Studying the AI & fully dynamic game sessions

[suBB]

Originally Posted by MarkShot

At the moment, I have put my grand schemes on back burner, since I have yet to play all the SC and DW missions hosted at SubGuru. Right now, I am still playing through the user missions for SC 1.08. Then, I will move on to SCX, SCU, and finally, LWAMI.

First, other people's scenarios although usually not very dynamic are dynamic, so to speak, the very first time you play them, and I have lots ahead of me.

Second, I am learning a lot about scenario design and the game engine by playing what others have done.

However, I still think a dynamic campaign set of scenarios or pre/post-processor would be a rebirth for these two sims. For example, SH2, was a relatively closed game system. The PA team through work done with Kriegstanz (a scenario templating engine) and pre-processors created SH/PA (Pacific Aces) and turned it into a true dynamic campaign. Sadly, like most great mods it was still hampered by original games bugs (like save games not working) and finally the arrival of SH3 killed off large scale community interest in the mod. At least, those of us who play SC/DW don't really need to worry about anything usurping these games. Yes, UBI could bring out a Cold War subsim, but it will be a hack on the SH3/4 engine focusing on graphics first and simulation second.

Before I begin, I can only account for LW/AMI, DW and MP missions and not SC & SP scenarios. Maybe this will help in your efforts of scenario design.

Here is what we’ve done in less than 2 years time regarding creating a scenario with in-game dynamics. But missing some vital points recently learned and verified. More points will be included in a features list of a MP map set currently WIP that will go into greater detail of new lessons learned and what we did to accomplish in-game dynamics.

The following points are:

Suppression of Predictability / Optimized In-Game Dynamics - I’ve learned that the only way to optimize in-game dynamics is to suppress predictability as much as possible. One big no-no that plays into predictability is spawning platforms as such to where detection WILL BE made by the same opposing platforms each time the map is loaded, or some period after heading into a general direction, or both. This doesn’t give players many options to choose from, which equates to limitation of tactics and strategy. Sooner or later the player will know the map as such that there really isn’t a need for tactics or strategy, only deployment of weapons. There are other ways to suppress predictability based on scenario design. The more you suppress predictability while balancing playability, the more you optimize in-game dynamics, something I call ‘optimized in-game dynamics’

LWAMI & AI – in achieving in-game dynamics I want to say we’ve accomplished our objective in harnessing what LW/AWI does for us in terms of AI. Sometimes AI has a mind of its own and at best to make ai contribute to in-game dynamics is use of RSB / DL combo along with RS tactic. The wider the area the more dynamic the experience, but be mindful of SSP type and balancing (more later). Thorough scripting can make AI just as effective as humans and use of scripts can make AI even more robust. It could be argued that AI can’t react and adapt to situations as a human can. But since LWAMI allows AI to be influenced by humans, in some regard AI can adapt to an ever-changing situation. IIRC that SC editor does not include scripts and AI is crappy, for those two reasons SC scenarios will fall on their face w/o use of lw/ami & DW. Since AI is AI, AI is limited to lack of reaction time and adaptation to an ever-changing situation. For that reason any scenario involving AI w/o at least 1 human ally would assume a linear performance curve.

Human Intervention – Simply put human intervention definitely makes up for Ais’ ability to react and adapt to the situation. Contributions of AI to in-game dynamics are limited by how robust you make them in editing, and nothing more.

Spawning & SSP / Sense of Freedom – Spawning should be as such to where it offsets ‘guaranteed detection’ so that form of predictability is reduced. In the same token be mindful of the SSP and not create a scenario that isn’t playable. By means of extensive testing under normal conditions of the map, a ‘base detection’ can be determined where SSP and spawning over the area are contributing factors to this number. Base detection should be optimized for 50%, which gives both sides 50% chance of countering each others’ tactics and strategy. What this does it gives players a true sense of freedom and room for their personal skill sets to contribute to the scenario and vary the opposing sides chance of success.

Hope this helps, good luck

[suBB]

Originally Posted by FERdeBOER

OK... I've been doing a small test.



kage, can You (or SuBB) tell us some of how you create your "dynamics"?

Thanks

I didn't notice earlier but of course, I'll be happy to share what was learned in this 'quest for truth' behind in-game dynamics.

