot, so ergonomically the cockpit should be better than the Hip which is design for a crew of 3.

 

Not sure how restrictive the cockpit view will be or how easy it'll be to "throw around", I'm guessing it'll still have a lot of similarities to the Hip but with a different focus.

[cordite]

 

Direct answer to the thread title.

Edited June 4, 2021 by cordite

[Northstar98]

On 6/4/2021 at 2:23 AM, Johnr666 said:

I have the Huey, Gazelle and the Mi8 and enjoy them all because each is different in every way.

 

Should I buy the Hind which appears to be very similar to the Mi8 so may disappoint me.

 

What are the thoughts of Mi8 owners about buying the Hind?

 

Cheers

 

I'm using the Hip as a stepping stone for the Mi-24P.

 

Both aircraft complement each other basically perfectly for the Cold War, though while you can say that the Mi-24P is technically a post mid 90s/2000s RuAF one, if you weapons restrict you've basically got a Cold War aircraft.

Edited June 5, 2021 by Northstar98

[MAXsenna]

I have the Huey, Gazelle and the Mi8 and enjoy them all because each is different in every way.
 
Should I buy the Hind which appears to be very similar to the Mi8 so may disappoint me.
 
What are the thoughts of Mi8 owners about buying the Hind?
 
Cheers

I you like the Hip you'll love the Hind. Just get it!

Sent from my MAR-LX1A using Tapatalk

[Johnr666]

Thanks for all the replies.

 

The  more I learnt about the Hind the more I wanted it so I have pre ordered the Hind.

 

Although I expect many similarities between the Mi8 and the Hind, there appears to be enough differences to make flying the Hind a new experience.

 

I considered the Apache as an alternative to the Hind but it did not appeal to me. Maybe because the Apache is in very early development so it is not possible to demonstrate all its qualities.

 

Cheers

[Stratos]

The more I think about the DCS Hind missions, the more I feel we need a CRASH change in AI behaviour, a change to be released with the Hind, pending a more deep update, but such CRASH change will allow us to:

 

- Make DCS combat envinroment more realistic.

 

- To make missions more enjoyable.

 

How? Just two points.

 

- Reduce APC/IFV/MBT gunner 360º view, tracking hability and gunnery hability. Make it random, see you/don't see you, or make them the same as now but with a greatly reducedcone view. But don't make them a insta Sa-10 radar as they are now.

 

- Make infantry scatter/go prone (or kneel) when a enemy air unit is nearby. They keep firing and standing whit a rocket volley flying their way.

 

- A trench line or earth berm to put infantry would be amazing.

 

That's my idea, feel free to add your opinions.

 

 

[BitMaster]

I haven't fought in the Mi-8 for quite a while, last time I did it it a few rifle bullets could cripple the Hip. Is this still the case ?  If yes, I hope they fix it so it can accompany the Hind and hopefully the Hind can take more punishment.

 

Anyway, I am really excited to try it out, preordered

[Avimimus]

2 hours ago, Stratos said:

The more I think about the DCS Hind missions, the more I feel we need a CRASH change in AI behaviour, a change to be released with the Hind, pending a more deep update, but such CRASH change will allow us to:

 

- Make DCS combat envinroment more realistic.

 

- To make missions more enjoyable.

 

How? Just two points.

 

- Reduce APC/IFV/MBT gunner 360º view, tracking hability and gunnery hability. Make it random, see you/don't see you, or make them the same as now but with a greatly reducedcone view. But don't make them a insta Sa-10 radar as they are now.

 

- Make infantry scatter/go prone (or kneel) when a enemy air unit is nearby. They keep firing and standing whit a rocket volley flying their way.

 

- A trench line or earth berm to put infantry would be amazing.

 

That's my idea, feel free to add your opinions.

 

 

 

Those would be rather nice things. Also, making it so that more detailed modelling of rocket warhead fragments was done for this first ten or twenty rockets fired (there might be performance issues with doing so for firing all 80 - so it might make sense to switch to simplified modelling for particularly large bursts).

