MKSheppard: It was now significantly faster compared to the Floppy version on the same hardware.
I remember the same thing (got it as CDROM classics much later), but played the original floppy verson on an (ugh!) intel SX 20MHz cpu with 4MB RAM.
The machines harddrive after the installation basically had strike commander and dos on it and was almost full !
I think I got a cd rom drive some years later and had a dx/2 80 MHz and 8MB RAM, so of course it was faster then, so I have the same memory , but also know that I had a faster machine.
Later upgrading to 166 MHz, I still remember asking someone (I think he still had the disk version) back THEN if his missiles where hitting at all... so I was hit by this issue around 1996...., and thous began the 'guns only' play through era for me....
In dosbox with identical cycles , I dont see the difference , BUT the cdrom version had an extra setup for vesa local bus graphics cards, in dosbox this does not seem to do anything , wish someone with real hardware could test what that option actually does speed wise.
A lot of code branches of the disk and cdrom version are really VERY similar , but I hust am not capable enough to deduct much of anything there... played with the timers a bit (only made the game crash or run at around 3x the speed), but this was more of 'I have no idea what I am doing' and trying things out.
My guess (also not a game developer, but developer for other stuff) is that it can be a simple bug that uncapped some loop that was previously on a timer.
Now it just steps into the code on every cycle of the 'main sim loop' so to speak.
It can just be some simple branching condition and hitting code that calculates some probabilities for loosing tracking, maybe if invoked it is calculating 2/10 times tracking loss for condition X in Y amount of time/cycles, but in the CDROM version this is hit so often its like 9/10
Going through the code is like finding the needle in a haystack, even my simple stuff took forever to find, and did not fix the origin of the problem (tracking probability calculation), but only the counter measure spam (which brakes some gameplay as mentioned).
Maybe one day we have a dos assembly trained AI we can instruct to compare the EXE files and point out the differences with explanation , it can only get easier in the future....
Harr-Nuta: Try these changes in campaign, make saves and play the missions again alternating between patches, focus on dogfight heavy mission and play it some times with each patch to get a feel if its better and worth uploading, I can only decide for myself that it feels like a better "game" this way. Since the file on PCGamingWiki is from July 27, 2017, I guess these changes were never uploaded?
