any work planned on dogfighting AI?

[Stonehouse]

23 minutes ago, AEthelraedUnraed said:

I haven't tried it yet, but I must say it looks very impressive and well thought-out! I agree that it sounds more like what historical accounts say about aces vs rookies. However I think this thread is mostly about the tactics rather than gunnery of the AI

 

I thought perhaps the changes to HittoDodge rate might have been relevant to the discussion since it seems to change the AI behaviour but agree about your gunnery v's tactics comment and this thread.  Although the mod does change things somewhat because the changes in firing ranges tends to alter how dogfights plays out a bit. Anyway it was just for whatever it was worth in case it helped.  

[Ribbon]

On 4/26/2022 at 4:19 PM, AEthelraedUnraed said:

Well, he also said recently that they're more or less on schedule. So I guess that everything will be alright, but it does mean that they have little room for extra stuff that cannot be postponed.

 

I don't agree. When set to Ace level, the AI poses quite a challenge (when flying an aircraft of similar performance than you). Perhaps you are a better pilot than I am and that's not the case for you, but given that the average IL2 player has much more dogfighting experience than the average WW2 pilot, I think the AI is fine.

 

 

I'm also fine with some delays, hope they are fine too.....even i'm really looking forward to Channel map!

 

Well maybe i'm used too much on MP challenge but AI IMO should be more aggressive, for example if i'm outnumbered  all those enemy fighters should gang bang me, giving me very little chance to survive.

Now that is not the case, they take they time and attack one by one over long period of time between attacks like only one enemy is assigned to engage me....while others wait for that one to be destroyed or complete whole circle to engage me.....and only then the other one decide to engage me.

They should just give me less air to breathe, this way i shoot them one by one even i'm alone vs 4 or more of them.

I try to keep my expetations realistic and i don't expect AI being capable as human but more aggressivness from them are needed!

[PatrickAWlson]

2 hours ago, =VARP=Ribbon said:

I'm also fine with some delays, hope they are fine too.....even i'm really looking forward to Channel map!

 

Well maybe i'm used too much on MP challenge but AI IMO should be more aggressive, for example if i'm outnumbered  all those enemy fighters should gang bang me, giving me very little chance to survive.

Now that is not the case, they take they time and attack one by one over long period of time between attacks like only one enemy is assigned to engage me....while others wait for that one to be destroyed or complete whole circle to engage me.....and only then the other one decide to engage me.

They should just give me less air to breathe, this way i shoot them one by one even i'm alone vs 4 or more of them.

I try to keep my expetations realistic and i don't expect AI being capable as human but more aggressivness from them are needed!

 

I question how much, if any, cooperation is coded into the AI.  There are times I swear it is doing drag and bag but that could also be me getting fixated and then killed as opposed to the AI intentionally doing it.  More complex stuff, like a simulated "break right" while the flight mate maneuvers with the expectation that this will happen, or something like a cooperative Thatch Weave,  ... I am guessing not.  Organized bouncing of a target - also guessing not.

[oc2209]

8 hours ago, =VARP=Ribbon said:

Well maybe i'm used too much on MP challenge but AI IMO should be more aggressive, for example if i'm outnumbered  all those enemy fighters should gang bang me, giving me very little chance to survive.

I try to keep my expetations realistic and i don't expect AI being capable as human but more aggressivness from them are needed!

 

I rather strongly disagree.

 

If all humans were equally aggressive, then the Germans never could've survived in Russia, and American/British flights never would've survived being outnumbered by gaggles of mostly undisciplined German fighters on the occasions the Germans happened to have local superiority.

 

The multiplayer environment is a mostly false representation of average air combat. Multiplayer is the equivalent of having, out of a lobby with, say, 30 people in it and assuming all are fighter pilots, 10 of them being 50+ kill aces, 10 of them being 10-20 kill aces, and the other 10 being 5-10 kill aces.

 

All the people there are kill hungry and competitive.