You’ll find some limitations beyond control such as 'suicide attack' script. but with some creativity and sharing ideas with others you can come up with a way to workaround it or something better.

For example I ran into a fuel problem with su-27 interceptors flying a 4.5 hour CAP after leaving the airport. The workaround was ‘converting’ a tu142 bear or IL May into ‘in-flight refueling aircraft’ for the flankers using some triggering w duration and scripting.

The tanker and CAP will fly formation and refuel every 2 hours (1:30 dw time w 30 minutes duration 15 minutes / plane) but if the 27’s make contact w hostile, they engage afterburners for 6 minutes breaking formation w tanker, engage contact w either missiles and guns, and when the contact is destroyed they return to formation w the tanker and resume CAP.

I think I can be far more descriptive in ‘how’ we did it when the features list is complete for the beta I’m releasing soon. But for now use the use this as reference and any questions you have about scripts, triggering & whatnot I’ll be glad to offer assistance.

[MarkShot]

So, are the only serious mission/hardcore players all playing DW MP?

DW is regarded to have better general characteristics than SCXIIc?

SP is regarded as no way to create really and interesting/challenging scenarios?

That's what I am getting from reading between the lines. I must say that I have gone through 30-50 SC 1.08 scenarios and I have generally been dissappointed with play and challenge. Thus, my inclination that despite having a written a scenario or two that I could do, at least, as well if not better.

[GrayOwl]

Very good doctrine - Has no been equal in reality.

Compatible only with AI Helos - NOT FOR CONTROL HELO FROM FFG!

The helicopter will not be clued any more never.


; Reinforce Alert Add On AI Helo Doctrine.
; -- HeloDipping.txt Original Name Doctrine ---
; Compatible for LWAMI 3.08
var DipTimer
var TargetCount
var WaypointReached
var Pinging
var Dipping
VAR TIME_ZVON
VAR FLAG_PADENIE
VAR T_NOPADENIE
VAR OLD_SPD
var LAST_CONT
var GaugesOn
var GaugesOff
var TR1
var TakeOff
var TimeInit
VAR ROK
VAR RX
VAR RY
VAR RokRng
VAR PROZVON
VAR TEMP
VAR Delay
var Lock
IF INIT THEN {
; DebugOut "AI Helo_Dipping Doctrine INIT"
PROZVON = 0
ROK = 0
GaugesOn = -1
GaugesOff = 0
TR1 = True
TakeOff = True
TimeInit = ( Time + 60 )
LAST_CONT = 0
FLAG_PADENIE = 0
TIME_ZVON = -1
Lock = 0
Delay = 0
TargetCount = 0
WaypointReached = 0
DipTimer = -1
} ENDIF
IF ( GaugesOn != -1 ) THEN {
IF ( ( OwnAlt < 50 ) AND ( NOT TakeOff ) AND ( OwnSpd <= 5 ) ) THEN {
Enable
Lock = 1
DebugOut OwnName
Debugout "SENSORS ON"
GaugesOn = -1
GaugesOff = 0
} ENDIF
} ENDIF
IF ( GaugesOff != -1 ) THEN {
IF ( ( OwnAlt > 50 ) OR ( OwnSpd > 5 ) ) THEN {
Preenable
DebugOut OwnName
DebugOut "SENSORS OFF"
Lock = 0
GaugesOff = -1
GaugesOn = 0
} ENDIF
} ENDIF

