Game Speed Q?

[infamousjbe 592]

Curious to know, over a specific amount of time, would the same events take place no matter what speed the sim is running?

 

Example: Say between the dates 1/1/00-1/1/01 running on cheetah there was a fire, if I ran the sim on turtle between the same dates, do you think that the same fire would take place?

 

Some more info that might help explain why I ask. Been testing a mod that will increase the amount of random fires, so when I start testing new changes I've made in Reader, I'll usually run the sim on cheetah just to make sure I get some kind of results before testing it on turtle. I play SC4 in turtle mode 99% of the time and testing in this speed can take more time than I have to wait around to see how frequent, if any, random fires pop up. 

 

I would like to know if I could use the frequency on cheetah to figure out the frequency if the sim was set on turtle.

 

Any info would be appreciated.

 

 

EDIT: (That is, if you didn't alter ANYTHING at all from the time the city loads)

[z1 566]

As I have been doing a lot of testing, starting the same city over and over from the same date and with the same starting conditions, I can tell you that there's a certain random element in the game, even if you're running at the same speed.  This makes testing for subtle results rather difficult, because variations in outcome can be fairly large.  Generally, it takes repeated tests in the same environment to establish how often an occasionally occurring event takes place.  I wouldn't expect running speed to affect this, as the game slows down when it needs to, such as when it needs to run the traffic simulator in a large city.

[AcronShinra 60]

If outcome of every simulator action was static, with almost no room for randomness, I would call it bad coding.
But on other hand, have you played Civilization 3? There was "Preserve Random Seed" option that made every battle outcome or other calculation being same even with save/load due to random number being kept in game save file. In SimCity4, because of its simulation complexity it rather wouldn't work.

[A Nonny Moose 8,261]

Well, it is a stochastic simulation, and if you could keep the same set of random seeds, computer programs being what they are, the same results should eventuate no matter what speed you ran.  However, we don't even know if there is only one random seed, and we don't know how that seed is chosen.

 

At the time of the game writing, one common method for seeding a random generator was to take the date-timer value, square it and take the middle.  Then as the session expanded, a timer would expire for the program to select a new seed.  This generally causes a pretty good random distribution.

 

One of the failures of the random system seems to be a preference to choose a default lot over one in the plugin pool.  This can be annoying at times, but what the heck, this is a game, and we need to just have fun.

[infamousjbe 592]

Thanks all for the information. The randomness aspect isn't something I've thought in depth about. It certainly makes sense to not have it to static or it would definitely get boring once the pattern was figured and broke.

 

Most of my testing/de-bugging experiences have been with console games, which most of the time can be replicated the exact way time and time again, which makes them extremely easy to figure out.

 

Was trying to avoid having to do extensive testing for random frequency in turtle since sometimes the randomness can be 40 minutes long. Would like to get the frequency to about 20-30 minutes on turtle.

 

The enjoyments of getting to the finite phase of testing a mod. Since I've been spending most of my time trying to narrow down which factors to adjust to achieve the goal of the mod, haven't had much time to actually play. Guess I'll have to combine my play and test time and hope not to mess things up in the mean time.     No more auto saving for me. 

 

[A Nonny Moose 8,261]

If you don't use the launcher you won't get autosave.  Use the options and remember to toss in a CTRL+s in the mix, especially if you are using puzzle pieces.

 

I tend to avoid indirect things, especially game launchers I didn't write.  I have my own "launcher" if you like in the form of a BASH script.  Here it is:

 

#!/bin/bash
date
env WINEPREFIX="/home/john/.wine" \
wine "C:/Program Files (x86)/Maxis/SimCity 4 Deluxe/Apps/SimCity 4.exe" \
 -intro:off -CustomResolution:enabled -r1920x1080x32 -d:software \
 -f -CPUcount:1 &>~/sc4.log
date
 

This operates two ways.  I have it as a GUI launcher, in which case the console output goes to the bit bucket, but if I start this from a console, then get the the starting date and ending date.  I always get the log which will show any wine problems.

 

The difference between Linux and Windows philosophy of end-user operation is that Linux doesn't hide the fact from its users that it uses scripting to launch any program.  Windows pretends launching a program is some kind of mysterious miracle.