Understanding ME : AI logic or bug ?

[CougarFFW04]

Hi everybody,

 

Don't know if it is the right place to put this...

Moderator might move it to the right forum room if needed.

 

Still playing/learning ME and I have observed a very strange behavior that I cannot explain. I don't know if this is any kind of AI logic or if this is a bug... Bug IMO.

 

 

 

So I started with a very basic mission where 4 M2000, 4 F15E and 1 C130 assigned with "None task" simply fly their FP and RTB (there are also a few other flights that we can ignore as they behave the same in both missions (see after)).

 

I have managed that the 3 flights orbit WPT1 till they all reach it and then they push alltogether to WPT 2, 3,... I have also managed to "fire a message" when they reach WPT1 or 2 (for exemple "M2000 reached WPT1" or "F15 reached WP1" or "M2000 reached WPT2"...)

 

It works pretty well and when I run mission in F10 view, they do what they were supposed to do :) As shown in Pic1 the F15 and M2000 are orbiting there respective WPT1 (waiting for the C130) before proceeding normally to next WPTs... I also get the WPT messages correctly when the different flights pass WPT1 or 2. Everything fine. :thumbup:

 

 

http://img110.xooimage.com/files/a/0/d/testfp_samfar_small-54dee25.jpg

 

 

Then I have added a SA-3 battery a bit further.

 

Please note that I did NOT change anything concerning any flights.

The SA3 engagement circle does NOT cover the planes FP although FPs are within the Search and Tracking SA3 circles.

 

Basically when I run the "Sa3 mission variant", all planes behave mostly the same except the M2000 : after TO they rush to WPT1 at 800knt (see pic 2, in the no-Sam they were at 300knt to be on time at WPT1) and then orbit WPT2 (instead of 1). As soon as the F15 and C130 reach WPT1, all flights proceed to next WPTs as they are supposed to do. Although the F15 and C130 respect their GS along their respective FPs (as in the no-SA3 mission) the 2000 rush again from WPT to WPTs...

 

 

I have also noticed something weird : As one can see in Pict 3, although F15 orbits their WPT1 (as they are supposed to do) they are linked (dashed line) to WPT2 :cry:. On the other hand, the M2000 which are orbiting WPT2 are linked to their WPT1... Very very strange...

Last thing that might be of interest : As soon as the M2000 come close to WPT1 the game trigger both the M2000 WPT1 and WPT2 messages together.:joystick:

 

 

http://img110.xooimage.com/files/9/e/6/testfp_samcloserush_small-54dee31.jpg

 

 

http://img110.xooimage.com/files/1/8/4/testfp_samcloseorbit_small-54dee39.jpg

 

 

 

 

So basically 2 questions :

- Why do the M2000 rush to WPT1 in the "SA3 mission" ?

- Why do they orbit WPT2 although they are supposed to orbit WPT1...

 

I can provide the 2 .mis if necessary.

 

Thanks in advance.

Edited July 30, 2018 by CougarFFW04

[feefifofum]

This is getting specific enough that it would probably be helpful to post your .miz :thumbup:

 

What may be coming into play is the "REACTION TO THREAT" option; by default, this is set to "ALLOW ABORT MISSION" which permits AI to ignore their programmed taskings if they detect they are inside of a threat zone.

 

You can try changing this option to EVADE FIRE (WP0-> advanced waypoint actions ->set option->reaction to threat->evade fire) to see if that resolves the issue.

[CougarFFW04]

Hi Feefifofum,

 

 

 

 

You can try changing this option to EVADE FIRE (WP0-> advanced waypoint actions ->set option->reaction to threat->evade fire) to see if that resolves the issue.

 

 

Did that but didn't notice any improvements.

 

 

 

This is getting specific enough that it would probably be helpful to post your .miz

 

 

Finally only one .miz is enough to test the "Mad" behavior.

I have observed that when the SA3 is far enough everything is good.

However, when the battery is closer, at a certain point they become mad although far from the SA3 engagement circle...

So I suggest to run first the provided miz where I put the SA3 far enough.

Then edit the mission and move the SA3 battery inside the Su 27 FP (it is close enough to make M2000 Mad) and test it again (point2).