IF ( Lock == 0 ) AND ( GaugesOff == -1 ) THEN { Delay = ( Delay + 1 ) } ENDIF

IF ( Delay > 2 ) THEN {
Enable
SensorEnable "AI Dip Active" 0
SensorEnable "AI Dip PassiveS" 0
SensorEnable "AI Dip PassiveD" 0
DEBUGOUT "RADAR,ESM,MAD,Visual - ENABLED / Acoustic SENSOR DISABLED"
Delay = 0
Lock = 1
} ENDIF
; Covert TakeOff, Don't Ping Near Parent Ship!
IF ( ( Time > TimeInit ) and TR1 ) THEN { TR1 = false TakeOff = false } ENDIF
IF NEWTRACK AND ( ( OwnAlt <= 55 ) AND ( ( TgtSource $= "Passive Sonar" ) OR ( TgtSource $= "Active Sonar" ) ) ) THEN {
IF ( TgtClass $= "SUB" ) THEN {
SETTACTIC "HeloDipping"
} ENDIF
} ENDIF
IF ( ( OwnSpd <= 13 ) AND ( OWNALT > 150 ) AND ( ( TIME_ZVON == -1 ) OR ( TIME > TIME_ZVON ) ) ) AND NOT TakeOff THEN {
DEBUGOUT "Dipping Sonars"
SETPRIORITY 250
SETPERSIST 90
SETSPD 0
SETALT 45
DipTimer = ( TIME + 90 )
TIME_ZVON = ( TIME + 300 )
} ENDIF
; Assign Dipping Tactic
IF NEWTRACK THEN {
; Sub?
IF ( ( ( TgtClass $= "SUB" ) OR ( TgtClass $= "SURF" ) ) AND ( ( NOT ( ORDER $= "TRANSIT" ) ) ) ) THEN {
DEBUGOUT "Assigned Dipping Sonar"
TACTICCENTER
LAST_CONT = ( TIME )
ROK = 0
SETTACTIC "DippingSonar"
Pinging = 0
WaypointReached = 0
DipTimer = -1
TargetCount += 1
} ENDIF
; Missile?
IF ( ( TgtClass $= "MISSILE" ) AND ( ( TgtMissileClass "AAM" ) OR ( TgtMissileClass "SAM" ) ) ) THEN {
SETTACTIC "AirEvadeMissile"
} ELSEIF ( ( TgtClass $= "MISSILE" ) AND ( TgtMissileClass "SSM" ) ) THEN {
DEBUGOUT "Missile CheckUp"
IF ( ROK == 0 ) THEN {
RX = TGTX
RY = TGTY
ROK = 1
} ENDIF
} ENDIF
; Enemy Plane?
IF ( ( TgtClass $= "AIR" ) AND ( ( TgtID $= "HOSTILE" ) OR ( TgtID $= "UAE" ) OR ( TgtID $= "UPD" ) ) ) THEN {
SETTACTIC "AirEvadeAir"
} ENDIF
} ENDIF
; lost track
IF LOSTTRACK THEN {
DEBUGOUT "Helo Doctrine, Lost Track!"
; decrement target count
IF ( TgtSource $= "Active Sonar" ) OR ( TgtSource $= "Passive Sonar" ) THEN {
TargetCount -= 1
; lets make sure our variable doesn't go negative!
IF TargetCount < 0 THEN {
TargetCount = 0
} ENDIF
} ENDIF
} ENDIF
; Check Up Point Missile Start
IF ( ( TargetCount == 0 ) AND ( ROK == 1 ) AND ( NOT ( ORDER $= "TRANSIT" ) ) ) THEN {
TEMP = XYBRG ( RX - OWNX ) ( RY - OWNY )
STEERXY ( RX - OWNX ) ( RY - OWNY )
SETSPD ( 110 )
ROKRNG = xyrng ( OwnX - RX ) ( OwnY - RY )
IF ( ROKRNG < 500 ) THEN {
ROK = 0
PROZVON = 1
} ENDIF
} ENDIF
; Search only
IF ( NOT ( ORDER $= "Transit" ) ) THEN {
; Dip
IF ( WaypointReached == 1 ) OR ( PROZVON == 1 ) THEN {
; game has just clued us we've reached a waypoint
DEBUGOUT "Waypoint Reached,Lower Dipping Sonars"
SETPRIORITY 249
SETPERSIST 90
SETSPD 0
SETALT 45
Pinging = 0
WaypointReached = 0
PROZVON = 0
Dipping = True
} ENDIF
; Start pinging
IF ( ( Pinging == 0 ) AND ( Dipping == 1 ) AND ( OwnAlt < 50 ) ) THEN {
DEBUGOUT "Commence Check Point"
SETPRIORITY 249
SETPERSIST 90
SETSPD 0
SETALT 45
Pinging = 1
DipTimer = ( TIME + 90 )
} ENDIF
; Stop Dip
IF ( ( DipTimer != -1 ) AND ( TIME > DipTimer ) ) THEN {
DEBUGOUT "Check Is Completed, Raise Dipping Sonars"
SETPRIORITY 250
SETPERSIST 60
SETSPD MAXSPD
SETALT 300
Pinging = 0
DipTimer = -1
Dipping = False
TACTICCENTER
} ENDIF
} ENDIF
; Don't Dive!
IF ( ( OwnSpd > 100 ) AND ( OwnAlt < 80 ) AND ( FLAG_PADENIE == 0 ) ) THEN {
OLD_SPD = OwnSpd
; DEBUGOUT "Helo Stall"
FLAG_PADENIE = 1
T_NOPADENIE = 0
SetPriority 251
SetPersist 20
SetAlt 150
SETSPD ( OWNSPD - 15 )
} ENDIF

IF ( FLAG_PADENIE == 1 ) THEN {
T_NOPADENIE = ( T_NOPADENIE + 1 ) } ENDIF
IF T_NOPADENIE > 30 THEN {
SetPriority 251
SetPersist 40
SetAlt 200
SETSPD OLD_SPD
T_NOPADENIE = 0
FLAG_PADENIE = 0
; DEBUGOUT "Refresh Spd"
} ENDIF





================================================== ==================================

Second Doctrine For Helo....

; HeloDipping --- Dipping Sonar.txt --- Search Doctrine For -Reinforce Alert- Addon /KPV_1974 design/
; Compatible for LWAMI 3.08
; -- DippingSonar.txt Original Name Doctrine ---
var DipTimer
var ReattackTimer
var ValidTarget
var GOGO
var T_G
var G
var TEMP
var L_C
var F
var LCX
var LCY
var TLC
IF ( INIT ) THEN {
LCX = TGTX
LCY = TGTY
L_C = ( GetEntVar "LAST_CONT" )
TLC = L_C
G = 0
T_G = 0
F = 1
SensorEnable "MAD" ON
GOGO = 0
DipTimer = -1
ReattackTimer = -1
IF ( ( TgtID $= "HOSTILE" ) OR ( TgtID $= "UPD" ) ) THEN {
ValidTarget = -1
} ELSE {
ValidTarget = 0
} ENDIF
} ELSE {
IF ( LOSTTRACK ) THEN { F = 1 } ENDIF
L_C = ( GetEntVar "LAST_CONT" )
IF ( DipTimer == -1 ) AND ( ValidTarget == -1 ) THEN {
TEMP = ( TIME - L_C )
IF ( ( ( TgtID $= "UPD" ) OR ( ( TgtID $= "HOSTILE" ) AND ( TEMP > 100 ) ) ) AND ( TgtRng <= 1200 ) ) THEN {
; DEBUGOUT "TIME:"
; DebugValueOut TIME
; DEBUGOUT "L_C:"
; DebugValueOut L_C
; DEBUGOUT "TEMP:"
; DebugValueOut TEMP
IF ( TgtID $= "HOSTILE" ) THEN {
; DEBUGOUT "Drop Contact1"
DROP } ENDIF
F = 1
DipTimer = ( TIME + 90 )
SETPRIORITY 251
SETPERSIST 80
SETCRS TGTBRG
SETSPD 0
SETALT 45
} ELSEIF TgtRng > 500 THEN {
; Close
SETPRIORITY 251
STEERTO
SetSpd MAXSPD
SETALT 300
SensorEnable "MAD" ON
} ELSE {
DEBUGOUT "Smell Him!"
F = 1
DipTimer = ( TIME + 90 )
SETPRIORITY 251
SETPERSIST 80
SETCRS TGTBRG
SETSPD 2
SETALT 45
} ENDIF
} ELSEIF ( ValidTarget == 0 ) THEN {
SETPRIORITY 251
SETSPD MAXSPD
SETALT 300
IF TgtRng > 13715 THEN {
; DEBUGOUT "DropContact2"
Drop
} ENDIF
} ELSEIF ( TIME > DipTimer ) THEN {
; Stop dip
DEBUGOUT "Stopping prosecution dip"
SETPRIORITY 251
TACTICCENTER
SETSPD MAXSPD
SETALT 300
SensorEnable "MAD" ON
DipTimer = -1
} ENDIF
; Engage
TEMP = ( TIME - L_C )
IF ( ( TEMP <= 100 ) AND ( TgtRng <= 1100 ) AND ( TgtID $= "HOSTILE" ) AND ( ( ReattackTimer == -1 ) OR ( TIME > ReattackTimer ) ) ) THEN {
GOGO = 1
G = 1
SETPRIORITY 252
SETPERSIST 10
SETSPD MAXSPD
SETCRS TGTBRG
SETALT 200
SETPRIORITY 250
SETPERSIST 200
ATTACKBEST
DipTimer = -1
ReattackTimer = ( TIME + 300 )
} ELSEIF ( ( G == 0 ) AND ( TgtRng > 1200 ) AND ( ( TgtID $= "HOSTILE" ) OR ( TgtID $= "UPD" ) ) ) THEN {
GOGO = 1
SETPRIORITY 252
SETALT 300
SETSPD 110
SETCRS TGTBRG
DipTimer = -1
} ENDIF
IF ( G == 1 ) THEN {
T_G = ( T_G + 1 )
IF ( T_G > 1200 ) THEN { G = 0
T_G = 0
LAND
} ENDIF
} ENDIF
LCX = TGTX
LCY = TGTY
} ENDIF