 

In reality, what's the actual percentage of trained pilots who saw combat and achieved ace status? 5%? 10% at most (depending on the nation)?

 

Being outnumbered by AI shouldn't be an automatic death sentence, because in real life, it simply wasn't the case.

 

Even setting the AI to 'Ace' level shouldn't make them all hyper-aggressive. All the cumulative effect of fighting ace-level AI is, you're less likely to get through a sortie without taking at least a few hits. It is a subtle, gradual effect of lowering your long-term chances of survival. This is most noticeable in a hard career on ironman.

[stevinne]

And none of the 30 people in that lobby are worried about dying, which I suspect had a lot to do with many of the combat pilots in WWII never actually shooting down an enemy. If my ability to continue playing flight sims rested on my ability to survive each encounter, I'd be one very timid pilot. Yet, online I'm eager to dive back into the fray.

[Ribbon]

3 hours ago, oc2209 said:

 

I rather strongly disagree.

 

If all humans were equally aggressive, then the Germans never could've survived in Russia, and American/British flights never would've survived being outnumbered by gaggles of mostly undisciplined German fighters on the occasions the Germans happened to have local superiority.

 

The multiplayer environment is a mostly false representation of average air combat. Multiplayer is the equivalent of having, out of a lobby with, say, 30 people in it and assuming all are fighter pilots, 10 of them being 50+ kill aces, 10 of them being 10-20 kill aces, and the other 10 being 5-10 kill aces.

 

All the people there are kill hungry and competitive.

 

In reality, what's the actual percentage of trained pilots who saw combat and achieved ace status? 5%? 10% at most (depending on the nation)?

 

Being outnumbered by AI shouldn't be an automatic death sentence, because in real life, it simply wasn't the case.

 

Even setting the AI to 'Ace' level shouldn't make them all hyper-aggressive. All the cumulative effect of fighting ace-level AI is, you're less likely to get through a sortie without taking at least a few hits. It is a subtle, gradual effect of lowering your long-term chances of survival. This is most noticeable in a hard career on ironman.

So in IRL situation you alone vs at least 4 enemy fighters whom with higher energy, better performing planes and the rest advantages it is realistic you down them all in 99,9% cases?

 

You don't need to reply, really!

 

To close my point, like in any game it's all about challenge and i'm here simply giving my feedback (cos i like this game otherwise i wouldn't bother writing this and giving my point of view how to improve it)...devs take it or leave it ?‍♂️

 

 

 

 

9 hours ago, PatrickAWlson said:

 

I question how much, if any, cooperation is coded into the AI.  There are times I swear it is doing drag and bag but that could also be me getting fixated and then killed as opposed to the AI intentionally doing it.  More complex stuff, like a simulated "break right" while the flight mate maneuvers with the expectation that this will happen, or something like a cooperative Thatch Weave,  ... I am guessing not.  Organized bouncing of a target - also guessing not.

Now when you're saying it, i've had similar feeling even i'm not into it as much as you are (as a mod developer i bet you know it's patterns better than any of us).

AI is limited in every game, especially in flight sims where beside stationary aim bot more cruical thing is it's path/movement vector in 3d enviroment vs players position and movement so i don't have high expetations also given it's not AAA studio with unlimited resurces and manppwer, add to it limitations and dev working on AI that continues other guy's work/code (my pure speculation here)....it doesn't seem simple, easy or fast fix...here i just gave feedback how i'd make it more challenging or even cover other AI flaws.

Here time between No.1 enemy fighter engagement and No.2 should be reduced (while No.1 emeny is in engagement others should adjust it's paths to player's 6 and gaining advantage) giving tense situation to player and narrowing it's capability to adjust and defend. 

 

 

Edited April 28, 2022 by =VARP=Ribbon

[oc2209]

7 hours ago, =VARP=Ribbon said:

So in IRL situation you alone vs at least 4 enemy fighters whom with higher energy, better performing planes and the rest advantages it is realistic you down them all in 99,9% cases?