 

 

 

Please also note that there are certainly many ways to achieve what I wanted to test here i.e. basically flight pushing together and CAP behavior.

You have already pointed to me somme different ways to do this (and it is what I am testing now) but the question here is NOT how to do this differently but with do one observe this "mad" behavior and if its a bug to submit to DCS if possible.

 

 

1) So please first take a few minutes to run the .miz attached as it is and observe the general picture (observer F10 view) : M2000 (Mike), F15 (Farmer) and C130 (Cowboy) takeoff and join WPT1. As soon as they all reach WPT1 they push with a delay of 10s for Farmer and 20s for Cowboy and proceed to next WPTs at the right speed as defined in the ME. In particular pay attention to the messages that "fire" at WPT1 and 2 and concentrate on Mike speed and behavior. The sequence is more or less something like that :

Mike WPT1 -> Farmer WPT1 -> Cowboy WPT1

Mike pushing WPT2 -10s-> Farmer pushing WPT2 -10s-> Cowboy pushing WPT2

Then the sequence depends more or less of their respective positions when they push and travel to WPT2 (at 300knt) that they reach more or less at the same time. When they are WPT2 new messages are fired :

Mike (Farmer, Coyboy) WPT2

 

and they proceed alltogether to next WPTs at ~330knt

That is good and what I wanted :pilotfly:

 

 

 

2) Now test the mad scenario (move the SA3 closer and do NOT change anything else). Look how their behavior completly change. in particular the M2000 are the most affected. Concentrate on the M2000 :

- After takeoff they travel half the way WPT1 before rushing at 800knt (in scenario 1 they were at ~300knt just to be on tos WPT1)

- both "Mike WPT1" and "Mike WPT2" are fired at the same time :cry:

- they orbit WPT2 and not WPT1 as they are supposed to do...

- After Farmer and Cowboy reach WPT1 they push but, althoug this is not systematic, sometime Mike rush at 800knt to WPT3 (330knt in the normal scenario).

 

 

I have also observed some inconsistancy of the F15 speed but it is not systematic...

 

 

Please also observe the dashed flight path and how it generally stay locked on WPT1 till WPT3 :joystick:

 

 

Looking forward for your test and comments.

 

 

 

Big thanks in advance,

 

 

Regards,

 

 

Cougar

LearningME_Test_CAPDebug.miz

[BlacleyCole]

Something similar I’ve noticed in the hueys instead of taking off from pad then climbing to altitude on the way to wpt1 they climb to altitude then head to wpt1 this climb may be in any direction? Is that standard or a bug?

[CougarFFW04]

Hi,

 

Tested with a new mission from scratch and same behavior.

 

When introducing the Sam, Flights become crazy and do not respect speed.

They rush from WPT to WPT at 800knt :(:joystick::cry:

 

 

Bug ?

[feefifofum]

Been busy trying to put something together for my group for this weekend. I'll take a peek ASAP but yeah, doesn't sound right.

[feefifofum]

I'm not going to speculate as to why the AI chose to react in the way it did, but the issue had to do with REACTION TO THREAT as expected.

 

By setting the REACTION TO THREAT at WP1, after executing all your other actions, they were left at their default "ALLOW ABORT MISSION" setting until after the other scripts and commands were executed.

 

Setting REACTION TO THREAT: PASSIVE DEFENSE (or EVADE FIRE) at WP0, as soon as the AC spawn in, corrected the issue with the early push and orbit at WP2 instead of WP1.

 

 

EDIT:

Part of the issue you're experiencing with speed is also that you are not factoring in taxi times. If you tell them to be at their push point at 8:07, the mission starts at 8:00, and the last member of the flight isn't in the air until almost 8:06, they're going to need to haul ass to make it to the push point on time which is why they're flying at 800kts in the case of the F-15s, for example. The extremely short distances you're working with here are going to further complicate the issues and require timing things out quite a bit on your end with respect to how long it takes them to get in the air etc.

Edited August 4, 2018 by feefifofum

[CougarFFW04]

I'm not going to speculate as to why the AI chose to react in the way it did,

isn't it something ED developpers might answer ?

 

Setting REACTION TO THREAT: PASSIVE DEFENSE (or EVADE FIRE) at WP0, as soon as the AC spawn in, corrected the issue with the early push and orbit at WP2 instead of WP1.