[suBB]

Originally Posted by MarkShot

So, are the only serious mission/hardcore players all playing DW MP?

DW is regarded to have better general characteristics than SCXIIc?

SP is regarded as no way to create really and interesting/challenging scenarios?

That's what I am getting from reading between the lines. I must say that I have gone through 30-50 SC 1.08 scenarios and I have generally been dissappointed with play and challenge. Thus, my inclination that despite having a written a scenario or two that I could do, at least, as well if not better.

Again I can only account for goings on in MP, in particular mission objective MP in DW. And when a mission is on the line, especially with some key features coded into a map(like the requirement of staying alive & reaching safe distance unharmed) you better believe you’ll take it serious in terms of out-smarting human opponents. Other than myself, keep your eye open for: Molon Labe, Sonoboy, MahuJa(kage), To Be (btw good dive guys did you prep the 53cm correctly?), admbleeannoying, foogleman, -moose- and about 15 others from gamespy by means of achtachtel who all play DW mission MP. Really, you don’t need a lot of people, you just need the appropriate scenario for the headcount. Typically we will dive scenarios based on headcount.

When you mention SCX still SC remains as source of editor. DW / LWAMI offers far more flexibility in editing (use of scripts & triggers for sake of AI than SC. So the answer is yes. AFAIK SCXIIC only gives you more platforms and that is it. I’ve even reinstalled SC for sake of SCX platforms and our concept of in-game dynamics, but when I learned that SC didn’t offer use of scripts, there really isn’t a point dealing with SC. Someone confirm for me that scripts do not exist with SC / SCXIIC running. If so I’ll change my tune about SC / SCXIIC.

SP isn’t my department and I’m sure some good scenarios can be created, like RSR for example. But when it comes to being challenged, AI even after coded to be as effective as humans still can’t cut it w/o some human intervention involved. It’s common sense. Plus some SP scenarios literally spawn platforms out of thin air, which isn’t realistic in my book. Instead I account for departure of platforms and random arrival times for both humans and AI alike.

[MarkShot]

SC's editor does not have scripts.

I think the one thing it has missing in DW is user actions. You can check if particular controls have been pressed or certain masts have been raised.

It was probably implemented in order to script the interactive tutorials that came with SC. DW just had videos, but no interactive tutorials.

The SC ones were similar to the high quality interactive tutorials that came with games like Long Bow. Come to think of it, it was probably a requirement of Electronic Arts at that time.

[Molon Labe]

Originally Posted by MarkShot

So, are the only serious mission/hardcore players all playing DW MP?

DW is regarded to have better general characteristics than SCXIIc?

SP is regarded as no way to create really and interesting/challenging scenarios?

That's what I am getting from reading between the lines. I must say that I have gone through 30-50 SC 1.08 scenarios and I have generally been dissappointed with play and challenge. Thus, my inclination that despite having a written a scenario or two that I could do, at least, as well if not better.

AI is always going to have its limitations, even with quality doctrines controlling it. The general rule for SP is that challenge and dynamics are opposing factors. To create a serious threat for the player, you have to set up the mission to increase the likelihood of detection, put the platform up against its natural enemies, and/or put the platform at a tactical disadvantage against its enemies. By making the mission dynamic, the player is able to dictate part of the tactical picture because the player can think ahead, plan, and adapt. The player is also going to win in a "fair fight" by virtue of better detection and engagement capability.