 

In real life, you would rarely find yourself alone with 4 enemy planes. For that to happen, many things would need to go wrong.

 

What you're outlining isn't even a real world scenario that's frequent enough (I really want to emphasize that point) to matter; to adjust the AI for specifically.

 

As I said, the AI, on ace level, is not meant to hand us our asses 1v1 or even 4v1. It's meant to wear us down over time. That's what makes a career challenging.

 

You might be able to defeat it with ease in 10 combat encounters; maybe a few dozen. But can you do it for 100? 300? Statistically, the odds aren't in your favor.

 

Also, having an insanely aggressive AI, with all of your enemies acting exactly alike in terms of aggression levels, would instantly feel very unrealistic and video gamey.

 

Having the AI lazily circle over you while one or two of them occasionally dives down to attack you--that's actually how real pilot accounts sound. Multiple attackers of a single target get in the way of each other, and they spend more time avoiding colliding with each other than they do shooting at the enemy. It's smarter to have the one guy who knows what he's doing, go in and make the attack, while everyone else provides top cover and cuts off escape routes.

[oc2209]

Here's a quick example of a bounce scenario. I'm in a 190A-5, the 4 AI are set to ace in Tempests. We meet head on at 300m; 2 of the AI turn directly for me and pass overhead; 2 others approach obliquely, and these 2 are the ones that latch on to my tail first.

 

In 1 minute and 13 seconds, I'm shot down (fatally) from this distance:

 

Spoiler

 

A little over 700m.

 

I'm not sure how much more aggressive the AI could be here. Just because the AI can't always handle a maneuvering human, doesn't mean it can't recreate a normal bounce situation effectively.

 

And bounces are how the AI will take you down in career, more often than not.

[tattywelshie]

23 minutes ago, oc2209 said:

Here's a quick example of a bounce scenario. I'm in a 190A-5, the 4 AI are set to ace in Tempests. We meet head on at 300m; 2 of the AI turn directly for me and pass overhead; 2 others approach obliquely, and these 2 are the ones that latch on to my tail first.

 

In 1 minute and 13 seconds, I'm shot down (fatally) from this distance:

 

  Reveal hidden contents

 

A little over 700m.

 

I'm not sure how much more aggressive the AI could be here. Just because the AI can't always handle a maneuvering human, doesn't mean it can't recreate a normal bounce situation effectively.

 

And bounces are how the AI will take you down in career, more often than not.

Yeah this is the thing, what are the chances of making it through all 4 careers on ace level and Ironman mode? Tiny I reckon! Sure there’s a chance of being downed by flak but the chances are you’ll cop it sooner or later from the ai.

[AEthelraedUnraed]

8 hours ago, =VARP=Ribbon said:

So in IRL situation you alone vs at least 4 enemy fighters whom with higher energy, better performing planes and the rest advantages it is realistic you down them all in 99,9% cases?

How many hours of dogfighting experience do you have, in how many dogfights?

 

I think you have a warped view of how good any AI should be. Fact is that many of us are much better dogfighters than even the best of WW2 pilots were. You have simply out-trained them.

[Ribbon]

2 hours ago, oc2209 said:

 

In real life, you would rarely find yourself alone with 4 enemy planes. For that to happen, many things would need to go wrong.

 

What you're outlining isn't even a real world scenario that's frequent enough (I really want to emphasize that point) to matter; to adjust the AI for specifically.

 

As I said, the AI, on ace level, is not meant to hand us our asses 1v1 or even 4v1. It's meant to wear us down over time. That's what makes a career challenging.

 

You might be able to defeat it with ease in 10 combat encounters; maybe a few dozen. But can you do it for 100? 300? Statistically, the odds aren't in your favor.

 

Also, having an insanely aggressive AI, with all of your enemies acting exactly alike in terms of aggression levels, would instantly feel very unrealistic and video gamey.

 

Having the AI lazily circle over you while one or two of them occasionally dives down to attack you--that's actually how real pilot accounts sound. Multiple attackers of a single target get in the way of each other, and they spend more time avoiding colliding with each other than they do shooting at the enemy. It's smarter to have the one guy who knows what he's doing, go in and make the attack, while everyone else provides top cover and cuts off escape routes.

Yeah flying above and watching how i bust their arses one by one not helping their wingmans, you're here simply to provide argument looking and justifying at very narrow AI problem ignoring everything else which wouldn't happen IRL.....where pilots would adjust to situation and you wouldn't have a chance vs multiple enemies with better planes and higher energy.....fact!

You're ignoring everything else in your arguments, logic and facts, just for the sake of justifying single fraction of problem.

 

I see no point discussing it any further!

 

 

 

[Dragon1-1]

Pilots do ignore everything else and not help their wingmen all the time. Good pilots wouldn't do it. Most pilots in WWII weren't good. A WWII ace could (and a few of them did) survive an encounter where 4 planes had a drop on him. Cooperating was encouraged, but in real situations, it was often forgotten about. Thatch Weave was so effective in part because it involved aircraft cooperating against an enemy that usually wouldn't work together. Today, working with your wingman is one of the main tenets of effective air combat, but this lesson was only learned in WWII (not that it wasn't known by the end of WWI, it just didn't quite sink in). We fly according to modern tenets of E-M theory and post-Vietnam tactics. WWII pilots didn't have most of that.

[Gambit21]

5 hours ago, Dragon1-1 said:

Pilots do ignore everything else and not help their wingmen all the time. Good pilots wouldn't do it.

 

Once a melee started, it was basically every man for himself...good pilots and otherwise.

[oc2209]

4 hours ago, Gambit21 said:

Once a melee started, it was basically every man for himself...good pilots and otherwise.

 

Statements this broad need some qualifications.

 

Americans circa 1942-43 and Eastern Front Germans absolutely needed to think in group terms--at least 2 man elements--before and during an engagement.

 

In other words, when you're fighting at either a numerical or qualitative disadvantage, you need to plan and coordinate or you'll take unsustainable losses.

 

The Japanese and Russians didn't have a choice--no radios meant no real coordination to speak of. But British, Americans, and Germans did have a choice to some extent.

 

10 hours ago, =VARP=Ribbon said:

you're here simply to provide argument looking and justifying at very narrow AI problem ignoring everything else which wouldn't happen IRL.....where pilots would adjust to situation and you wouldn't have a chance vs multiple enemies with better planes and higher energy.....fact!

 

Okay. If you only deal in facts, then this didn't happen:

 

"Hartmann had flown five missions against North American P-51 Mustangs in the spring of 1944 and briefly later on, downing a total of seven Mustangs. If he had flown more, it would have been extremely difficult for him to survive the war. In the encounter with the eight Mustangs chasing him at Ploiesti, the Americans had fired wildly and Hartmann had narrowly missed being shot down. If the Americans had employed his method of getting up close, it's unlikely he would have survived. In fact, they chased him until he ran out of fuel. After he bailed out, one Mustang passed by to photograph him and waved as he passed by. During this encounter, nearly half of the Gruppe's aircraft had been shot down, with two pilots dead and a number of others wounded."

 

(courtesy of http://www.aviation-history.com/airmen/Erich_Hartmann.htm)

 

And Sadaaki Akamatsu didn't engage "75 Mustangs" solo in a Zero, and live. This account is from Henry Sakaida from Osprey books.

 

The problem here is that you're overestimating the average quality of human pilots in WWII, and then you're trying to apply that overblown estimate to AI, which would be even more difficult for programmers to implement than 'normal' AI behavior.

We sim players are much closer to Hartmanns and Akamatsus, relative to AI capabilities, than we are to the average real pilot in WWII.

 

What you're asking is not to have the AI approximate human behavior; but having it approximate the upper 1% of human behavior. And that's something no AI can do in a flight sim.