OK will try. Thanks.

Nevertheless this seems to me like a workaround to something that sounds like a "bug"

 

EDIT:

Part of the issue you're experiencing with speed is also that you are not factoring in taxi times. If you tell them to be at their push point at 8:07, the mission starts at 8:00, and the last member of the flight isn't in the air until almost 8:06, they're going to need to haul ass to make it to the push point on time which is why they're flying at 800kts in the case of the F-15s, for example. The extremely short distances you're working with here are going to further complicate the issues and require timing things out quite a bit on your end with respect to how long it takes them to get in the air etc.

That's clear that they have to rush to reach WPT1 on time. But as you might have seen, they can do it in the "No Sam" scenario ( the M2000 leader fly at about 300knt to be on TOS@WPT1). The point was the different behavior bewtween the two scenario.

 

Part of the issue you're experiencing with speed is also that you are not factoring in taxi times

Assuming that the FP is "as it is", I do not understand what I could do. I have already lower at the minimum the speed (unlocked) to reach WPT1 to maximise their chance to be on TOS. As far as I see there is no way to take into account the taxi time... Changing from "departure from runway" to "cold start" (or whatever) do NOT change anything in the timing as computed by the ME software. Any advises ?

 

Thanks again.

 

++

[feefifofum]

1. Maybe, but they don't often come in here and your mission takes a few minutes of staring at to figure out what's supposed to be happening since it's all written in the SSE...so unlikely unless you write a super simple example and post it in the AI bugs forum.

 

2. Maybe, maybe not. Once I figure out how to make something work the way I want I generally stop asking questions...I find I stay a lot more sane that way. ;) Adjusting that reaction to threat value is a part of my routine whenever I create an AI flight because they often refuse to follow their scripting in a high threat scenario if given permission to ignore it. If I want them to try to do what they're scripted to do no matter what, I always change that value to EVADE FIRE at their start point.

 

3. Re: speeds they are obviously reacting to the SAM. Changing the reaction to threat parameter restores the original behavior...that's where I drop it. :dunno: You'll have to ask ED or RAZBAM if you want to know why they react to a SAM threat by tearassing around.

 

4. Re: taxi time it's just something that you as a designer need to think about and test, especially if you're looking for a seven minute time on station to a waypoint 10 miles from the airport. There is basically zero margin for error in this scenario. Waypoints in the flight plans are so close together that flights almost pass them in their orbits; it's going to create +/- a minute or two depending on where in their orbit they are when they get the push order and the AI just isn't that precise. Give them a little wiggle room and things will fall in line.

Edited August 8, 2018 by feefifofum

[CougarFFW04]

1. Maybe, but they don't often come in here and your mission takes a few minutes of staring at to figure out what's supposed to be happening since it's all written in the SSE...so unlikely unless you write a super simple example and post it in the AI bugs forum.

Ok. I Will see if it's worth trying...

 

2. Maybe, maybe not. Once I figure out how to make something work the way I want I generally stop asking questions...I find I stay a lot more sane that way. Adjusting that reaction to threat value is a part of my routine whenever I create an AI flight because they often refuse to follow their scripting in a high threat scenario if given permission to ignore it. If I want them to try to do what they're scripted to do no matter what, I always change that value to EVADE FIRE at their start point.

Good to know. Thanks

 

 

3. Re: speeds they are obviously reacting to the SAM. Changing the reaction to threat parameter restores the original behavior...that's where I drop it. You'll have to ask ED or RAZBAM if you want to know why they react to a SAM threat by tearassing around.

OK. I will see if it's worth asking...

I generally like to understand how things work, how and why...

 

4. Re: taxi time it's just something that you as a designer need to think about and test, especially if you're looking for a seven minute time on station to a waypoint 10 miles from the airport. There is basically zero margin for error in this scenario. Waypoints in the flight plans are so close together that flights almost pass them in their orbits; it's going to create +/- a minute or two depending on where in their orbit they are when they get the push order and the AI just isn't that precise. Give them a little wiggle room and things will fall in line.

Sure. My goal was to test how they behave under "stressing time constrains"

 

Thanks again :thumbup: