Just wanted to share that I successfully tested a random encounter spawner that "rolls a dice" to see if the encounter happens. This example is a 50/50 chance of the encounter with a timer that resets the zone trigger (should the player return later, the roll can happen again if you so choose).

Enjoy!

Here's a diagram of it



or link:

http://i622.photobucket.com/albums/tt305/Gustang/basic_random_spawn_diagram.jpg

Edit: Just wanted to mention that the only reason the 3.5 second timer is in there is to fire the report that the roll failed. Otherwise, it's unnecessary.

Now THAT'S a good thing ... just what's needed for a real patrol mission.

Well done.

WH_JoeBob

Here's another example of a 4-way choice of encounter. Still testing this one...
It works, but it's not apparent to me that it works correctly. Separating the timers by milliseconds produced a lot of roll failures, where 1 second separation did produce better results (on my computer). This has the timers separated by 3 seconds.

http://i622.photobucket.com/albums/tt305/Gustang/Basic_25_Percent_random_spawn_diagr.jpg

Nice! A lot of possibilities with these random timers...

I think we need to be cautious with using msec timers for this kind of stuff, like you said. For my random recon patrol mission I used 200, 400, 600, 800, 1000 msec timers and that seems to work just fine. But who knows, this could work out not so well on another PC?

Specially for single player/campaign missions I think we need to work with at least "seconds" timers. When I was testing my mission in single player mode I used time acceleration to quickly see the results. But the acceleration completely messed up the randomness! I either got the no1 or no5 timer!
Luckily, in co-op missions you can't accelerate time and the server usually has (or it wouldn't be a server) enough processing power to do this correctly.

Watching this with great interest!

Thanks Gustang good stuff, naming a mission-building notes folder in your honour and backing up the thread

Ming

Thanks!

Also when 'there is time' enough, perhaps the 'divide et impera' method can be used in a cascade fashion, using just a similar structure of '2 self-canceling timers' at a time

Very cool Gustang.

Looks like you guys are developing some real potential with the ME.

Well, a couple of things to hash out here. I tested the 3-second separation approach last night and it didn't produce any better results. I got a lot of Fokkers or fails, so it's not bullet proof. Perhaps what we need to try next is to change the steps. Currently, it's:

Timer 1: 25 percent
Timer 2: 25 percent
Timer 3: 25 percent
Timer 4: 100 percent

But perhaps it could be:

Timer 1: 50 percent
Timer 2: 50 percent
Timer 3: 50 percent
Timer 4:100 percent

or increasing:

Timer 1: 25 percent
Timer 2: 50 percent
Timer 3: 75 percent
Timer 4:100 percent

I'm not well versed on statistics (rather probability), so I'm looking for the best way to produce (as close as possible to) an equal chance (like a 4-sided dice) for each step to occur. Any ideas?

I like Laser's suggestion of cascading pairs at 50 percent too. Is that a better approach?

Edit: It's clear now that the 25 percent/timer stepping isn't what I intended. Even at 50 percent/timer stepping, at 2 rolls, the probability of an encounter is already at 25%. Although that doesn't necessarily mean that 25 percent stepping isn't useful.

Gustang, is it possible to combine this with your Russian Roulette for Two mission so that the enemies spawned are randomly generated?