[Gambit21]

25 minutes ago, oc2209 said:

 

Statements this broad need some qualifications.

 

Americans circa 1942-43 and Eastern Front Germans absolutely needed to think in group terms--at least 2 man elements--before and during an engagement.

 

In other words, when you're fighting at either a numerical or qualitative disadvantage, you need to plan and coordinate or you'll take unsustainable losses.

 

The Japanese and Russians didn't have a choice--no radios meant no real coordination to speak of. But British, Americans, and Germans did have a choice to some extent.

 

 

Okay. If you only deal in facts, then this didn't happen:

 

"Hartmann had flown five missions against North American P-51 Mustangs in the spring of 1944 and briefly later on, downing a total of seven Mustangs. If he had flown more, it would have been extremely difficult for him to survive the war. In the encounter with the eight Mustangs chasing him at Ploiesti, the Americans had fired wildly and Hartmann had narrowly missed being shot down. If the Americans had employed his method of getting up close, it's unlikely he would have survived. In fact, they chased him until he ran out of fuel. After he bailed out, one Mustang passed by to photograph him and waved as he passed by. During this encounter, nearly half of the Gruppe's aircraft had been shot down, with two pilots dead and a number of others wounded."

 

(courtesy of http://www.aviation-history.com/airmen/Erich_Hartmann.htm)

 

And Sadaaki Akamatsu didn't engage "75 Mustangs" solo in a Zero, and live. This account is from Henry Sakaida from Osprey books.

 

The problem here is that you're overestimating the average quality of human pilots in WWII, and then you're trying to apply that overblown estimate to AI, which would be even more difficult for programmers to implement than 'normal' AI behavior.

We sim players are much closer to Hartmanns and Akamatsus, relative to AI capabilities, than we are to the average real pilot in WWII.

 

What you're asking is not to have the AI approximate human behavior; but having it approximate the upper 1% of human behavior. And that's something no AI can do in a flight sim.

I was thinking of (but neglected to type) U.S. escort pilots. Every Mustang/Jug pilot I’ve chatted with or interviewed indicated this was generally the case.

[Ribbon]

2 hours ago, oc2209 said:

 

Statements this broad need some qualifications.

 

Americans circa 1942-43 and Eastern Front Germans absolutely needed to think in group terms--at least 2 man elements--before and during an engagement.

 

In other words, when you're fighting at either a numerical or qualitative disadvantage, you need to plan and coordinate or you'll take unsustainable losses.

 

The Japanese and Russians didn't have a choice--no radios meant no real coordination to speak of. But British, Americans, and Germans did have a choice to some extent.

 

 

Okay. If you only deal in facts, then this didn't happen:

 

"Hartmann had flown five missions against North American P-51 Mustangs in the spring of 1944 and briefly later on, downing a total of seven Mustangs. If he had flown more, it would have been extremely difficult for him to survive the war. In the encounter with the eight Mustangs chasing him at Ploiesti, the Americans had fired wildly and Hartmann had narrowly missed being shot down. If the Americans had employed his method of getting up close, it's unlikely he would have survived. In fact, they chased him until he ran out of fuel. After he bailed out, one Mustang passed by to photograph him and waved as he passed by. During this encounter, nearly half of the Gruppe's aircraft had been shot down, with two pilots dead and a number of others wounded."

 

(courtesy of http://www.aviation-history.com/airmen/Erich_Hartmann.htm)

 

And Sadaaki Akamatsu didn't engage "75 Mustangs" solo in a Zero, and live. This account is from Henry Sakaida from Osprey books.

 

The problem here is that you're overestimating the average quality of human pilots in WWII, and then you're trying to apply that overblown estimate to AI, which would be even more difficult for programmers to implement than 'normal' AI behavior.

We sim players are much closer to Hartmanns and Akamatsus, relative to AI capabilities, than we are to the average real pilot in WWII.

 

What you're asking is not to have the AI approximate human behavior; but having it approximate the upper 1% of human behavior. And that's something no AI can do in a flight sim.

Ypu might read what you write, think about it and compare it to il2 AI...

Did Harmann flew in circles and mustangs engae him from above one by one? No, it was a pursuit where he barely made it alive, downing two of them probably at start of encounter by suprise and than running away, same with zero pilot.

Not to mention in il2 ypu'll go in circles and down them all compared to two Harmanns downed p51 he got by suprise.

 

You keep replying to me with bad arguments, AI is known flaw of il2, in dogfights not doing proper maneuvers and my suggestion was to increse number and aggressivness to cover those flaws.

 

Now this is really last from me on this subject!

 

Bye?

Edited April 29, 2022 by =VARP=Ribbon

[Willy__]

If the AI behaved like the old 1946 I'm sure there would not be so many complaints like we have... And I know the AI back then "cheated", but at least it behaved in a realistic way. Also I really liked that you could break your flight in pairs, sending the second pair to attack while you and your wingman cover them (or vice-versa), right now the commands that you can do for the AI in IL2 GB is barely adequate. And thats with air-to-air in mind. If you start adding ground attack orders it really is lacking compared to the old title.

[oc2209]

I think the AI heard us talking about it, and decided to spontaneously self-improve.

 

Spoiler

 

 

This was about 1 minute after the combat started. My pilot was heavily wounded, which normally precludes bailing out; so I was trying to get some altitude, roll the plane over, and fall out of the cockpit.

 

And here's me getting my wing broken off from a half a kilometer away as I turn:

 

Spoiler

 

Edited April 30, 2022 by oc2209

[Eisenfaustus]

10 hours ago, oc2209 said:

What you're asking is not to have the AI approximate human behavior; but having it approximate the upper 1% of human behavior. And that's something no AI can do in a flight sim.

It’s even nothing the AI should do. 
 

Yet of course the AI is far from perfect. 
 

While I‘m pretty happy with how the single plane behaves I still have two major wishes:

- proper group tactics and formations 

- More precise radio orders, that are followed! (Including receiving sensible commands from an AI flight leader if you are a wingman)

 

[Yogiflight]

2 hours ago, Eisenfaustus said:

While I‘m pretty happy with how the single plane behaves I still have two major wishes:

- proper group tactics and formations 

- More precise radio orders, that are followed! (Including receiving sensible commands from an AI flight leader if you are a wingman)

Especially the first one is currently my major issue with AI. Always turning hard for heading corrections, even if it is only a few degrees, even with heavy bombload and full fuel at low speed, seriously damaged aircraft (how often do damaged aircrafts crash in landing pattern, because they try to turn hard and simply fall out of the sky), flying in echelon right formation, bombers and Stuka flying in a nine ship V-formation (except enemy bombers in intercept missions) instead of three ship V-formations, the AI, flying in front of the player constantly throttling up/braking, and all that kind of stuff, which takes away immersion completely.

And even worse the ground attack behaviour, diving down to 500m, 10km before reaching the target, throttling back to not get too fast, then 1-2km before reaching the target starting climbing back to 1000m, climbing with low speed directly over the AA defended target to attack it from another direction and after the attack run always turn to the surprisingly next to the target positioned enemy airfield to get the next AAA firing at them.

These kinds of behaviour are for me currently the most urgent serious issues with AI. The dogfighting of the AI has improved quite a lot and is not too bad. There needs to be some fine tuning, to give the AI the ability to use the strengths of every aircraft types, like the La-5 or I-16, which currently are too easy preys, even for the Bf 110 E2. But for me the focus currently should be on friendly AI, because dogfighting is not everything for a combat flight simulation. It is the only thing for QMB, but that is not simulation, but first person shooter, starting mission and getting action as fast as possible. This is not a bad thing either, especially for players, who don't have the time to fly long missions, but it shouldn't be the only focus of the developers.