So if you're looking for a dynamic fight that is also challenging, your best bet is human opposition in the context of a rather open-ended MP scenario.

SP is still good for creating specific scenarios, though. There's nothing wrong with thinking up new situations (or hearing about real life ones) and seeing how they would play out.

[MarkShot]

Interesting point. I had not considered that dynamic behavior and challenge or pulling in opposite directions, but you are right.

Also, there are mainly two ways to challenge the player:

(1) Overwhelm - numbers exceed the players ability to counter or events happen too fast.

(2) Outclass - the player is at a disadvantage; maybe has a platform with half the sensor range as the enemy.

So, going back to dynamic that leaves the designer with not really using dynamic creation/placement of objects, but instead having a small fixed set precisely scripted missions where only one is actually instantiated and the player doesn't know and cannot quickly determine which.

[Molon Labe]

Originally Posted by MarkShot

Interesting point. I had not considered that dynamic behavior and challenge or pulling in opposite directions, but you are right.

Also, there are mainly two ways to challenge the player:

(1) Overwhelm - numbers exceed the players ability to counter or events happen too fast.

(2) Outclass - the player is at a disadvantage; maybe has a platform with half the sensor range as the enemy.

If you want to see a perfect example of all three factors I mentioned before used to screw a SP player in the ass, hard, check out Bill's Race of the Cripples. Use LW/Ami so that the helo that launches has a working MAD and so the Udaloy doesn't overshoot when it fires its Silexes.

So, going back to dynamic that leaves the designer with not really using dynamic creation/placement of objects, but instead having a small fixed set precisely scripted missions where only one is actually instantiated and the player doesn't know and cannot quickly determine which.

I tried using this design philosophy for an MP FoF scenario about sinking supply ships... after sinking more than 12 hours into it to complete only about 1/6 of the work (just to set it up, not counting testing and revisions) I tabled it indefinitely.

I think RSBs and DLs are great time savers whenever you can use them, but obviously their utility is very limited, especially when you have a specific destination in mind (including moving targets) or when you're trying to spawn formations instead of lone platforms. Thus, DGs tend to be the best option for creating different starting positions. Rather than create entirely separate instantiations, whenever possible, I would try to use layered DGs and probability inclusions to randomly select dynamic design elements individually. You get more dynamicity (yes, I made that word up) for your buck that way.

[FERdeBOER]

Originally Posted by Molon Labe

AI is always going to have its limitations, even with quality doctrines controlling it. The general rule for SP is that challenge and dynamics are opposing factors. To create a serious threat for the player, you have to set up the mission to increase the likelihood of detection, put the platform up against its natural enemies, and/or put the platform at a tactical disadvantage against its enemies. By making the mission dynamic, the player is able to dictate part of the tactical picture because the player can think ahead, plan, and adapt. The player is also going to win in a "fair fight" by virtue of better detection and engagement capability.

Of course fighting aginst a human is allways better, and in single player, a good mission is great... but only a few times, then You know all You need about the scenario You will not play it anymore (except if you really love it). But the point is to find a way to create an scenario (or better, a campaign) so dynamic that can be played many times, even by its creator.

I'm talking as submarine (Akula) player:
Is by far more dificult to design, but I think a dynamic mission can be challenge without overwelming the player or putting him in disadvantage. The point is that only the uncertainy of who, where and how many enemies are out there puts you in a little disadvantage. Also the fact that you have to choose a variety of weapons because you don't know what will you need untill the mission message arrives, or that you may have to choose different ways to reach one location, gives the mission a good challenge.
Also the uncertainy of knowing that maybe that enemy sub you found last time can be close... or maybe is at 20 miles... or maybe it isn't this time... or maybe it is, but this time the mission of that submarine is other than killing you... maybe an ally this time? :hmm:

It will be hard, VERY hard, but a good dynamic scenario will be great

(I'm going now to play now those Bill scenarios )

edit: what does RSBs and DLs means? Sorry, I'm not English...

[To be]

Originally Posted by suBB

To Be (btw good dive guys did you prep the 53cm correctly?)

Yes!

[Molon Labe]

Originally Posted by FERdeBOER

edit: what does RSBs and DLs means? Sorry, I'm not English...

Random start boxes and Dynamic locations.