[Snails161]

I wasn’t going to bother with this module as i have most of the others but very rarely fly most of them (f14,f5,gazellebeing the exception) but i watched a hind walk round video on you tube and thought COOL, so just signed up :)

 

[BeastyBaiter]

Honestly, massive AI changes are not realistic expectation or request at this stage. For right now it is what it is. I do hope they redesign the AI at some point but that's likely a long ways off. I don't see it as the massive problem for the Hind as a few people in this thread though. I think the Hind will do just fine in the game as it is.

[Lurker]

1 hour ago, BeastyBaiter said:

Honestly, massive AI changes are not realistic expectation or request at this stage. For right now it is what it is. I do hope they redesign the AI at some point but that's likely a long ways off. I don't see it as the massive problem for the Hind as a few people in this thread though. I think the Hind will do just fine in the game as it is.

 

Massive AI changes are not needed. Small changes in behavior and damage modeling that would take a dedicated programmer maybe one day of work would go a long way towards making the ground based units in DCS a little more believable. We don't really need sweeping massive changes, give the AI some blind spots and slower reaction times (depending on the unit of course) and maybe even introduce a bit of error in their targeting calculations, introduce some kind of very simplistic fragmentation damage modeling to rockets that only affect infantry and unarmored units, and voilla: You have solved literally 85% of the problems people keep bringing up. 

[Avimimus]

28 minutes ago, Lurker said:

Massive AI changes are not needed. Small changes in behavior and damage modeling that would take a dedicated programmer maybe one day of work would go a long way towards making the ground based units in DCS a little more believable. We don't really need sweeping massive changes, give the AI some blind spots and slower reaction times (depending on the unit of course) and maybe even introduce a bit of error in their targeting calculations, introduce some kind of very simplistic fragmentation damage modeling to rockets that only affect infantry and unarmored units, and voilla: You have solved literally 85% of the problems people keep bringing up. 

 

Yes... although that could still be a month or two of work on the AI (to get it efficient, realistic looking, and debugged). The same goes for implementing a better fragmentation model.

 

These tasks might seem logical and solutions easy to imagine (e.g. me, writing below) but they may not be in practice - especially in such a complex piece of software... and development effort required is ever unpredictable.

 

Anyway, some thoughts:

 

The problem with rocket fragments is that they aren't perfectly evenly distributed. However, I could see that one could do the following instead of actually modelling fragments: Checking the distance of objects. Assigning a probability of them receiving damage based on that distance (i.e. of a fragment hitting). Apply damage if the vehicle fails that probability roll.

 

One could even have the table include the probability of receiving multiple fragments (and multiplying damage by the number of fragments)... so it'd be a single check per warhead.

 

If resources permitted one could even have the probability/damage distributions (over-pressure with an ~100% hit rate but a rapid decrease in damage with distance, small fragments which are more numerous, and large fragments that do more damage but 'travel' further) - all done simply by checking the distance vs. a probability of being hit by a fragment (and the amount of damage done by that fragment).

 

So one check per enemy per rocket - or if one wanted to model multiple effect warheads (e.g. Vikhr with over-pressure, main warhead fragments, and fragmentation belt) like just described... that'd be 3 checks per missile per enemy.

 

It sounds doable. However, I wonder if the devs might refuse to use such a layered probability approach for the simple reason that they eventually want to give ground units more advanced damage models (and want hit locations to matter). In that case rockets might remain without fragmentation models until computers get fast enough to give every ground unit a complex damage model and model every fragment from an 80 rocket burst... so ten years without improvements.

Edited June 7, 2021 by Avimimus

[Fri13]

56 minutes ago, Lurker said:

Massive AI changes are not needed.

 

It is really required. 

Not now on the Mi-24 release, but very soon. You can't have proper combat with the current ground units.

 

Quote

Small changes in behavior and damage modeling that would take a dedicated programmer maybe one day of work would go a long way towards making the ground based units in DCS a little more believable.

 

First you would need to make that behavior and damage modeling.

 

Quote

We don't really need sweeping massive changes, give the AI some blind spots and slower reaction times (depending on the unit of course) and maybe even introduce a bit of error in their targeting calculations, introduce some kind of very simplistic fragmentation damage modeling to rockets that only affect infantry and unarmored units, and voilla: You have solved literally 85% of the problems people keep bringing up. 

 

That is like 5% of the problems that there are. 

 

Do you know what would happen when the enemy aircraft is detected by air defense and its heading etc is found out?

The generic air warning is issued for the troops on the expected area.

 

That means, every single unit on that area will change their situation and prepare for a possible air alarm.

Vehicles are hidden, troops possible locations are changed, anti-air units are manned and put in full alert status, communications are prepared and warnings are followed.

 

Just to get the blind spots requires to have visual cones and simulated LOS searching and tracking:

 

Simulation between the crew members is required (delayed communication)

Simulation between platoon etc is required (delayed communication, misunderstandings etc)

 

 

The inaccuracy for the units to engage requires as well moral, stress, fear, obedience etc to be simulated, and that requires that every ground unit has basic training and tactics simulated so they know how to fight in simple form like when to open fire based to such simple things as commands, pre-planned sectors and ranges, enemy type and actions etc.

As often you do not want to open fire even when you have a change to shoot once of the targets as your commands are to staying undetected for a scouting enemy units etc.

 

We need proper behavior for the ground units, like they would stay inside forest waiting that enemy comes close, and then drive forward to acquire visual line of sight to enemy and engage them from concealment. Detach from the short engagement and possibly move to alternative position. Support infantry platoon they are attached to and perform more complex tactics than just shoot back and sit around. 

 

If someone wants just some suppression effects, they can just use simple scripts already available:

 

 

What we have now doesn't cut anymore (has not for a decade). Every unit is required to have already a proper stats and factors that are utilized for their behavior, that is based to real military tactics and command structure. That is too much to ask for now, and what Mi-24P will be fighting against for few years is what we have now. 

 

The dynamic campaign is expected to be released 2021, so it is possible we get some changes. 

Edited June 7, 2021 by Fri13

[Stratos]

Fri I agree with you about the massive changes needed, BUT I think a quick change is mandatory to fully enjoy the Hind. Something that need to be added NOW, not in some years. Something good is far better than something perfect the next year.

What we shouldnt tolerate is the poor status of the AI behaviour and the lack of frag damage.

[BitMaster]

2 hours ago, Fri13 said:

 

.........

 

The dynamic campaign is expected to be released 2021, so it is possible we get some changes. 

 

 

Tbh, that would be awesome. It's only 6 months till end of '21...I have my doubts it will happen that fast.

 

 

[MAXsenna]

Fri I agree with you about the massive changes needed, BUT I think a quick change is mandatory to fully enjoy the Hind. Something that need to be added NOW, not in some years. Something good is far better than something perfect the next year.
What we shouldnt tolerate is the poor status of the AI behaviour and the lack of frag damage.

I agree with@Fri13 too, very good points. But@Lurker has a point, even if he might be a little too enthusiastic about the development needed. I rather see some baby steps compared to a huge workover that will take ten years.

Sent from my MAR-LX1A using Tapatalk

[Fri13]

11 minutes ago, Stratos said:

Fri I agree with you about the massive changes needed, BUT I think a quick change is mandatory to fully enjoy the Hind. Something that need to be added NOW, not in some years. Something good is far better than something perfect the next year.

What we shouldnt tolerate is the poor status of the AI behaviour and the lack of frag damage.

 

The main change that would be required on day 1 is the units to have a simulated view zones and LOS. We already have the basic delays based skill level, as well we have proper turret turn rates etc. But we have every unit with a perfect view around them all the time and perfect information about every unit inside their detection range, and they react immediately when they get LOS on them. 

 

Someone said that you could surprise a unit (green state) from the rear as they can't see there, but I have no other information about that. 

 

But those are very minor things that would make something but nothing serious after sometime with any helicopter etc. 

 

 

[Stratos]

2 hours ago, Fri13 said:

 

The main change that would be required on day 1 is the units to have a simulated view zones and LOS. We already have the basic delays based skill level, as well we have proper turret turn rates etc. But we have every unit with a perfect view around them all the time and perfect information about every unit inside their detection range, and they react immediately when they get LOS on them. 

This point only, would make a dramatic improvement.

[cow_art]

10 hours ago, Stratos said:

This point only, would make a dramatic improvement.

 

^This and a quick patch to address rocket damage (perhaps just add a second, larger blast radius for non/lightly armored targets?) and I think I'd be a happy camper.

[Fri13]

12 hours ago, Stratos said:

This point only, would make a dramatic improvement.

 

Problem is that is such a big thing that doing it would make easy to implement almost all rest of the my listed features.

 

If we would add just the LOS system, it would mean that once someone in group detects an enemy, every single unit in the group knows that position.

So everyone would turn their gun on the spotted threat and end result is not better than before.

 

It would require a new communication system where units in the group would communicate properly and requires everyone to search the target by themselves even when one spots the target, this would mean that individual unit spotting a threat would only assist others to look for a generic direction to find it. 

So from example mechanized infantry platoon a one vehicle could likely be able see the unit. 

This would make every ground unit even worse as they would likely never spot anyone ever.

 

What requires that communication system is developed so that groups can share information between each other, so Early Warning radar issues the warning to units in area when it spots threat 200 km from their position and heading to that direction. The another unit between threat and the target area would hear the threat and issue warning to group so they know where to look and where to take a cover.

 

So now we are back to situation that units requires new AI that understands the terrain (trees, buildings etc) around it and knows how to take cover, issue a air alarm and prepare for self-defense by deploying infantry that has MANPAD etc. To get this the new AI is required that knows how it can perform such a action dynamically, and can it like middle of combat or on route to the combat zone etc.

Taking just a another route can fool the threat so much that it can't find the ground units in time and is required to leave the area back to base as 10 min fuel for loitering is not much.

And now suddenly proper damage modeling is required as well to go with everything. 

 

It is more like a package deal where everything needs to be done ready in basic level first, and then improve everything over time. Like first make the AI self-protection mode such that it knows where are trees and move under them to conceal themselves, and have a understanding how to unmount infantry and get the generic MANPAD units scattered to edges of forest. Later add buildings and more complex tactics. 

To do that, basic information network is required like now that everyone would know everything but just extend the delay, but it leads to current situation that all units shoots threat as all knows where it is, so adding some random delays here and there, adjusting little more the aiming accuracy and it would be similar what we have today (where units even shoot behind the air unit where it was like 2 seconds ago... Completely idiotic behavior because current CEP aiming principle). 

 

So basically anything that would be done, would require first the foundation to be done properly. And then start to add small things here and there using those new AI functions etc. 

[Lurker]

All of those things that you list Fri13 would require a massive rewrite of the AI code in DCS World. While your enthusiasm is commendable, this is something that would take years to code properly. We need a quick, temporary fix. Something that can be done with the existing codebase. Not sure why you have an issue with that, it's not like one of those things precludes work from being done on the other. I'm also not sure if you consider the the amount of processing power that such a detailed simulation would require, wherever possible developers need to use simplifications in these cases. If the resulting case of an AI rework is that DCS World would run like a slideshow then this is simply not a feasible solution. 

Edited June 8, 2021 by Lurker

[lmp]

I don't think we can realistically expect the AI to be able to behave intelligently beyond some very basic things. Go prone, turn front of tank towards enemy - yes. Communicate your contact to the rest of the platoon and talk them onto it with a realistic delay and degree of accuracy - no. That's way too much work for not enough difference to anybody but maybe CA players. The Hind is a big, loud helicopter flying against a sky background. They'll all see us before we see them anyway. Same goes for fragmentation damage. Sure, it's nice to have, but will doing a bit more damage (and unreliably) to unarmored targets be such a game changer? I think not.

 

IMHO the low hanging fruit would be to refine and clean up some of the tools we already have. I can "suppress" a unit or force it to "retreat" with the use of zones, triggers, go to waypoint and ROEs, but it's a huge pain in the rear. If we had branching, conditional waypoints, reliable ways to measure suppression, awareness of enemy, etc., we could easily create complex game plans for large numbers of units without being script wizards. We could easily make the AI units feel a lot less dumb and a lot more human.

 

1 hour ago, Lurker said:

All of those things that you list Fri13 would require a massive rewrite of the AI code in DCS World.

 

I would argue the same is true for your changes. There's no way to go from the current health bar damage model to anything even close to realistic in "maybe one day of work". Same goes for spotting. Sure, you can make the AI ignore what's not within the few degrees of FOV of the periscope, but then you need the units to realistically scan their surroundings and coordinate that scanning at least within a platoon... see how quickly that grew? All that is at least weeks of coding and weeks of research, data entry and testing for each of the great many units in the game.

[Lurker]

I was using hyperbole when I said it would take one day, but adding a simple "invisible hitbox" for units to take fragmentation damage from nearby exploding rockets in addition to their regular hitbox should be relatively simple to implement. Likewise, adding some sort of randomly generated percentage chance for an AI unit to engage a certain amount of time after spotting a unit (instead of as soon as possible like it is currently implemented) as well as some kind of random generated percentage chance for an AI to actually shoot some rounds not on it's currently implemented perfectly calculated trajectories, should not be that hard, and I argue would go a long way towards making them feel a little less like AI and a bit more like organic living units. 

 

I'm not saying that my solution is the be-all-end-all solution or that it's anywhere near close to a good solution, let alone an ideal solution, but in a game where missile chaff rejection is calculated in the exact same way, it could at least be an interim "hot-fix" to a problem that would otherwise be left to languish in Status-Quo development hell for who knows how long. 

Edited June 8, 2021 by Lurker

[Fri13]

Thank you ED for your new forum that deletes everything that is written and even just selects couple of the attachements randomly. 

Edited June 8, 2021 by Fri13

[BeastyBaiter]

2 hours ago, Lurker said:

I was using hyperbole when I said it would take one day, but adding a simple "invisible hitbox" for units to take fragmentation damage from nearby exploding rockets in addition to their regular hitbox should be relatively simple to implement. Likewise, adding some sort of randomly generated percentage chance for an AI unit to engage a certain amount of time after spotting a unit (instead of as soon as possible like it is currently implemented) as well as some kind of random generated percentage chance for an AI to actually shoot some rounds not on it's currently implemented perfectly calculated trajectories, should not be that hard, and I argue would go a long way towards making them feel a little less like AI and a bit more like organic living units. 

 

I'm not saying that my solution is the be-all-end-all solution or that it's anywhere near close to a good solution, let alone an ideal solution, but in a game where missile chaff rejection is calculated in the exact same way, it could at least be an interim "hot-fix" to a problem that would otherwise be left to languish in Status-Quo development hell for who knows how long. 

 

 

I remember the AI shooting with random inaccuracy from the perfect trajectory in a certain WW2 flight sim released in 2001, it was pretty hilarious. They would aim on the perfect trajectory and laser beam an arc, 1m short or 90 degrees off in the wrong direction, once again with perfect tracking not counting the offset. I  understand these sound like simple requests, but they really aren't. I would like to see the AI improved in general but I don't think there are any quick fixes, at least none that would be an actual improvement. This is a core game engine rework we are talking about, and for me the higher priority item is updating the game engine to use our modern 12-16 thread CPU's instead of being optimized for a Pentium 4. Imp's idea of improving the mission editor might actually be a good option. ED could add some template type scripts and other similar things without completely redesigning anything. Those are more straight additions rather than modifying existing items (which may require modifying everything that references them, and references the references, etc).

 

With all of that said, I don't think the AI is the problem you're making it out to be. Done plenty of Huey and Hip missions and survived against infantry and other soft targets. You just can't expect them to go head on with a tank division right now. But we aren't talking about a transport chopper., the Hind should be able to go after tanks much like the Hokum currently. The only restriction is it will be limited to 8 tank kills per arming instead of 12. Is that really such a big deal?