 

EDIT: One point, I forgot to mention is, in both, dogfights and ground attacks, as soon as the action starts, the flight leader gets slower and lets the AI pass, so he is only the pathfinder, who leads the flight to the action and then lets his flight fight on their own, following in a large distance. This make him an easy target for the enemy, fighters and AAA, which causes him usually being the first victim in the player flight. The flightleader should be the leader, especially in combat, not only for finding the waypoints.

Edited April 30, 2022 by Yogiflight

[enyak]

I see this topic has now been been picked up but I just want to reiterate that I think the individual flying / combat skills of the AI is fine by itself. It can always be improved of course.

 

However what's really missing is formation and group tactics. At leaster higher skill AI should actively maneuver to stay in formation and use group tactics to engage or defend from the enemy. This is a core principle in ACM that is completely missing in SP for now.

[49Topper]

On the general topic of AI, I've noticed (to my frequent demise) that when AIs spawn, they frequently spawn directly on my six. Given that I am a constant six checker, it is apparently not often enough as I get nailed by an AI that was nowhere to be seen a few seconds prior to my flaming death. Does anyone else notice this or is it just me being a lousy fighter pilot? What am I doing wrong?

 

Thanks.

[ShamrockOneFive]

1 hour ago, 49Topper said:

On the general topic of AI, I've noticed (to my frequent demise) that when AIs spawn, they frequently spawn directly on my six. Given that I am a constant six checker, it is apparently not often enough as I get nailed by an AI that was nowhere to be seen a few seconds prior to my flaming death. Does anyone else notice this or is it just me being a lousy fighter pilot? What am I doing wrong?

 

Thanks.

 

In the 7-8 years that I've been flying this iteration of IL-2, I have never ever had an AI aircraft spawn nearby and be on my six. Some missions do spawn AI aircraft in but they are usually doing this at 20-30km out.

 

Record tracks and review what's going on. It seems unlikely to me that this is happening... and rather they were a bit sneaky and got you. It happens to the best of us

[oc2209]

3 hours ago, ShamrockOneFive said:

In the 7-8 years that I've been flying this iteration of IL-2, I have never ever had an AI aircraft spawn nearby and be on my six. Some missions do spawn AI aircraft in but they are usually doing this at 20-30km out.

 

This is true in my experience.

 

Enemy AI also rarely (the exceptions being bombers) spawns as single planes.

 

Fighters always come in pairs, at least. And some of the fighters that do spawn during a given sortie are quite far from your operating area; you won't encounter them unless you go out looking for them.

[KevPBur]

9 hours ago, 49Topper said:

On the general topic of AI, I've noticed (to my frequent demise) that when AIs spawn, they frequently spawn directly on my six. Given that I am a constant six checker, it is apparently not often enough as I get nailed by an AI that was nowhere to be seen a few seconds prior to my flaming death. Does anyone else notice this or is it just me being a lousy fighter pilot? What am I doing wrong?

 

Thanks.

 

I've never noticed this. I often fly with the little map visible as my old eyes can't spot anything! I have never noticed ai spawning within that range.

 

If your problem is with pesky 109's they tend to always have a significant height advantage to the point that without labels on I can't see them whilst bumbling along in an La-5 unable to get above 2k without loosing all performance. They then dive down seemingly appearing from nowhere.

[smink1701]

Spent about 10 minutes in SP yesterday, me in a P38 against five 109s.
Two collided and two more cratered. 
Typical AI. 
Many great things about IL2 but AI is not one of them. 
Back to MSFS. 

[KevPBur]

39 minutes ago, smink1701 said:

Spent about 10 minutes in SP yesterday, me in a P38 against five 109s.
Two collided and two more cratered. 
Typical AI. 
Many great things about IL2 but AI is not one of them. 
Back to MSFS. 

 

What type of mission did you fly?

