Just rebooted this game and wow!

[Mobhack]

Originally Posted by Dion

"WOW!" is a common commit among players that have returned to the game. I myself haven't played in quite awhile, looking forward to playing again. I took a leave of absence for almost a year because my computer broke down. Not only that but I got sick of the cheating A/I. I thought the A/I was pretty good until it started cheating. Probably wasn't the poor old A/I's fault, it was probably poor sportsmanship on my part.

I have been playing as well as coding the games for 20+ odd years now, and so would by now have twigged any such cheat code if it existed.

Now the original MSDOS SSI code (SP1, 2 and so on) acually did have a little cheat for the AI artillery, which we found and deleted. That cheat code would teleport artillery strikes to a slightly "better" impact zone much nearer your troops if they were spotted by it and the arty was about to fall nearby but not quite near enough. Again, all gone now in our games since we got rid of that original bit of SSI DOS code ages ago.

Additionally, the SSI code used your actual troop positions as a hint matrix for plotting any pre-game stonks, so those tended to land mostly inside or nearby the area you had deployed inside (and yes, having everyone sitting at say the South of the map with a spotter waay off in the North, totally messed with that algorithm!). Not there now, our AI plot uses randomness and things like road junctions, objective hexes and the intended assault axis of advance etc in its arty plotting and has since we got the DOS code and found that and culled it.

There was also a bit of code in the AI routines which it used to decide things like whether to drop off its passengers, or maybe to stop for now based on nearness of enemy units. That worked even on non-spotted enemy, but not now - it only responds based on known enemy units nearby. No more "AI spidey sense"!

And finally The old SSI AI buy routine bought some troops for the AI that came for free (but would be worth VP if killed of course) - typically a mortar and an AA/SPAA section, not much but the old SSI games were only 100 units each or whatever. Now replaced naturally, the AI only has the correct points to spend in our versions of the games.

So - what exactly are these "cheats" the AI allegedly does?

I'd really like to know, so it could be addressed, if real.

[MarkSheppard]

Originally Posted by Mobhack

Now the original MSDOS SSI code (SP1, 2 and so on) acually did have a little cheat for the AI artillery...

Additionally, the SSI code used your actual troop positions as a hint matrix for plotting any pre-game stonks...

There was also a bit of code in the AI routines which it used to decide things like whether to drop off its passengers, or maybe to stop for now based on nearness of enemy units...

Are there any other "war stories" you can tell us?

I can't help but wonder if the AI cheating in the old SSI code was due to:

A.) Not enough CPU cycles in 1995 to do an AI that did a lot of deep thinking -- the SP2 reqs were for a 486 DX/2 66 MHz -- but I think it was "kinda sorta" playable on 386s.

B.) Not enough time to code everything, remember the team also had to now make OBAT files for fifty years (1950-2000) for about two dozen nations; in an era before the internet.

[Mobhack]

Originally Posted by MarkSheppard

Are there any other "war stories" you can tell us?

I can't help but wonder if the AI cheating in the old SSI code was due to:

A.) Not enough CPU cycles in 1995 to do an AI that did a lot of deep thinking -- the SP2 reqs were for a 486 DX/2 66 MHz -- but I think it was "kinda sorta" playable on 386s.

B.) Not enough time to code everything, remember the team also had to now make OBAT files for fifty years (1950-2000) for about two dozen nations; in an era before the internet.

Likely a shortcut to make the AI more competetive with restricted CPUs of the day and also memory. PBEM replay as I recollect was brought in as a patch to SP1 after a year or so and it required a hefty (for the day!) RAM upgrade to - IIRC - 2MB.

Those were the days when you had multiple MSDOS boots using different calls to HIMEM and other such routines (there was some other one than HIMEM) for handling of memory above 640K, sound card setups etc. And DR_DOS was briefly the king O/S over Microsofts since you could make a boot loader script that asked you which config.sys setup you wanted, Microsofts one (of the day) needed fiddling IIRC with config.sys(es) since it couldnt ask a question on bootup, so you had to have several on disk and renamed the file(s) and rebooted for the particular game (like Task Force 1942). And DR_DOS was cheaper than MSDOS as well if I remember it right.

[MarkSheppard]

[quote=Mobhack;853077]

Originally Posted by MarkSheppard

Likely a shortcut to make the AI more competetive with restricted CPUs of the day and also memory. PBEM replay as I recollect was brought in as a patch to SP1 after a year or so and it required a hefty (for the day!) RAM upgrade to - IIRC - 2MB.

Also, back then development cycles were severely compressed -- we've gotten used to games taking three or four years to be released today; something possible only because video game consoles took over enough of the market to force graphical stasis.

Back then, you might have a year or less to code a game. So shortcuts had to be taken.

Those were the days when you had multiple MSDOS boots using different calls to HIMEM and other such routines (there was some other one than HIMEM) for handling of memory above 640K, sound card setups etc.

Now there's stuff I haven't heard in a long time...I do remember QEMM being the best for automagically moving everything into different areas to magically open up enough convnetional memory to run jst about everything with no need for boot disks.