GermanyThe map generation depends on the start system date and time of the main program.
I figured out that you can get to play the same map when you you can set date and time in a batch script before calling civ.. However you need to be precise with the screen, sound and mouse settings as this screen - the RNG seed is chosen by your date & time just after you do that. After a bit of training, I can get the same seed pretty reliably this way.
Are we going to allow setting this in the leaderboard? It's not something you can easily detect if we disallow it.
I have another suggestion if we end up with more categories.
I'd like to experiment with setting the time exactly for RNG manipulation. I did a bit of research in how the game functions in my free time, trying to reverse engineer the game. The random seed is set by the system clock (date + time), but after the preloader is finished (the 3 screens that configures display, sound and input device). I found that it is hard to do this by a batch script and time it correctly, but I got some success having the same seed in a bunch of tries.
I bet there's a way to get the timer set exactly when the main game loads (and this will be super interesting to make a TAS run of).
I think using a tool to manipulate the RNG by setting a seed should be allowed in separate category, with the rule of posting the method and exact date/timestamp used with the video.
I'm playing with a x230 machine, about 10 years old. For most games, 15-18k cycles works absolutely fine even when I record stuff via obs (instead of dosbox which I usually do for speedruns)
Okay, let's talk about a frustration I have running the game: Dosbox default cycles.
My WR run was extremely lucky, both AI players founded one city then didn't manage to expand. This happens sometimes and I am not sure if there's a deterministic way (apart from setting the clock as random seed, which is another topic).
This is the ideal situation, but it happens rarely. There's a way to notice it around mid-run: waiting times between turns. At least it is with a dosbox set to 3000 cycles.
Depending on the AI they can expand and find you (that is basically run over at that point), though I have the hypothesis that you can still get a really good times if the turns were a lot faster when they expand / fight each other.
I'd propose an increase in dosbox cycles, ideally to a fixed value. I don't like it when the people with the fastest hardware get an unfair advantage.
I'd go with something resembling a fast 486 machine. Maybe around 12-20k cycles.
Yeah we consider dosbox as emulation. I wish I had a real hardware (and I believe a fast 386 or even 486 will shave off time of the space race category at least compared to 3000 cycles in dosbox).
I have a CD version of Civ1, it's not the version I've been running on. I don't think that would stop me from playing it on real hardware though.
Hey, I'd like to get into Speedrunning this game and I'd highly prefer to use OpenRCT2 (mainly for the early completion setting) but I highly dislike the game speed setting rule.
I just watched one of the runners get about 1.5 in-game days per second, all At turbo speed all I can get is about one day every 3 seconds on my older hardware laptop.
That basically suggests in order for me to be competitive, I need to throw money at the problem.
Can we get a normal speed or some kind of in-game time category that will make it less hardware dependent?
I agree. I think we should at least have another category without fast-forward.
So after playing 80+ runs so far, I'd propose to use first appearance of this screen as starting position: https://imgur.com/a/FzhmXbi
And end for the end this frame that appears just before the end cinematic: https://imgur.com/2WCPMPN
Reason why not to wait for the sound to start - if you manage to get a glitched ending state, this frame can appear for longer than usual, also on the beginning of the space cinematic I've had it glitched that you would start seeing the spaceship 'dot' & sound starting after a long pause.
I switched to 474.01 (from .4) in my past ~15 runs or so. It has the glitch sometimes but it's different. So far I have encountered a different glitched ending, I call it the "pillar glitch":
Every couple of the space race runs I do I end up with a glitched ending cinematic.
Graphics are glitched and it takes longer for the spacecraft (dot) to appear, sometimes the sound appears though but sometimes it takes a minute or so to start, just showing a dark blue screen.
What I noticed is that dosbox freezes after trying to quit the game and when that happens I have to kill dosbox (and corrupt the video it's recording) unless I hit my 'stop recording'-key before it ends.
What I couldn't find out so far is, why it happens and if there's a way to prevent it. I am not sure if it's a bug with dosbox or a memory leak issue of civ. Any input is appreciated!
When looking at your runs, how would you get the cue on when the game exactly starts?
For my own reference, I ended up starting my runs on appearance of "you're stranded..." as hitting enter makes it slowly fade out.
