It seems that you're using an outdated browser. Some things may not work as they should (or don't work at all).
We suggest you upgrade newer and better browser like: Chrome, Firefox, Internet Explorer or Opera

×
Hey guys, just wondering if anyone else has this same issue. I've never found AAMs to be much good in this game, because they're way too easy to spoof. Launch a missile at someone and the other guy always shakes it off once he releases a single packet of chaff or flares. Is there some sort of setting somewhere I'm missing, where I can change this?
Missiles mostly require luck, especially the juliettes must be launched from much closer range than you would normally use missiles for (rear aspect also).
I think the missiles are either bugged or they deliberately programmed it this way, anyhow missiles are mostly dead weight you carry on your pylons in this game.
Post edited May 10, 2013 by Strijkbout
In case anyone's still reading this, I've done my research on the subject since it did annoy me anti-air missiles were largely useless. I found out it's likely a problem related to how fast the game's processed: the AI's reaction speed is related to it in some way.

That means that if the CPU cycles are high enough in the DOSBox emulation to make the game super smooth, that has the unfortunate side-effect of giving AI pilots very high reaction times. Which results in large amounts of countermeasures being deployed very soon, and consequently nab missiles constantly unless the target's almost literally within arm's reach.

Strike Commander's design accounted for a degree of choppiness, given its groundbreaking engine wasn't fully manageable by even the most advanced hardware of its time.

Bottomline is, dial down the CPU cycles a bit and you'll notice the AI being more reasonable when it comes to (not) evading missiles. I don't remember what's the optimal setting, but I think it was in the vicinity of 10,000 cycles. Probably a bit less than that.
avatar
Shadow86: In case anyone's still reading this, I've done my research on the subject since it did annoy me anti-air missiles were largely useless. I found out it's likely a problem related to how fast the game's processed: the AI's reaction speed is related to it in some way.

That means that if the CPU cycles are high enough in the DOSBox emulation to make the game super smooth, that has the unfortunate side-effect of giving AI pilots very high reaction times. Which results in large amounts of countermeasures being deployed very soon, and consequently nab missiles constantly unless the target's almost literally within arm's reach.

Strike Commander's design accounted for a degree of choppiness, given its groundbreaking engine wasn't fully manageable by even the most advanced hardware of its time.

Bottomline is, dial down the CPU cycles a bit and you'll notice the AI being more reasonable when it comes to (not) evading missiles. I don't remember what's the optimal setting, but I think it was in the vicinity of 10,000 cycles. Probably a bit less than that.
Thanks Shadow, I haven't given this a try yet, but it sounds plausible. It seems that it also affects my own countermeasures, enemy IR missiles always go after my own flares. Beating up on disabled Eagle pilots is quite entertaining in its own way!
I know this is pretty much a necro for an old thread, but I wanted to confirm that Shadow86 is correct. 10,000 cycles makes the game choppy but playable and eliminates the missile problem. I've flown several instant actions against AI pilots at 10,000 cycles and have consistently gotten multiple kills per instance with the crappiest missile, the AIM-9J. While occasionally the AI will dump flairs and evade them, it's often realistically slow on the draw at shorter ranges, as it should be. I've even seen the occasional J ignore flairs and hit the MiG anyway, which never happened at higher cycle settings. In other words, yes, the AI is tied to how fast the CPU can cycle and slowing this down closer to 90's era performance makes it considerably slower on the draw. This holds true in the campaign as well.

Missiles are fun again :)
Early AAMs were pretty easy to spoof; they could even veer off in the direction of the sun. That may be one factor in it.
high rated
The theory about AI reaction time being related to the CPU cycles speed is unfortunately 100% accurate.

However, it only applies to the CD version of Strike Commander.

I was pretty sure that the missiles in this game were much more fun back then, so I found my original dusted floppy SC disks back from 1993, and decided to do some testing:

After addition of floppy speech pack and floppy Tactical Operations pack the game gets updated to the version 2.1 (which can be checked with ctrl-v in-game). In this state the game has moreless all the features of cd version apart from the digital character voices between missions (which may not necessarily be a bad thing ;)

However, what makes the REAL difference is the gameplay!