In Quick Combat the ai is single mindedly hell bent on a rolling dogfight and all of them want to get on your tail at the same time, hence the collisions and to some extent the lawndarts as they refuse to leave your tail.

 

In QMB they are better behaved, especially if you have some friendly ai too.

I find Campaigns to be the best as every plane has objectives to achieve beyond your destruction and they then do a far better job of actually challenging you.

 

There has to be a little give and take with the ai. The more realistic the scenario the more believable their action. You taking on 5 109's in a single P38 would, I suspect, be a rather bold move for even the cockiest young flyboy with the realisation he doesn't get to press reset if it doesn't go his way.

 

As has been mentioned before in the thread, the ai will only use period apropriate  manouvres, again we may be prone to using more recent knowledge to our advantage and not flying by the book as we don't care if we lawndart or blow or damage an engine (or 2 in case of the P38).

 

 

[smink1701]

It was quick combat.

When I load IL-2 I typically have a 30 minute window of stick time. I do a quick combat scenario, get bored and frustrated after about five minutes and then spend 25 minutes with MSFS and knock out some landing challenges. 

[Eisenfaustus]

15 minutes ago, smink1701 said:

It was quick combat.

When I load IL-2 I typically have a 30 minute window of stick time. I do a quick combat scenario, get bored and frustrated after about five minutes and then spend 25 minutes with MSFS and knock out some landing challenges. 

If I may give you an advice: create a career as squadron leader (not Ironman) - let air start selected and manually move waypoint 2 (air spawn point) as close to the action as you can. Now you start very close to the action. After it you can leave in air as long as you are over friendly territory - or if you still have the time fly home and land. This way you can enjoy career missions within your limited time window - a problem I can very much relate to ^^

[KevPBur]

55 minutes ago, smink1701 said:

It was quick combat.

When I load IL-2 I typically have a 30 minute window of stick time. I do a quick combat scenario, get bored and frustrated after about five minutes and then spend 25 minutes with MSFS and knock out some landing challenges. 

 

Try the QMB those missions are usually 30 mins and you start typically 5 minutes from the action. You are always flight leader and you can select the type of combat you require, Escort, attacking bombers/attack planes, etc. If all you fancy is a dogfight, leave the rest of your flight to tackle the bombers whilst you entertain the escorts

 

I fly these when I have less time or am trying a new or unused plane. Again, the ai behaves, generally, better than in Quick Combat.

 

[FuriousMeow]

On 4/27/2022 at 2:31 PM, PatrickAWlson said:

 

Spoiler

True and not true

 

Not true.  There have been a steady stream of improvements funded by, but not directly related to purchased product.  AI is the perfect example - you can't help but notice how much better the AI is compared to three years ago.  Various graphics improvements, clouds and plenty more.  

 

True.  AI still has a long way to go.  Performance improvements have been inadequate.  Bugs can linger for a very long time.

 

Here's my take: 1C has to create and deliver the modules to make money.  No money, no business, no product - the end.  They would like to do all of these other things and they do try, but they have to prioritize the modules that actually make money.  The money made is not enough for a staff of hundreds or even dozens, so things linger.  1C does not know how to monetize the other improvements.  Maybe they can't.  This is a complex sim, not easily accessible.  That means limited potential revenue streams.  

 

The team is too small.  The revenue to grow it will never really be there due to limited potential revenue streams.  The only way that I can think to solve the problem is almost certainly a non-starter.  Open source.  Put the code online and accept contributions.  Meanwhile, focus the team on the modules that make money.  Anybody want to take bets on that happening?  

 

 

While this is old, I don't visit the forums that much these days, I think you missed what I was replying to. Ideally, funding continues, all sorts of "fixes"/changes/etc happen to the existing software which is what you alluded to.

 

Funding doesn't continue, which is what I was replying to by virtue of that person saying they won't buy anymore and expect the software to be fixed up and more features added to make it more of what they expect, further development would simply never happen.

 

Edited May 21, 2022 by FuriousMeow
double