In floppy version, regardles of CPU speed, missiles work exactly as they should, and provide real fun :)

If I set enemy AI to rookie, then they are killable with even 9Js (properly targeted from behind etc of course).
If AI is set to "ace" then the bandits' reaction time is way quicker and killing them becomes really challenging, but still possible; they do not just spam chaffs/flares upon shooting a missile, which happens in CD version all the time.
My CPU cycles in dosbox were set to 75.000 during testing, which provided super-smooth gameplay.
The cd version started to behave properly (similar AI reaction to the floppy version) only after I reduced cycles down to 10.000 which unfortunately also made the game very choppy, close to unplayable and surely not enjoyable..

So, if You are willing to really enjoy this game, get some great missile kills and all of it in smooth high-cycle gameplay, your answer is: the floppy version :-)

Also checked this on my old-gaming PC rig with P3-800 and real MS-DOS inside, and everything is the same there :
cd version - AI 100% perfect reaction regardless of the difficulty => 0 fun :-/
floppy version - normal AI behaviour depending on the set difficulty => 100% fun :)
I suspect that on something closer to 486 DX2-66 the cd version would start behaving properly.

Analyzing the file structure of these two versions gives impression that these are two different games, each programmed and coded independently from the other. Which may also explain why there is such difference in their behaviour.

Which leads to a conclusion: the original floppy version should really be added by the GoG to their SC offer, because the CD version is "broken" somehow. Not intentionally, but still.
avatar
matello: After addition of floppy speech pack and floppy Tactical Operations pack the game gets updated to the version 2.1 (which can be checked with ctrl-v in-game). In this state the game has moreless all the features of cd version apart from the digital character voices between missions (which may not necessarily be a bad thing ;)
Sadly the only TO disks I "found" updated to version 2.3 , exact same behavior as the CD version.

There was someone working on a map and airplane viewer some time ago, he had some useful code to extract the game datafiles/assets (same structure than privateer1 apparently).

After some days off digging and trial and error this is what I found out so far:
- when extracting data files from bigtre.tre and putting them in the sc directory and removing bigtre from SC.CFG , the data files are loaded instead (even every mission restart from alt-o screen, sc directory needs to mounted as C:\ )
- the files within liltre.tre / bigtre.tre and the disk versions are in many parts identical
- a2a missile files seem to be the same across all versions
- the dogfight module loads ROOKIE / VETERAN / ACE AI profiles according to ui choice
- there is a NULL AI file that cant maneuver and only flies straight , even that one is immediately dropping chaff/flare on missile launch but in oddly 1.5 sec intervals when replacing ROOKIE AI with it
- manipulation of AI files did not yield any result so far, maybe playtesters guide AII variables will help
- renaming airplanes works as well , flying the gray f16 / mirage2000 / a10 worked out ok, f16gray has a little less thrust than players machine, invalid weapons will be stripped on mission start.
- aerodynamic properties and thrust can be manipulated (reached mach 9 by mistake -> but no controls)
- while default weapons are defined in planes there are loadout profiles that get stacked on top of each other (logged file access to see the order)
- chaff / flare quantity can be changed, values in hex is 1E after the keyword DCOY
- editing the STRIKE.EXE with an hex editor and renaming DCOY to anything else activates hardcore mode: no one has "decoys", missile evasion only possible through maneuvers :O

Also when in close range the enemy planes seem to be able to tank missiles in CD version , e.g. 3-4 sidewinders for mirages is a bit much, not sure about disc version...
Could be a collision detection problem as well as some sidewinders from 6 o clock fired by enemy planes seem to fly through my cockpit if I do not maneuver

Maybe I will find something comparing TO version 2.3 to pure disk regarding missiles, if its anything in the data files one could probably build a TRE file which only patches those files...
If its really in the exe file I fear one would need a dos debug / disassembler / assembler god, I do not understand anything when using debug version of dosbox yet.
avatar
matello: The theory about AI reaction time being related to the CPU cycles speed is unfortunately 100% accurate.

However, it only applies to the CD version of Strike Commander.
This is a golden tip, and you're indeed correct.

Even though I've only been able to find TacOps versions which update the game to version 2.3, I've played a couple of missions in the floppy version (with speech pack and tac ops) and haven't noticed any AI overreactions to missiles.
Post edited January 28, 2017 by Shadow86
high rated
After MANY hours learning/googling assembler and using IDA Pro Demo version I think I finally found a fix.
Surprisingly enough the code of the CD version and tacops is 100% identical in this area, but the passed variables are loaded in completely different orders (to complicated for me to retrace), BUT manipulating a bitshift operation that loads the variable that gets checked if spamming, moderate or almost no chaff / flares should be ejected did indeed fix something.

The combination of ACE / VETERAN / ROOKIE AI in instant mission and the 3 gameplay levels in the in flight menu seem to yield different results for each combination, almost like in the taops 2.3 disk version now.

I had almost given up after finding so much places in the code I could manipulate (autopilot landing with broken engine / landing gear anyone) , but never getting closer to that spamming AI....

So far I am still testing some way to build a generic patcher, if anyone wants to hex edit their STRIKE.EXE to test:

change offset 0xDDBC
from 08 to 04

Tested using instant mission (be sure to RESTART or quit and start another one after first game start) with 3 MIRAGES2000 and 4xAIM9M 4x120AMRAAM and 50000 cycles.

Seems there are many other bugs with high cycles as well , missile hit detection is way off if close in CD and disk versions sometimes, also AGM65 Maverick missiles sometimes do not track at all in CD version on high cycles, also I think the CD/tacops buffed flare / chaff effects , evasion is much more instant there than in the older disk version (dropping 2 flares against AI missile is almost overkill), maybe missile tracking issues are linked , no idea.
Post edited February 04, 2017 by Harr-Nuta
Thank you for your investigations and posts.
Flying some campaign missions it looks like I am not finished yet , most of them still behave like ACE pilots from instant action coupled with the selected gameplay settings, will try a thing or to I tried before that had a little effect (disabling some code that skips some AI setup in CD version that is not present in the disk version(.

At least I saw my wingman scoring a AIM9J hit, first time in years... or maybe decades ? (got my CD version around 1995 and shortly after a Pentium 166 mmx which would render missiles near useless)
Resorted to test within campaign when altering code and using the debugger, this helped a lot.

If anyone is still interrested in testing (choose veteran in gameplay or below, otherwise they are still very fast with dumping chaff/flares):

Changing Hex offset 0xD4BC
74 -> 75 seems to help a lot in the campaign

Changing Hex offset 0xDDBC
08 -> 04 seems to help a lot in training / inconclusive in campaign (other AI profiles than training missions I presume)

Testing is done with 50000 cycles , keep in mind that with 10000 its a murderfest now.

I am currently testing/playing the campaing with both changes to see if its playable.
Thanks for this awesome commitment.

Recently I finally managed to get an original Thrustmaster gear combo (Mark II: joy, throttle and rudder) and currently waiting for the delivery. When it is here I'm definitely going back to do some more SC testing.
With that opportunity I will also check Your suggested hex modifications on CD version to see whether it actually helps with missiles in campaign.
The following code can be pasted into PowerShell (Included in Windows)
Run it from the SCCD directory. Backup STRIKE.EXE first!

#This is based on Harr-Nuta's comments
# Adjust AI
$strike = Get-Content -Raw -Encoding Byte .\STRIKE.EXE
$strike[0xddbc] = 4
$strike[0xd4bc] = 117
$strike[0x4535D] = 68
Set-Content -Encoding Byte -Path STRIKE.EXE -Value $strike

# Set back to normal
$strike = Get-Content -Raw -Encoding Byte .\STRIKE.EXE
$strike[0xddbc] = 8
$strike[0xd4bc] = 116
$strike[0x4535D] = 68
$strike[0x4535D] = 67
$strike[0x4535D] = 79
$strike[0x4535D] = 89
Set-Content -Encoding Byte -Path STRIKE.EXE -Value $strike

# Disable Counter Measures
$strike = Get-Content -Raw -Encoding Byte .\STRIKE.EXE
$strike[0x4535D]=69
Set-Content -Encoding Byte -Path STRIKE.EXE -Value $strike
Post edited April 26, 2017 by BurialVermin68